Method, policy controller and computer readable medium for computer network

By using policy controllers in the data center to generate and distribute intent-based application firewall policies, the problem of manual generation of policies is solved, and more efficient and accurate firewall management is achieved.

CN114844841BActive Publication Date: 2024-11-19HEWLETT PACKARD ENTERPRISE DEV LP
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210467175.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2018-09-28
Filing Date
2019-06-24
Publication Date
2024-11-19
Estimated Expiration
2039-06-24

AI Technical Summary

Technical Problem

In complex data center environments, manually generating firewall policies is not feasible, error-prone, and difficult to adjust and upgrade. It is difficult for existing technology to achieve automatic generation, comprehensive, accurate and flexible firewall policies.

Method used

Collect business statistics for application workloads through the policy controller, generate intent-based application firewall policies, and distribute them to virtual router agents, and automatically apply them to business flows to allow or deny specific business flows.

Benefits of technology

It realizes automatic generation of more comprehensive, accurate and flexible firewall policies in the data center, reduces manual errors, and improves policy adaptability and management efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114844841B_ABST
    Figure CN114844841B_ABST
Patent Text Reader

Abstract

The present application relates to a method, a policy controller and a computer-readable medium for a computer network. A technology for generating an intent-based policy and applying the policy to a business of a computer network is disclosed. In one example, a policy controller of a computer network receives business statistics of business flows between multiple application workloads executed by a first group of computing devices. The policy controller associates the business statistics information with session records of multiple application workloads. The policy controller generates an application firewall policy for the application workload based on the session records of the application workload. Each application firewall policy defines whether the business between the application workloads is allowed or denied. The policy controller distributes the application firewall policy to a second group of one or more computing devices for application to business flows between instances of the application workload.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] This application is a divisional application of the Chinese patent application with application number 201910548861.6. Technical Field

[0002] The present disclosure relates to computer networks, and more particularly, to business policies for computer networks. Background Art

[0003] Virtualized data centers are becoming the core foundation of modern information technology (IT) infrastructure. In particular, modern data centers have been widely used in virtualized environments, in which virtual hosts (such as virtual machines or containers) are deployed and executed on the underlying computing platform of physical computing devices. Using large-scale data centers for virtualization can provide multiple advantages. One advantage is that virtualization can significantly improve efficiency. With the emergence of multi-core microprocessor architectures with a large number of cores per physical CPU, the underlying physical computing devices (i.e., servers) are becoming more and more powerful, and virtualization becomes easier and more effective. The second advantage is that virtualization provides important control over infrastructure. As physical computing resources become replaceable resources, such as in cloud-based computing environments, the supply and management of computing infrastructure becomes easier. Therefore, in addition to the efficiency and increased return on investment (ROI) provided by virtualization, enterprise IT staff usually prefer virtualized computing clusters in data centers to achieve their management advantages.

[0004] For example, a data center may physically house all infrastructure equipment, such as network and storage systems, redundant power supplies, and environmental controls. In a typical data center, clusters of storage systems and application servers are interconnected via a switching fabric provided by one or more layers of physical network switches and routers. More complex data centers provide a global infrastructure where subscriber support equipment is located in various physical hosting facilities. In many examples of data centers, the infrastructure may include a combination of physical devices, which may be referred to as "underlying resources," which are linked to and communicate with various virtual resources (which may be referred to as "overlay resources") such as virtual servers, agents, and / or policy controllers. Summary of the invention

[0005] Disclosed is a technology for generating an intent-based firewall policy for a computer network based on an analysis of the network business of the computer network and then applying the policy to the network business. In one example, before being deployed in a production environment, a customer may deploy one or more applications in a development environment for testing. A business collector may collect multiple session records of business statistics of multiple business flows of multiple application workloads operating in the development environment, and forward the session records to a policy controller. The policy controller may associate multiple session records of business statistics of multiple business flows with multiple session records of business statistics of multiple application workloads. The policy controller generates an application firewall policy for the application workload based on multiple session records of business statistics of multiple application workloads. Each of the multiple application firewall policies includes one or more application firewall policy rules, and each of the one or more application firewall policy rules defines whether a virtual router agent operating on a host computing device should allow or deny the business flow of the corresponding application workload.

[0006] The policy controller can distribute the application firewall policy to multiple virtual router agents for application to business flows. Therefore, when one or more applications are deployed in a production environment, the virtual router agent can apply the application firewall policy to the business flow. Therefore, the technology of the present disclosure can allow the policy controller to automatically generate intent-based application firewall policies for multiple application workloads. In other words, the policy controller can use the session records of the business statistics of the application workload to identify the "expected" business relationship between the application workloads, and use the "expected" relationship to generate application firewall policies for multiple application workloads that allow such "expected" business and block other forms of business.

[0007] The computer network of a data center is a complex environment that may include hundreds or thousands of components, such as applications, virtual machines, virtual routers, etc., that communicate with each other and with external devices. Developing a firewall policy for each component that communicates within a data center may be impractical, infeasible, error-prone, and cumbersome to adjust or upgrade. Compared to manually generating a firewall policy, the technology of the present disclosure may provide a technical advantage of automatically generating a firewall policy that is more comprehensive, accurate, flexible, and efficiently generated.

[0008] In one example, the present disclosure describes a method, comprising: receiving, by a policy controller of a computer network, business statistics of multiple business flows between multiple application workloads, the multiple application workloads being executed by a first group of one or more computing devices of the computer network; associating, by the policy controller, the business statistics of the multiple business flows to session records of the business statistics of the multiple application workloads; generating, by the policy controller and based on the session records of the business statistics of the multiple application workloads, one or more application firewall policies for the multiple application workloads, wherein the one or more application firewall policies define whether business flows between the application workloads of the multiple application workloads are allowed or denied; and distributing, by the policy controller, the one or more application firewall policies to a second group of one or more computing devices for application to business flows between instances of the application workloads.

[0009] In another example, the present disclosure describes a policy controller for a computer network, which is executed on a processing circuit and is configured to: receive business statistics of multiple business flows between multiple application workloads, the multiple application workloads being executed by a first group of one or more computing devices of the computer network; associate the business statistics of the multiple business flows to session records of the business statistics of the multiple application workloads; generate one or more application firewall policies for the multiple application workloads based on the session records of the business statistics of the multiple application workloads, wherein the one or more application firewall policies define whether business flows between the application workloads of the multiple application workloads are allowed or denied; and distribute the one or more application firewall policies to a second group of one or more computing devices for application to business flows between instances of the application workloads.

[0010] In another example, the present disclosure describes a non-transitory computer-readable medium comprising instructions that, when executed, cause a processing circuit to execute a policy controller for a computer network to: receive business statistics for multiple business flows between multiple application workloads, the multiple application workloads being executed by a first group of one or more computing devices of the computer network; associate the business statistics for the multiple business flows with session records of the business statistics for the multiple application workloads; generate one or more application firewall policies for the multiple application workloads based on the session records of the business statistics for the multiple application workloads, wherein the one or more application firewall policies define whether business flows between the application workloads of the multiple application workloads are allowed or denied, and distribute the one or more application firewall policies to a second group of one or more computing devices for application to business flows between instances of the application workloads.

[0011] The details of one or more examples of the disclosure are set forth in the accompanying drawings and the description below. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 is a block diagram illustrating an example computer network system in accordance with the techniques described herein.

[0013] Figure 2 To show in more detail Figure 1 A block diagram of an example implementation of a data center.

[0014] Figure 3 is a block diagram illustrating an example of multiple data centers according to the techniques described herein.

[0015] Figure 4 A block diagram illustrating an example computing device executing a virtual router for a virtual network according to one or more aspects of the present disclosure.

[0016] Figure 5 is a block diagram of an example network in which a controller manages servers or computing nodes and provides visualization of various aspects related to workflows and policy objects associated with a computer network through a dashboard, according to one or more aspects of the present disclosure.

[0017] Figure 6 is a table showing a number of example session records for traffic statistics of traffic flows according to the techniques of this disclosure.

[0018] Figure 7 This is to show that the technology according to the present disclosure will Figure 6 A graphical representation of multiple example session records of traffic statistics for a traffic flow converted into multiple firewall policies.

[0019] Figure 8 is a flow diagram illustrating example operations for generating multiple firewall policies in accordance with the techniques of this disclosure.

[0020] Fig. 9 is a flow diagram illustrating example operations for generating multiple firewall policies in accordance with the techniques of this disclosure.

[0021] The drawings and descriptions provided herein illustrate and describe various examples of the inventive methods, devices, and systems of the present disclosure. However, the methods, devices, and systems of the present disclosure are not limited to the specific examples shown and described herein, and other examples and variations of the methods, devices, and systems of the present disclosure, as understood by those of ordinary skill in the art, are considered to be within the scope of the present application. DETAILED DESCRIPTION

[0022] Figure 1 is a block diagram illustrating an example computer network system 8 in accordance with the techniques described herein. Figure 1The computer network system 8 in the example includes data centers 10A-10X (collectively referred to as “data center 10”) interconnected with each other via a service provider network 7 and with a customer network associated with a customer 11. Figure 1 An example implementation of a computer network system 8 and a data center 10A is shown, which hosts one or more cloud-based computing networks, computing domains, or projects, generally referred to herein as cloud computing clusters. Cloud-based computing clusters can be co-located in a common overall computing environment (such as a single data center), or distributed across environments (such as distributed across different data centers). For example, the cloud-based computing clusters can be different cloud environments, such as an OpenStack cloud environment, a Kubernetes cloud environment, or various combinations of other computing clusters, domains, networks, etc. Other implementations of the computer network system 8 and data center 10A may be appropriate in other situations. Such implementations may include Figure 1 Examples include a subset of the components and / or may include Figure 1 The data centers 10B-10X may include the same or similar features and be configured to perform the same or similar functions as described herein with respect to the data center 10A.

[0023] exist Figure 1 In the example shown, data center 10A provides an operating environment for providing applications and services to customers 11 coupled to data center 10A by service provider network 7 through gateway 108. Figure 1 The functions and operations described in the computer network system 8 can be shown as distributed in Figure 1 , but in other examples, the Figure 1 The features and techniques of one or more devices in the present invention may be performed internally by local components of one or more such devices. Similarly, one or more of such devices may include certain components and perform various techniques, or these techniques may be attributed to one or more other devices in the description herein. In addition, certain operations, techniques, features, and / or functions may be combined with Figure 1 In some embodiments, the present invention may be described as a component, device, or module, or may be performed in other ways by a specific component, device, and / or module. In other examples, such operations, techniques, features, and / or functions may be performed by other components, devices, or modules. Therefore, some operations, techniques, features, and / or functions attributed to one or more components, devices, or modules may be attributed to other components, devices, and / or modules, even if not specifically described in this manner herein.

[0024] Data center 10A hosts infrastructure equipment such as network and storage systems, redundant power supplies, and environmental controls. Service provider network 7 may be coupled to one or more networks managed by other providers and may therefore form part of a large-scale public network infrastructure (e.g., the Internet). In some examples, data center 10A may represent one of many geographically distributed network data centers. Figure 1 As shown in the example of FIG. 1 , data center 10A is a facility that provides network services to customers 11. Customers 11 may be collective entities, such as businesses and governments or individuals. For example, a network data center may host web services for multiple businesses and end users. Other exemplary services may include data storage, virtual private networks, business engineering, file services, data mining, science or supercomputing, etc. In some examples, data center 10A is a separate network server, network peer or other institution.

[0025] exist Figure 1 In the example of FIG. 1 , data center 10A includes a group of storage systems and application servers, including servers 12A to 12X (collectively referred to as “servers 12”) interconnected via a high-speed switching fabric 20 provided by one or more layers of physical network switches and routers. Servers 12 serve as physical computing nodes in the data center. For example, each server 12 may provide a processor for executing one or more application workloads 37 (in Figure 1 As described herein, the terms "application workload 37" or "workload 37" may be used interchangeably to refer to an application workload 37. A workload 37 may be executed on a virtualized environment, such as a virtual machine, a container, or some type of virtualized instance, or in some cases, on a bare metal server that directly executes the workload rather than indirectly executing the workload in a virtualized environment. Each server 12 may alternatively be referred to as a host computing device, or more simply a host. A server 12 may execute one or more workloads 37 on one or more virtualized instances, such as a virtual machine, a container, or other virtual execution environment for running one or more services, such as a virtualized network function (VNF).

[0026] The switching fabric 20 may include top-of-rack (TOR) switches 16A-16N coupled to a distribution layer of rack switches 18A-18M, and the data center 10A may include one or more non-edge switches, routers, hubs, gateways, security devices (such as firewalls, intrusion detection and / or intrusion prevention devices), servers, computer terminals, laptops, printers, databases, wireless mobile devices (such as cell phones or personal digital assistants), wireless access points, bridges, cable modems, application accelerators, or other network devices. The data center 10A includes servers 12A-12X interconnected via a high-speed switching fabric 20 provided by one or more layers of physical network switches and routers. The switching fabric 20 is provided by a group of interconnected top-of-rack (TOR) switches 16A-16N (collectively referred to as "TOR switches 16") coupled to a distribution layer of rack switches 18A-18M (collectively referred to as "rack switches 18"). Although not shown, data center 10A may also include, for example, one or more non-edge switches, routers, hubs, gateways (security devices such as firewalls, intrusion detection and / or intrusion prevention devices), servers, computer terminals, laptops, printers, databases, wireless mobile devices (such as cellular phones or personal digital assistants), wireless access points, bridges, cable modems, application accelerators, or other network devices.

[0027] In this example, the TOR switch 16 and the rack switch 18 provide redundant (multi-homed) connections to the gateway 108 and the service provider network 7 to the server 12. The rack switch 18 aggregates traffic flows and provides high-speed connections between the TOR switches 16. The TOR switch 16 can be a network device that provides Layer 2 (MAC) and / or Layer 3 (e.g., IP) routing and / or switching functions. The TOR switch 16 and the rack switch 18 can each include one or more processors and memories and be capable of executing one or more software programs. The rack switch 18 is coupled to the gateway 108, which can perform Layer 3 routing to route network traffic between the data center 10A and the customer 11 through the service provider network 7.

[0028] The switching fabric 20 can perform layer 3 routing to route network traffic between the data center 10A and the customer 11 through the service provider network 7. The gateway 108 is used to forward and receive packets between the switching fabric 20 and the service provider network 7. The data center 10A includes an overlay network that extends the switching fabric 20 from physical switches 18, 16 to software or "virtual" switches. For example, virtual routers 30A-30X located in servers 12A-12X, respectively, can extend the switching fabric 20 by communicatively coupling with one or more physical switches located within the switching fabric 20. The virtual switch can dynamically create and manage one or more virtual networks that can be used for communication between application instances. In one example, the virtual routers 30A-30X execute the virtual network as an overlay network, which provides the ability to decouple the virtual address of the application from the physical address (e.g., IP address) of one of the servers 12A-12X on which the application is executing. Each virtual network can use its own addressing and security scheme and can be considered orthogonal to the physical network and its addressing scheme. Various techniques can be used to transmit packets within and across virtual networks on a physical network.

[0029] Software defined network ("SDN") controller 132 provides a logically and in some cases physically centralized controller for facilitating the operation of one or more virtual networks within data center 10A according to one or more examples of the present disclosure. Throughout the present disclosure, the terms SDN controller and virtual network controller ("VNC") may be used interchangeably. In some examples, SDN controller 132 operates in response to configuration input received from orchestration engine 130 via northbound API 131, which in turn operates in response to configuration input received from administrator 24 operating user interface device 129. Additional information regarding the SDN controller 132 operating in conjunction with other devices of the data center 10A or other software defined networks can be found in International Application No. PCT / US2013 / 044378, filed on June 5, 2013, entitled “PHYSICAL PATH DETERMINATION FOR VIRTUALNETWORK PACKET FLOWS,” and in U.S. Patent Application Serial No. 15 / 476,136, filed on March 31, 2017, entitled “SESSION-BASED TRAFFIC STATISTICS LOGGING FOR VIRTUAL ROUTERS.”

[0030] For example, an SDN platform may be used in data center 10 to control and manage network behavior. In some cases, the SDN platform includes a logically centralized and physically distributed SDN controller (such as SDN controller 132), and a distributed forwarding plane in the form of a virtual router that extends the network from physical routers and switches in the data center switching fabric to a virtual overlay network hosted in a virtualized server.

[0031] In some examples, the SDN controller 132 manages the network and networking services, such as load balancing, security, and allocates resources from the servers 12 to various applications via the southbound API 133. That is, the southbound API 133 represents a set of communication protocols used by the SDN controller 132 to make the actual state of the network equal to the desired state specified by the orchestration engine 130. For example, one such communication protocol may include a message communication protocol, such as XMPP. For example, the SDN controller 132 implements high-level requests from the orchestration engine 130 by configuring the following: physical switches, such as the TOR switches 16, rack switches 18, and switch fabrics 20; physical routers; physical service nodes, such as firewalls and load balancers; and virtual services, such as virtual firewalls in a virtualized environment. The SDN controller 132 maintains routing, network, and configuration information within a state database. The SDN controller 132 transmits an appropriate subset of the routing information and configuration information from the state database to a virtual router (VR) 30A-30X or proxy 35A-35X ( Figure 1 Proxy in .

[0032] As described herein, each server 12 includes a corresponding forwarding component 39A-39X (hereinafter referred to as a "forwarding component 39") that executes the workflow executed on each server 12 ( Figure 1 "WF 37") data forwarding and business statistics collection functions. Figure 1 In the example of , each forwarding component is described as including a virtual router ( Figure 1 ” in the VR 30A-VR 30X) and a VR agent ( Figure 1 The VR agent 35 may be referred to as a policy agent. Alternatively, the policy agent may correspond to a subcomponent or function of the VR agent 35.

[0033] In this example, each virtual router 30A-30X implements at least one routing instance for a corresponding virtual network within the data center 10 and routes packets to the appropriate virtual machine, container, or other element executing within the operating environment provided by the server. Packets received by the virtual router of the server 12A, for example, packets received from the underlying physical network structure, may include an outer header to allow the physical network structure to tunnel the payload or "inner packet" to the physical network address of the network interface of the server 12A executing the virtual router. The outer header may include not only the physical network address of the network interface of the server, but also a virtual network identifier, such as a VxLAN tag or a multi-protocol label switching (MPLS) label, which identifies a virtual network executed by the virtual router and the corresponding routing instance. The inner packet includes an inner header having a destination network address that conforms to the virtual network addressing space of the virtual network identified by the virtual network identifier.

[0034] In some aspects, the virtual router buffers and aggregates multiple tunneled packets received from the underlying physical network structure before delivering to the appropriate routing instance for the packet. That is, the virtual router of the forwarding component executed on one of the servers 12 can receive inbound tunnel packets of the packet flow from the TOR switch 16, and before routing the tunnel packets to the locally executed virtual machine, process the tunnel packets to construct a single aggregated tunnel packet for forwarding to the virtual machine. That is, the virtual router can buffer multiple inbound tunnel packets and construct a single tunnel packet in which the payloads of the multiple tunnel packets are combined into a single payload and the outer / overlay header on the tunnel packet is removed and replaced with a single header virtual network identifier. In this way, the aggregated tunnel packet can be forwarded by the virtual router to the virtual machine as if a single inbound tunnel packet was received from the virtual network. In addition, in order to perform the aggregation operation, the virtual router can utilize a kernel-based offload engine that seamlessly and automatically guides the aggregation of tunnel packets. Further example techniques for a virtual router to forward traffic to customer-specific virtual machines executing on server 12 are described in U.S. patent application Ser. No. 14 / 228,844, filed on Mar. 28, 2014, and entitled “PACKET SEGMENTATION OFFLOAD FOR VIRTUAL NETWORKS.”

[0035] exist Figure 1In the example of FIG. 1 , the SDN controller 132 learns routes and other information (such as configuration) and distributes it to all computing nodes in the data center 10. Upon receiving the routing information from the SDN controller 132, the VR agent 35 of the forwarding component 39 running within the computing node typically programs the data forwarding element (virtual router 30) with the forwarding information. The SDN controller 132 sends routing and configuration information to the VR agent 35 using a message communication protocol such as the XMPP protocol semantics rather than a heavier protocol using a routing protocol such as BGP. In XMPP, the SDN controller 132 and the agent transmit routes and configurations through the same channel. When receiving routes from the VR agent 35, the SDN controller 132 acts as a message communication protocol client, and in this case the VR agent 35 acts as a message communication protocol server. Conversely, when the SDN controller sends routes to the VR agent 35, the SDN controller 132 acts as a message communication protocol server for the VR agent 35 that is a message communication protocol client. The SDN controller 132 can send security policies to the VR agent 35 for the virtual router 30 to apply.

[0036] In some examples, the system 8 performs session sampling and recording of traffic flows and traffic statistics. The virtual routers 30 of the forwarding component 39 perform routing services between WLs 37, such as forwarding network traffic between an initiator WL 37 of network traffic and a destination WL 37 for network traffic. The virtual routers 30 process the ingress and egress flows of network traffic forwarded by each virtual router 30 to generate traffic flow statistics. Traffic flow statistics may include metrics such as bandwidth consumption, traffic routing and paths, and application usage.

[0037] Each virtual router 30 transmits traffic flow statistics to a corresponding VR agent 35 of the forwarding component 39. Each VR agent 35 can process the traffic flow statistics of the ingress and egress flows to identify pairs of ingress and egress flows corresponding to common communication sessions between the workflows 37. For each identified pair, the VR agent 35 generates a session record of the traffic statistics of the network traffic forwarded by the virtual router 30. In some cases, the session record can be used for a unidirectional traffic flow. The VR agent 35 uploads the session record together with the tag information for the given flow to the traffic collector 38 of the SDN controller 32 for the cloud data center 10. The traffic collector 38 receives the session records from multiple VR agents 35 and stores these session records in an analysis database for use by the administrator 24 of the cloud data center 10. The user interface device 129 can access such metrics via an interface and generate a user interface to help the administrator 24 detect and diagnose network performance problems in the cloud data center. In addition, and discussed in more detail below, the traffic collector 38 can forward the session record to the policy controller 23 for analysis according to the technology of the present disclosure.

[0038] In some examples, the session record includes various information and metadata describing the business flow. For example, the session record may include 5-tuple information, such as the source address of the business flow, the source port of the business flow, the destination address of the business flow, the destination port of the business flow, and the protocol of the business flow. In other examples, the session record may also include other information, such as the virtual network of the business flow, the source virtual machine interface (VMI) and / or the destination VMI of the business flow, one or more tags of the source of the business flow, one or more tags of the destination of the business flow, the IP prefix of the business flow, the service port of the business flow, the service protocol of the business flow, the operation of the business flow, the virtual machine identifier of the business flow, the operating system of the application workload associated with the business flow, one or more software packages installed on the operating system for the application workload associated with the business flow, or the version of the application workload 37 associated with the business flow.

[0039] The user interface device 129 may be implemented as any suitable computing system, such as a mobile or non-mobile computing device operated by a user and / or administrator 24. The user interface device 129 may, for example, represent a workstation, a laptop or notebook computer, a desktop computer, a tablet computer, or any other computing device that may be operated by a user and / or present a user interface according to one or more aspects of the present disclosure. Further example techniques by which the user interface device 129 may present such metrics to assist the administrator 24 in detecting and diagnosing network performance issues in a cloud data center are described in U.S. patent application Ser. No. 16 / 022,464, entitled “INTER-APPLICATION WORKLOAD NETWORKTRAFFIC MONITORING AND VISUALIZATION”.

[0040] In some examples, orchestration engine 130 manages functions of data center 10A, such as computing, storage, network, and application resources. For example, orchestration engine 130 can create virtual networks for tenants within data center 10A or across data centers. Orchestration engine 130 can attach workloads (WLs) to tenants' virtual networks. Orchestration engine 130 can connect tenants' virtual networks to external networks, such as the Internet or VPNs. Orchestration engine 130 can implement security policies across the boundaries of a set of workloads or tenant networks. Orchestration engine 130 can deploy network services (e.g., load balancers) in tenants' virtual networks.

[0041] In some examples, SDN controller 132 manages the network and networking services, such as load balancing, security, and allocates resources from server 12 to various applications via southbound API 133. That is, southbound API 133 represents a set of communication protocols used by SDN controller 132 to make the actual state of the network equal to the desired state specified by orchestration engine 130. For example, SDN controller 132 implements high-level requests from orchestration engine 130 by configuring: physical switches, such as TOR switches 16, rack switches 18, and switch fabrics 20; physical routers; physical service nodes, such as firewalls and load balancers; and virtual services, such as virtual firewalls in VMs. SDN controller 132 maintains routing, networking, and configuration information within a state database.

[0042] Typically, traffic between any two network devices, such as traffic between network devices (not shown) within the switching fabric 20 or traffic between a server 12 and a client 11 or between servers 12, can traverse the physical network using many different paths. For example, there may be several different paths of equal cost between two network devices. In some cases, packets belonging to network traffic from one network device to another can be distributed among various possible paths using a routing strategy known as multipath routing at each network switching node. For example, Internet Engineering Task Force (IETF) RFC2992, "Analysis of Equal-Cost Multipath Algorithms," describes routing techniques for routing packets along multiple equal-cost paths. The technology of RFC2992 analyzes a specific multipath routing strategy that involves assigning flows to containers by hashing packet header fields that send all packets from a particular traffic flow on a single deterministic path.

[0043] For example, a "flow" may be defined by five values ​​or a "quintuple" used in a packet's header, namely, protocol, source IP address, destination IP address, source port, and destination port for routing the packet through a physical network. For example, protocol specifies a communication protocol, such as TCP or UDP, and source port and destination port refer to the source and destination ports of a connection. A set of one or more packet data units (PDUs) matching a particular flow ingress represents a flow. Flows may be broadly classified using any of the following parameters of the PDUs: such as source and destination data link (e.g., MAC) and network (e.g., IP) addresses, virtual local area network (VLAN) tags, transport layer information, Multiprotocol Label Switching (MPLS) or Generic MPLS (GMPLS) labels, and the ingress port of the network device receiving the flow. For example, a flow may be all PDUs sent in a Transmission Control Protocol (TCP) connection, all PDUs originating from a particular MAC address or IP address, all PDUs with the same VLAN tag, or all PDUs received at the same switch port. A flow may additionally or alternatively be defined by an application identifier (AppID) determined by a virtual router agent or other entity that identifies a type of service or application associated with the flow, e.g., using port and protocol lists or deep packet inspection (DPI), i.e., the flow transports application data for that type of service or application.

[0044] Virtual routers (virtual routers 30A to 30X, collectively referred to as Figure 1 The "virtual router 30" in the data center 10A executes multiple routing instances for corresponding virtual networks within the data center 10A and routes packets to appropriate workloads 37 executing within the operating environment provided by the server 12. Each server 12 may include a virtual router. Packets received by the virtual router 30A of the server 12A, for example, packets received from the underlying physical network structure may include an outer header to allow the physical network structure to tunnel the payload or "inner packet" to the physical network address of the network interface of the server 12A. The outer header may include not only the physical network address of the server's network interface, but also a virtual network identifier, such as a VxLAN tag or a multi-protocol label switching (MPLS) label, which identifies a virtual network executed by the virtual router and the corresponding routing instance. The inner packet includes an inner header having a destination network address that conforms to the virtual network addressing space of the virtual network identified by the virtual network identifier.

[0045] In some aspects, the virtual router buffers and aggregates multiple tunneled packets received from the underlying physical network structure before delivering to the appropriate routing instance for the packet. That is, a virtual router executed on one of the servers 12 can receive inbound tunnel packets of a packet flow from one or more TOR switches 16 within the switching structure 20, and before routing the tunnel packets to the locally executed virtual machines, process the tunnel packets to construct a single aggregated tunnel packet for forwarding to the virtual machines. That is, the virtual router can buffer multiple inbound tunnel packets and construct a single tunnel packet in which the payloads of the multiple tunnel packets are combined into a single payload and the outer / overlay header on the tunnel packet is removed and replaced with a single header virtual network identifier. In this way, the aggregated tunnel packet can be forwarded by the virtual router to the virtual machine as if a single inbound tunnel packet was received from the virtual network. In addition, to perform the aggregation operation, the virtual router can utilize a kernel-based offload engine that seamlessly and automatically guides the aggregation of tunnel packets. Further example techniques for a virtual router to forward traffic to customer specific virtual machines executing on server 12 are described in US patent application Ser. No. 14 / 228,844, entitled “PACKET SEGMENTATION OFFLOAD FOR VIRTUAL NETWORKS.”

[0046] In some example embodiments, a virtual router 30 executed on a server 12 manipulates inbound tunnel packets received between multiple processor cores to facilitate packet processing load balancing between the cores when processing packets for routing to one or more virtual and / or physical machines. As an example, a server 12A includes multiple network interface cards and multiple processor cores to execute a virtual router 30A, and manipulates received packets between multiple processor cores to facilitate packet processing load balancing between the cores. For example, a particular network interface card of the server 12A may be associated with a designated processor core to which the network interface card indicates all received packets. Instead of processing each received packet, various processor cores offload flows to one or more other processor cores for processing using available work cycles of the other processor cores, based on a hash function applied to at least one of the internal and external packet headers.

[0047] exist Figure 1In the example, the data center 10A also includes a policy controller 23, which provides monitoring, scheduling, and performance management for the data center 10A. The policy controller 23 interacts with policy agents 35A-35X (collectively referred to as "policy agents 35" or "agents 35"), which are deployed in at least some corresponding physical servers 12 to monitor resource usage of physical computing nodes and virtualization hosts (such as WL 37) executed on physical hosts. In this way, the policy agent 35 provides a distributed mechanism for collecting various usage metrics and local implementation of policies installed by the policy controller 23. In an example embodiment, the policy agent 35 runs on the lowest level "compute node" of the infrastructure of the data center 10A, which provides computable resources for executing application workloads. For example, a compute node can be a bare metal host of a server 12, a workload 37, a container, etc.

[0048] like Figure 1 As shown in the example of , the policy controller 23 can define a rule base as a set of policies 202 and maintain it based on multiple session records of business statistics as described below. The policy controller 23 can manage the control of each server 12 based on the policy set 202 maintained by the policy controller 23. The policy 202 can be created or derived in response to the input of the administrator 24 and / or in response to the operations performed by the policy controller 23 as described above. In some examples, the policy controller 23 can additionally, for example, observe the operation of the data center 10A over time and apply machine learning techniques to generate one or more policies 202. As further observations about the data center 10A are made, the policy controller 23 can improve the policy 202 periodically, occasionally, or continuously.

[0049] The policy controller 23 (e.g., an analysis engine within the policy controller 23) can determine how to deploy, implement, and / or trigger policies at one or more servers 12. For example, the policy controller 23 can be configured to push one or more policies 202 to one or more policy agents 35 executing on the server 12. The policy controller 23 can receive information about internal processor metrics from one or more policy agents 35 and determine whether the conditions of the rules for one or more metrics are met. The policy controller 23 can analyze the internal processor metrics received from the policy agent 35, and based on the analysis, instruct or cause one or more policy agents 35 to perform one or more operations to modify the operation of the server associated with the policy agent. For example, the policy controller 23 can push one or more policies, including the configuration of the communication protocol to be used, the interval for publishing communication probes, and / or the metrics monitored for use with agents that publish communication probes between agents. The information collected from the policy agent associated with the publication and monitoring of these communication probes can be used to generate a dashboard including a graphical user interface that visually depicts one or more states associated with the metrics associated with the collected information and / or derived from the collected information and data associated with the publication of the communication probes. In another example, policy controller 23 may be configured to aggregate information about devices and arrangements of devices included in switch fabric 20 and generate a set of graphical user interfaces that visually depict the devices and the interconnections between the devices based on a set of predefined and / or user-selectable rules.

[0050] In various examples, the policy controller 23 distributes policy rules including tags for objects at the project level. However, the policy controller 23 may additionally or alternatively distribute policy rules including tags specifying various different object levels, such as the global environment level, the project level, the virtual network level, the virtual machine level, or the interface level.

[0051] The policy controller 23 may be implemented as any suitable computing system, such as a mobile or non-mobile computing device operated by a user and / or by an administrator 24. The policy controller 23 may, for example, represent a workstation, a laptop or notebook computer, a desktop computer, a tablet computer, or any other computing device that may be operated by a user and / or present a user interface according to one or more aspects of the present disclosure. In some examples, the policy controller 23 is executed on one of the servers 12. In another example, the policy controller 23 is distributed on one or more computing devices. In some examples, the policy controller is a dedicated controller executed on dedicated hardware.

[0052] According to the technology of the present disclosure, the policy controller 23 can generate an intent-based firewall policy for the network 8 based on the analysis of the traffic flows of the network 8. As described above, the policy controller 23 can receive multiple session records of traffic statistics of multiple traffic flows served by multiple VR agents 35 from the traffic collector 38. The policy controller 23 associates the multiple session records of traffic statistics of the multiple traffic flows with the multiple session records of traffic statistics of the multiple workloads 37.

[0053] In one example, in order to associate the session records of the business flows into the session records of the workload 37, the policy controller 23 presents multiple session records of business statistics of multiple business flows to the user or administrator via the UI device 129. In addition, the policy controller 23 receives input via the UI device 129 specifying which business flows among the multiple business flows correspond to corresponding application workloads 37 among the multiple application workloads 37, and based on the input, aggregates the session records of the business flows into the session records of the workload 37.

[0054] In another example, policy controller 23 associates the session record of the business flow to the session record of workload 37 based on one or more tags specified by the session record of the business flow. For example, the session record of the business flow may specify one or more tags corresponding to the tags of the tagged workload 37. Policy controller 23 may use such tags to match workload 37 to a specific session record of the business flow, and aggregate the session record of the business flow into the session record of workload 37 based on the matching tags.

[0055] In another example, the policy controller 23 associates the session record of the business flow into the session record of the workload 37 described herein. In addition, the policy controller 23 may generate one or more tags corresponding to the tags of the tagged workload 37 from the session record of the business flow, and apply the generated tags to the session record. For example, the policy controller 23 may use metadata in the session record to generate one or more tags corresponding to the tags of the tagged workload 37, and apply the generated tags to the session record. The policy controller 23 may aggregate the session record of the business flow into the session record of the workload 37 based on the tags. In addition, as described in more detail below, the policy controller 23 may use the generated tags to generate one or more application policies for business flows applied to instances of the workload 37.

[0056] In another example, the policy controller 23 associates the session records of the service flow to the session records of the workload 37 by applying a clustering algorithm to the session records of the service flow. For example, the policy controller 23 may use a clustering algorithm to identify the relationship between the session records of the service flow and a specific workload 37, and based on these relationships, aggregate the session records of the service flow into the session records of the workload 37. In some examples, the clustering algorithm is one of a K-means clustering algorithm, a mean shift clustering algorithm, a density-based spatial clustering of noise applications (DBSCAN) clustering algorithm, an expectation maximization (EM) clustering algorithm that may use a Gaussian mixture model (GMM), and an agglomerative hierarchical clustering algorithm.

[0057] The policy controller 23 generates an application firewall policy 202 for multiple workloads 37 based on multiple session records of traffic statistics of the multiple workloads 37. Each of the multiple application firewall policies includes one or more application firewall policy rules, and each of the one or more application firewall policy rules defines whether one of the VR agents 35 should allow or deny the traffic flow of the corresponding workload 37. For example, the policy controller 23 can use a 5-tuple of session records for the workload 37 to define a 5-tuple of application firewall policy rules. In addition, the policy controller 23 can define rules for application firewall policy rules (e.g., allowing, blocking, logging, or reporting traffic corresponding to the rule) based on the session records. In some examples, each application firewall policy rule specifies a default rule that allows the corresponding network traffic and blocks all other network traffic that does not match the generated application firewall policy rule.

[0058] In an example implementation of the technology disclosed herein, the policy controller 23 uses multiple configuration objects to implement one or more policies. As an example, the policy controller 23 applies a first group of configuration objects at a global level. The first group of configuration objects includes a global application policy set, a global firewall policy, a global firewall rule, and a global tag across multiple levels and / or categories. The policy controller 23 distributes the first group of configuration objects to virtual routers, for example, at a global level. The policy controller 23 matches the global tag associated with the global application policy set, the global firewall policy, and the global firewall rule with the object marked with the global tag. Based on the global application policy set, the global firewall policy, and the global firewall rule, the policy controller 23 allows or blocks network services between interfaces of objects marked with the global tag. For example, the interface can be a virtual machine interface (VMI).

[0059] In addition, the policy controller 23 applies a second set of configuration objects at the project level. The second set of policy rules may include project-specific application policy sets, firewall policies, firewall rules, and tags across multiple levels. The policy controller 23 distributes the second set of configuration objects at the project level. The policy controller 23 matches the project tags associated with the project-specific application policy sets, firewall policies, and firewall rules with objects marked with the project tags. Based on the project-specific application policy sets, firewall policies, and firewall rules, the policy controller 23 allows or blocks network services between interfaces of objects marked with the project tags.

[0060] In a further example, the policy controller 23 can specify lower-level configuration objects, such as application policy sets, firewall policies, firewall rules, and tags defined at a virtual network-specific level, a virtual machine-specific level, and / or an interface-specific level. By doing so, the policy controller 23 can apply a hierarchical policy set to multiple objects within one or more data centers 10. Thus, the techniques of the present disclosure allow for the distribution of simplified business policies that are scalable and robust in many different types of deployment and execution environments. Additional descriptions can be found in U.S. patent application Ser. No. 15 / 819,522, filed on Nov. 22, 2017, entitled “Scalable Policy Management for Virtual Networks.”

[0061] The virtual router 30 may receive and apply application policies associated with executing workloads 37 on the corresponding servers 12, which generates traffic statistics for flows of the streaming sessions that are partially managed by the application policies. Such application policies may specify different actions on flows. For example, the application policies may specify that traffic should be allowed, blocked, logged, reported, etc. As described in further detail below, the virtual proxy 35 uploads the traffic statistics for the streaming sessions to the collector 38 or directly to the policy controller 23.

[0062] In some examples, the policy controller 23 presents the one or more application firewall policies to a user or administrator for review via the UI device 129. In response to receiving input from the user or administrator approving the one or more application firewall policy rules, the policy controller 23 distributes the one or more application firewall policy rules to the VR agent 35 for application to the network traffic.

[0063] In one example of the technology of the present disclosure, a customer may develop one or more applications in a development environment for testing before deploying in a production environment. The virtual router 30 processes the ingress and egress flows of network services generated by one or more applications and forwarded by each virtual router 30 to generate traffic flow statistics. Each virtual router 30 transmits traffic flow statistics information to a corresponding VR agent 35 of a forwarding component 39. Each VR agent 35 processes the traffic flow statistics of the ingress and egress flows to identify pairs of ingress and egress flows corresponding to a common communication session between workflows 37. For each identified pair, the VR agent 35 generates a session record of traffic statistics of the network services forwarded by the virtual router 30. The VR agent 35 uploads the session record together with the tag information of the given flow to the traffic collector 38 of the SDN controller 32 for the cloud data center 10. The traffic collector 38 receives the session records from multiple VR agents 35 and forwards the session records to the policy controller 23 for analysis according to the technology of the present disclosure.

[0064] The policy controller 23 receives session records of traffic statistics of traffic flows served by the VR agent 35 from the traffic collector 38. The policy controller 23 associates the multiple session records of traffic statistics of the multiple traffic flows with the multiple session records of traffic statistics of the multiple workloads 37 operating within the development environment. The policy controller 23 generates application firewall policies 202 for the multiple workloads 37 based on the multiple session records of traffic statistics of the multiple workloads 37. Each of the multiple application firewall policies includes one or more application firewall policy rules, and each of the one or more application firewall policy rules defines whether one of the VR agents 35 should allow or deny the traffic flow of the corresponding workload 37. The policy controller 23 distributes the application firewall policy to the VR agent 35 to apply to the traffic flow. Therefore, when deployed in the production environment, the VR agent 35 can apply the application firewall policy to the traffic flow within the production environment. Therefore, the technology of the present disclosure can allow the policy controller 23 to automatically generate intent-based application firewall policies for the workload 37. In other words, the policy controller 23 can use the session records of the business statistics of the workload 37 to identify the "expected" business relationship between the workloads 37 in the development environment, and use the "expected" relationship to generate an application firewall policy for the workload 37 for use in the production environment to allow such "expected" business and block other forms of business. Therefore, the technology of the present disclosure can allow the creation of firewall policies that are more comprehensive, accurate, flexible, and effectively generated than traditional methods.

[0065] Figure 2 To show in more detail Figure 1 A block diagram of an example implementation of a data center 10A is shown. Figure 2In an example of a data center 10A, an overlay network is included that extends the switching fabric 20 from the physical switches 16, 18 to software or "virtual" routers 30A-30X (again collectively referred to as "virtual routers 30"). The virtual routers 30 dynamically create and manage one or more virtual networks 34 that can be used for communication between application instances. In one example, the virtual router 30 implements the virtual network as an overlay network, which provides the ability to decouple the virtual address of an application from the physical address (e.g., IP address) of one of the servers 12A-12X ("servers 12") on which the application is executing. Each virtual network can use its own addressing and security scheme and can be considered orthogonal to the physical network and its addressing scheme. Various techniques can be used to transmit packets within and across virtual networks 34 on a physical network. In some examples, the techniques described in the present disclosure provide multicast services within a virtual network 34 without requiring multicast support in the underlying physical network.

[0066] Each virtual router 30 may be executed within a hypervisor, host operating system, or other component of each server 12. Each server 12 may represent an x86 or other general or specialized server capable of executing workload 37. Figure 2 In the example of FIG. 1 , the virtual router 30A is executed within a hypervisor 31, which is also commonly referred to as a virtual machine manager (VMM), which provides a virtualization platform that allows multiple operating systems to run simultaneously on one of the servers 12. Figure 2 In the example of , virtual router 30A manages virtual networks 34, each of which provides a network environment for executing one or more workloads (WL) 37 on a virtualization platform provided by hypervisor 31. Each WL 37 is associated with one of the virtual networks VN0-VN1, and can represent a tenant WL that runs a client application such as a Web server, a database server, an enterprise application, or a virtualized service that is hosted to create a service chain. In some cases, any one or more of server 12 and another computing device can directly host the client application, i.e., not as a virtual machine. In some cases, some of WL 37 can represent containers, which are another form of virtualized execution environment. That is, both virtual machines and containers are examples of virtualized execution environments for executing workloads.

[0067] In general, each WL 37 may be any type of software application and may be executed on a virtualized environment (e.g., a virtual machine or container) that is assigned a virtual address for use within a corresponding virtual network 34, where each virtual network may be a different virtual subnet provided by virtual router 30A. The virtualized environment may be assigned its own virtual layer 3 (L3) IP address, for example, for sending and receiving communications, but may not be aware of the IP address of the physical server 12A on which the virtualized environment is executing. In this manner, a "virtual address" is a different physical address from that of the underlying physical computer system (e.g., Figure 1 or Figure 2 The address of the logical address of server 12A in the example.

[0068] In one embodiment, each server 12 includes a respective one of the virtual network (VN) agents 35A-35X (collectively, "VN agents 35") that controls the overlay of the virtual network 34 and coordinates the routing of packets within the server 12. Typically, each VN agent 35 communicates with an SDN controller 132, which generates commands to control the routing of packets through the data center 10A. The VN agent 35 may operate as a proxy for control plane messages between the workload 37 and the SDN controller 132. For example, a WL 37 may request that a message be sent via a VN agent 35A using its virtual address, and the VN agent 35A may in turn send a message and request a response to a message originating from the first message received for the virtual address of the WL 37. In some cases, the WL 37 may invoke a program or function call presented by an application programming interface of the VN agent 35A, and the VN agent 35A may also handle the encapsulation of the message, including addressing.

[0069] In one example, a network packet (e.g., a Layer 3 (L3) IP packet or a Layer 2 (L2) Ethernet packet) generated or consumed by an instance of an application executed by a workload 37 within a virtual network domain can be encapsulated in another packet transmitted by a physical network (e.g., another IP or Ethernet packet). The packet transmitted in the virtual network may be referred to as an "internal packet" in this article, while the physical network packet may be referred to as an "external packet" or "tunnel packet" in this article. The encapsulation and / or decapsulation of the virtual network packet within the physical network packet may be performed within the virtual router 30, for example, within a hypervisor or host operating system running on each server 12. As another example, the encapsulation and decapsulation functions may be performed at the first-hop TOR switch 16 at the edge of the switching structure 14, the first-hop TOR switch 16 being a hop removed from the application instance that initiated the packet. This function is referred to as tunnel transmission in this article, and can be used within the data center 10A to create one or more overlay networks. In addition to IPinIP, other example tunnel protocols that can be used include IP over GRE, VxLAN, MPLS over GRE, MPLS over UDP, etc.

[0070] As described above, the SDN controller 132 provides a logically centralized controller to facilitate the operation of one or more virtual networks within the data center 10A. For example, the SDN controller 132 can maintain a routing information base, such as one or more routing tables that store routing information for physical networks and one or more overlay networks of the data center 10A. Similarly, the switches 16, 18 and the virtual router 30 maintain routing information, such as one or more routing and / or forwarding tables. In an example embodiment, the virtual router 30A of the hypervisor 31 implements a network forwarding table (NFT) 32 for each virtual network 34. Typically, each NFT 32 stores forwarding information for the corresponding virtual network 34 and identifies where the data packet is to be forwarded and whether to encapsulate the packet in a tunneling protocol, such as using a tunnel header that may include one or more headers for different layers of the virtual network protocol stack.

[0071] In some examples, the SDN controller 132 (or "virtual network controller 132") also includes a policy controller 23. In some examples, the SDN controller 132 and the policy controller 23 are executed on the same computing device (e.g., one of the servers 12). In some examples, the SDN controller 132 and the policy controller 23 are executed on different computing devices (e.g., different servers in the server 12). The policy controller 23 is configured to mark multiple objects ("policy objects") across multiple levels. In some examples, the multiple levels include levels of objects, such as global environment levels, project levels, virtual network levels, virtual machine levels, or interface levels of objects. The policy object can be associated with one or more workloads. In a further example, the policy controller 23 marks multiple objects across multiple categories, which can include a top category and one or more subcategories. In some examples, the multiple categories include applications, deployments, application layers, geographic sites, virtual networks, WL 37, interfaces, projects, security requirements, quality requirements, users, or compliance requirements executed within WL 37.

[0072] The virtual router 30A of the forwarding component 39A performs session monitoring and records traffic flows and traffic statistics. In one example, the virtual router 30A of the forwarding component 39A processes the forwarding traffic flows forwarded by the virtual router 30A and the reverse traffic flows forwarded by the virtual router 30A to generate traffic flow statistics for the traffic flows. In some examples, the traffic flow statistics for each flow include a five-tuple (e.g., protocol, source address, source port, destination address, and destination port), a count of packets forwarded for the flow, and / or a count of bytes forwarded for the flow. The virtual router 30A periodically transmits the generated traffic flow statistics information to the VR agent 35A so that the VR agent maintains similar records of traffic flow statistics for the traffic flowing through the virtual router 30A.

[0073] The VR agent 35A processes the generated traffic flow statistics to identify pairs of ingress traffic flows and egress traffic flows corresponding to a common communication session between one or more workloads 37. In one example, the VR agent 35A identifies pairs of ingress traffic flows and egress traffic flows corresponding to a common communication session. In one example, the VR agent 35A processes the traffic statistics generated by the virtual router 30A to compare the protocol, source address, source port, destination address, and destination port (e.g., a five-tuple) of each sampled ingress packet flow and each egress packet flow. For example, if the first ingress flow has the same protocol as the first egress flow, if the source address and source port of the first ingress flow are the same as the destination address and destination port of the first egress flow, and if the destination address and destination port of the first ingress flow are the same as the source address and source port of the first egress flow, then the VR agent 35A determines that the first ingress flow and the first egress flow correspond to a common communication session between two devices. However, if the second ingress flow does not have the same protocol as the second egress flow, or if the source address and source port of the second ingress flow are not the same as the destination address and destination port of the second egress flow, or if the destination address and destination port of the second ingress flow are not the same as the source address and source port of the second egress flow, then the VR agent 35A determines that the second ingress flow and the second egress flow do not correspond to a common communication session between the two devices.

[0074] For each identified pair of ingress and egress traffic flows, VR agent 35A generates a session record of traffic statistics for network traffic forwarded by virtual router 30 A. In other words, VR agent 35A generates a session record of traffic statistics for network traffic forwarded by virtual router 30 A, wherein each record includes traffic statistics related to a common communication session between two devices and includes traffic statistics for both the ingress traffic flow and the egress traffic flow of the common communication session.

[0075] In some examples, VR agent 35A generates such a session record of traffic statistics for each traffic flow received by virtual router 30A. In other examples, VR agent 35A generates a session record of traffic statistics only for a subset of traffic flows received by virtual router 30A to generate a session record of traffic statistics for a subset of traffic flowing through virtual router 30A. In some examples, the subset of traffic flows is randomly selected. In still other examples, the subset of traffic flows is selected according to an algorithm to provide a statistically accurate representation of the traffic flows received by virtual router 30A. In some examples, VR agent 35A receives an indication of one or more traffic flows including a subset of session records for generating traffic statistics from controller 132.

[0076] In one example, a single session record 52 may include a plurality of "difference bytes" for both the ingress and egress flows associated with the session. Such "difference bytes" describe the number of bytes received for the ingress flow and the number of bytes received for the egress flow during the previous monitoring period (e.g., the difference in bytes between the end of the previous monitoring period and the end of the current monitoring period). In addition, the session record 52 may include additional traffic flow and traffic statistics, such as the total number of bytes forwarded for both the ingress and egress flows, the available time for both the ingress and egress flows, addresses, ports, and forwarding information associated with each of the ingress and egress flows, and the like.

[0077] The VR agent 35A uploads the session records 52 to the traffic collector 38, which in this example is a component of the SDN controller 132. In some examples, the VR agent 35A periodically uploads a set of session records 52 generated during a previous sampling period (e.g., the VR agent uploads the set of session records 52 to the traffic collector 38 approximately once per second, once per minute, or once per day). The traffic collector 38 writes the session-related traffic flows and traffic statistics of the session records 52 into the analysis database 35.

[0078] According to the technology of the present disclosure, the policy controller 23 can generate an intent-based firewall policy 202 for the data center 10A based on the analysis of the business flows of the data center 10A. As described above, the policy controller 23 can receive multiple session records of business statistics of multiple business flows served by multiple VR agents 35 from the business collector 38. The policy controller 23 associates the multiple session records of business statistics of the multiple business flows with the multiple session records of business statistics of the multiple workloads 37.

[0079] In one example, in order to associate the session record of the service flow with the session record of the workload 37, the policy controller 23 Figure 1 The UI device 129 presents multiple session records of business statistics of multiple business flows to the user or administrator. In addition, the policy controller 23 receives input via the UI device 129 specifying which business flows among the multiple business flows correspond to corresponding application workloads 37 among the multiple application workloads 37, and based on the input, aggregates the session records of the business flows into the session records of the workload 37.

[0080] In another example, policy controller 23 associates the session record of the business flow to the session record of workload 37 based on one or more tags specified by the session record of the business flow. For example, the session record of the business flow may specify one or more tags corresponding to the tags that mark workload 37. Policy controller 23 may use such tags to match workload 37 to a specific session record of the business flow, and aggregate the session record of the business flow into the session record of workload 37 based on the matching tags.

[0081] In an example implementation of the technology disclosed herein, the policy controller 23 uses multiple configuration objects to aggregate session records of business flows into session records of workload 37. As an example, the policy controller 23 applies a first set of configuration objects at a global level. The first set of configuration objects includes a global application policy set, a global firewall policy, a global firewall rule, and a global tag across multiple levels and / or categories. The policy controller 23 distributes the first set of configuration objects to the virtual router 30, for example, at a global level. The policy controller 23 matches the global tag of the session record of the business flow associated with the global application policy set, the global firewall policy, and the global firewall rule with an object marked with the global tag. For example, the policy controller 23 uses this global tag of the session record of the business flow to match the workload 37 to a specific session record of the business flow, and based on the matching tag, aggregates the session record of the business flow into the session record of the workload 37.

[0082] In addition, the policy controller 23 applies a second set of configuration objects at the project level. The second set of policy rules includes project-specific application policy sets, firewall policies, firewall rules, and tags across multiple levels. The policy controller 23 distributes the second set of configuration objects at the project level. The policy controller 23 matches the project tags of the session records of the business flows associated with the project application policy sets, project firewall policies, and project firewall rules with objects marked with the project tags. For example, the policy controller 23 uses such project tags of the session records of the business flows to match the workload 37 to the specific session records of the business flows, and based on the matching tags, aggregates the session records of the business flows into the session records of the workload 37.

[0083] In a further example, the policy controller 23 may specify lower-level configuration objects, such as application policy sets, firewall policies, firewall rules, and tags defined at a virtual network-specific level, a virtual machine-specific level, and / or an interface-specific level. By doing so, the policy controller 23 may apply a set of hierarchical tags to multiple objects within one or more data centers 10, obtain session records for business flows including the tags, and aggregate such session records for business flows into session records for workloads 37 based on matching tags. Additional description of the distribution of simplified business policies can be found in U.S. patent application Ser. No. 15 / 819,522, filed Nov. 22, 2017, entitled “Scalable Policy Management for Virtual Networks.”

[0084] In another example, the policy controller 23 associates the session records of the service flow to the session records of the workload 37 by applying a clustering algorithm to the session records of the service flow. For example, the policy controller 23 may use a clustering algorithm to identify the relationship between the session records of the service flow and a particular workload 37, and based on these relationships, aggregate the session records of the service flow into the session records of the workload 37.

[0085] The policy controller 23 generates an application firewall policy 202 for the multiple workloads 37 based on the multiple session records of the traffic statistics of the multiple workloads 37. Each of the multiple application firewall policies includes one or more application firewall policy rules, and each of the one or more application firewall policy rules defines whether one of the VR agents 35 should allow or deny the traffic flow of the corresponding workload 37. For example, the policy controller 23 can use a 5-tuple of session records for the workload 37 to define a 5-tuple of application firewall policy rules. In addition, the policy controller 23 can define rules of the application firewall policy rules based on the session records (e.g., allow, block, log, or report the traffic corresponding to the rules). In some examples, each application firewall policy rule specifies a default rule that allows the corresponding network traffic and blocks all other network traffic that does not match the generated application firewall policy rule.

[0086] The virtual router 30 may receive and apply application policies associated with executing workloads 37 on the corresponding servers 12, which generates traffic statistics for flows of the streaming sessions that are partially managed by the application policies. Such application policies may specify different actions regarding flows. For example, the application policies may specify that traffic should be allowed, blocked, logged, reported, etc. As described in further detail below, the virtual proxy 35 uploads the traffic statistics for the streaming sessions to the collector 38. Figure 2In the example of FIG. 1 , the collector 38 is a subcomponent of the SDN controller 132. However, in other examples, the collector 38 may be a component separate from the SDN controller 132. In such an example where the collector 38 is separate from the SDN controller 132, the virtual agent 35 may upload the traffic statistics of the flow session to the collector 38 or directly to the policy controller 23.

[0087] In some examples, the policy controller 23 presents the one or more application firewall policies to a user or administrator for review via the UI device 129. In response to receiving input from the user or administrator approving the one or more application firewall policy rules, the policy controller 23 distributes the one or more application firewall policy rules to the VR agent 35 for application to the network traffic.

[0088] Therefore, the technology of the present disclosure can allow the policy controller 23 to automatically generate an intent-based application firewall policy for the workload 37. In other words, the policy controller 23 can use the session records of the business statistics of the workload 37 to identify the "expected" business relationship between the workloads 37, and use the "expected" relationship to generate an application firewall policy for the workload 37 for subsequent use to allow such "expected" business and block other forms of business. Therefore, the technology of the present disclosure can allow the creation of firewall policies that are more comprehensive, accurate, flexible, and effectively generated than traditional methods.

[0089] Figure 3 1 is a block diagram illustrating an example of multiple data centers 10 according to the techniques described herein. Each data center 10A-10X (collectively referred to as a "data center 10") may be configured to communicate with Figure 1 and Figure 2 The data centers 10 of FIG. 1 and FIG. 2 operate in a substantially similar manner. In some examples, one or more data centers 10 implement a cloud environment that executes Openstack, while in another example, one or more data centers 10 are a cloud environment that executes Kubernetes, and in yet another example, one or more data centers 10 implement a cloud environment that executes Amazon Web Services, and in another example, one or more data centers 10 execute on a "bare metal" infrastructure. In some examples, each data center 10 executes an infrastructure different from each other data center 10. Examples of data centers 10 may include a multi-cloud system having one or more clouds coupled to the data center 10, as shown in cloud 320. Cloud 320 may include one or more data centers, and in various examples, the data center includes a data center having the same or similar components and configured to perform some or all of the features and functions described herein with respect to the data center 10.

[0090] exist Figure 3In the example of FIG. 1 , each data center 10 includes a database 308 and three different environments: a development environment 310, a staging environment 312, and a production environment 314. Each environment of each data center 10 is capable of executing one or more applications 302A-302D (collectively referred to as "applications 302"), such as application 302A executed in development environment 310A of data center 10A, application 302B executed in production environment 314A of data center 10A, application 302C executed in development environment 310B of data center 10B, and application 302D executed in staging environment 312B of data center 10B. In one example, each application 302 is an instance of a financial application. The strategies described herein are robust and scalable so as to be applied in many different types of cloud environments as described above.

[0091] Within the environment, each application 302 may include multiple workloads. Figure 3 In the example of the present invention, application 302A of development environment 310A executes example workload web process 304A and application program interface (API) process 306A. In addition, application 302B of production environment 314A executes example workload web process 304B and API process 306B. Application 302C of development environment 310B executes example workload web process 304C and API process 306C. In addition, application 302D of staging environment 312B executes example workload web process 304D and API process 306D. However, in other examples, each application 302 of the corresponding environment 310A, 312A, 314A, 310B, 312B or 314B may execute additional processes and / or workloads or processes and / or workloads different from each other application 302. Furthermore, in some examples, the applications executing within each respective environment 310A, 310B, 312A, 312B, 314A, and 314B may be similar to or different from the applications executing within each other environment 310A, 310B, 312A, 312B, 314A, and 314B.

[0092] It may be desirable to ensure that the processes 304, 306 of each application 302 of the respective environment 310, 312, 314 do not exchange network traffic with the processes 304, 306 of the application 302 within the different environment 310, 312, 314, while still allowing network traffic to flow between different processes 304, 306 within the same application 302. For example, an administrator may wish to allow network traffic between a web process 304A and an API process 306A because each of the web process 304A and the API process 306 operate within the same application 302A within the development environment 310A. In addition, the administrator may wish to prevent network traffic between a web process 304A (e.g., operating within a development environment 310A of the data center 10A) and an API process 306B (e.g., operating within a production environment 314A of the data center 10A) or between a web process 304A (e.g., operating within a development environment 310A of the data center 10A) and an API process 306C (e.g., operating within a development environment 310B of the data center 10B). In addition, the administrator may wish to allow processes 304, 306 executing within an application 302 within the same data center 10 to access the same database 308, regardless of the particular environment 310, 312, or 314 in which the particular application 302 executes. For example, the administrator may wish to allow an API process 306A executing within an application 302A and an API process 306B executing within an application 302B to each access a database 308A.

[0093] If the administrator were to manage network service policies independently for each process 304, 306 within each application 302 within each environment 312, 312, and 314 within each data center 10, it may be difficult for the administrator to ensure that each network service policy meets security, reliability, and quality requirements. In addition, upgrading one or more policies across multiple data centers 10 may become cumbersome or difficult to manage.

[0094] In one example of the technology of the present disclosure, a customer may develop one or more applications 302 in a development environment for testing before deploying in a production environment. The policy controller 23 of the SDN controller 132 applies one or more tags specifying one or more categories to each process 304, 306. Figure 3As shown in the example of , the policy controller 23 of the SDN controller 132 applies a tag “1A” to the web process 304A, which specifies that the web process 304A belongs to the development environment 310A. Similarly, the policy controller 23 applies a tag “2A” to the API process 306A, which specifies that the web process 304A belongs to the development environment 310A. Conversely, the policy controller 23 applies a tag “1B” to the web process 304B and a tag “2B” to the API process 306B, each tag specifying that the corresponding application belongs to the production environment 314A. In addition, a tag “1C” is applied to the web process 304C, and a tag “2C” is applied to the API process 306C, each tag specifying that the corresponding application belongs to the development environment 310B; and a tag “1D” is applied to the web process 304D, and a tag “2D” is applied to the API process 306D, each tag specifying that the corresponding application belongs to the staging environment 312B.

[0095] Figure 2 The virtual routers 30 of the cloud data center 10 process the ingress and egress flows of network services generated by one or more applications and forwarded by each virtual router 30 to generate traffic flow statistics. Each virtual router 30 transmits the traffic flow statistics information to the corresponding VR agent 35 of the forwarding component 39. Each VR agent 35 processes the traffic flow statistics of the ingress and egress flows to identify the ingress and egress flow pairs corresponding to the common communication sessions between the workflows 37. For each identified pair, the VR agent 35 generates a session record of the traffic statistics of the network services forwarded by the virtual router 30. The VR agent 35 uploads the session record together with the tag information of the given flow to the traffic collector 38 of the SDN controller 32 for the cloud data center 10. Typically, the session record contains all tags associated with the source workload and the destination workload (or the source port and the destination port). The traffic collector 38 receives the session records from multiple VR agents 35 and forwards the session records to the policy controller 23 for analysis according to the technology of the present disclosure.

[0096] Policy controller 23 receives session records of traffic statistics for traffic flows serviced by VR agent 35 from traffic collector 38. Policy controller 23 associates the multiple session records of traffic statistics for multiple traffic flows with the multiple session records of traffic statistics for multiple workloads 37 operating within the development environment.

[0097] about Figure 3In the example of , the policy controller 23 receives a first session record of traffic statistics for a first traffic flow including a source tag 1A and a destination tag 2A. The policy controller 23 determines that the tag 1A is applied to the web process 304A and the tag 2A is applied to the API process 306A. Therefore, the policy controller 23 can associate the first session record of the first traffic flow with the first session record of traffic statistics corresponding to the workload of the web process 304A (e.g., as the source of the first traffic flow). In addition, the policy controller 23 can associate the first session record of the first traffic flow with the second session record of traffic statistics corresponding to the workload of the API process 306A (e.g., as the destination of the first traffic flow).

[0098] As another example, the policy controller 23 receives a second session record of traffic statistics for a second traffic flow including a source tag 2A and a destination of the database 308A. The policy controller 23 determines that the tag 2A applies to the API process 306A. Therefore, the policy controller 23 can associate the second session record of the second traffic flow with a second session record corresponding to the workload of the API process 306A (e.g., as the source of the second traffic flow). In addition to or in lieu of associating the first session record of traffic statistics for the traffic flow with the second session record of traffic statistics corresponding to the workload of the API process 306A (e.g., as the destination of the traffic flow), the policy controller 23 can also associate the second session record of the second traffic flow with the second session record corresponding to the workload of the API process 306A.

[0099] As about Figure 3 To further illustrate the example, the policy controller 23 associates the session record of the traffic statistics of the traffic flow including the source tag 1B and the destination tag 2B with the session record of the workload corresponding to the web process 304B and the session record of the workload corresponding to the API process 306B. The policy controller 23 associates the session record of the traffic statistics of the traffic flow including the source tag 2B and the destination tag 308A with the session record of the workload corresponding to the workload corresponding to the API process 306B.

[0100] Likewise, the policy controller 23 associates the session record of the traffic statistics of the traffic flow including the source tag 1C and the destination tag 2C with the session record of the workload corresponding to the web process 304C and the session record of the workload corresponding to the API process 306C. The policy controller 23 associates the session record of the traffic statistics of the traffic flow including the source tag 2C and the destination tag 308B with the session record of the workload corresponding to the workload corresponding to the API process 306C.

[0101] In addition, the policy controller 23 associates the session record of the traffic statistics of the traffic flow including the source tag 1D and the destination tag 2D with the session record of the workload corresponding to the web process 304D and the session record of the workload corresponding to the API process 306D. The policy controller 23 associates the session record of the traffic statistics of the traffic flow including the source tag 2D and the destination tag 308B with the session record of the workload corresponding to the workload corresponding to the API process 306D.

[0102] The policy controller 23 generates application firewall policies 202 for the multiple workloads 37 based on the multiple session records of the business statistics of the multiple workloads 37. Each of the multiple application firewall policies includes one or more application firewall policy rules, and each of the one or more application firewall policy rules defines whether one of the VR agents 35 should allow or deny the business flow of the corresponding workload 37.

[0103] For example, and about Figure 3 In the example of, the policy controller 23 determines the workload of the web process 304A corresponding to the web process 304A sending network business to the API process 306A according to the first session record of business statistics. Therefore, the policy controller 23 generates a first application firewall policy for the web process 304A, which includes an application firewall policy rule that allows network business originating from the web process 304A and sent to the API process 306A. In some examples, the first application firewall policy for the web process 304A also specifies an application firewall policy rule that blocks all other business originating from the web process 304A or sent to the web process 304A.

[0104] As another example, the policy controller 23 determines the workload of the API process 306A corresponding to the web process 304A sending network traffic to the API process 306A and the API process 306A sending network traffic to the database 308A based on the second session record of the traffic statistics. Therefore, the policy controller 23 generates a second application firewall policy for the API process 306A, which includes a first application firewall policy rule that allows network traffic originating from the web process 304A and sent to the API process 306A, and a second application firewall policy rule that allows network traffic originating from the API process 306A and sent to the database 308A. In some examples, the second application firewall policy for the API process 306A also specifies an application firewall policy rule that blocks all other traffic originating from or sent to the API process 306A.

[0105] Policy controller 23 can use the above operations to generate similar application firewall policies for web process 304B, API process 306B, web process 304C, API process 306C, web process 304D, and API process 306D. Therefore, policy controller 23 can create application firewall policy rules that allow network traffic between various workloads identified by the session records of the workloads.

[0106] In some examples, application firewall policy 202 also specifies application firewall policy rules that block all other traffic except for the specific traffic flow described by the session record of the workload. For example, because neither the session record of web process 304A nor the session record of API 306B identifies the traffic flow between web process 304A and API 306B, the application firewall policy of web process 304A or API 306B can prohibit such traffic between web process 304A and API 306B. Similarly, traffic between web process 304A and API 306C can be prohibited, and traffic between API 306D and database 308A can be prohibited.

[0107] The policy controller 23 distributes the application firewall policy 202 to the VR agent 35 to apply to the business flow. Therefore, when deployed in a production environment, the VR agent 35 or the policy agent 35 can apply the application firewall policy to the business flow in the production environment.

[0108] For example, policy agent 35A receives a first application firewall policy for web process 304A, which includes a first application firewall policy rule that allows network traffic originating from web process 304A and destined for API process 306A. Policy agent 35A may then allow network traffic flowing from an interface of a virtual machine executing web process 304A to an interface of a virtual machine executing API process 306A, and block other traffic flowing into or out of an interface of a virtual machine executing web process 304A that is not destined for API process 306A.

[0109] As another example, policy agent 35A receives a second application firewall policy for API process 306A, which includes a first application firewall policy rule that allows network traffic originating from web process 304A and destined for API process 306A, and a second application firewall policy rule that allows network traffic originating from API process 306A and destined for database 308A. Policy agent 35A may then allow network traffic flowing from an interface of a virtual machine executing web process 304A to an interface of a virtual machine executing API process 306A, or from an interface of a virtual machine executing API process 306A to an interface of database 308A, and block other traffic flowing into or out of an interface of a virtual machine executing API process 306A.

[0110] In the above example, the policy agent 35A implements a whitelist, wherein all traffic between designated interfaces of the virtual machine is preemptively blocked, and one or more policy rules whitelist specific traffic flows (e.g., specifying specific traffic flows between interfaces of the virtual machine that are allowed). In an alternative example, the policy agent 35 implements a blacklist, wherein all traffic between interfaces of the virtual machine is preemptively allowed, and one or more policy rules blacklist specific traffic flows (e.g., specifying specific traffic flows between interfaces of the blocked virtual machine). In an alternative example, one or more policy rules may include a combination of whitelisting or blacklisting specific traffic flows or traffic flow groups between interfaces of the virtual machine.

[0111] Therefore, the technology of the present disclosure can allow the policy controller 23 to automatically generate intent-based application firewall policies for workloads 37. In other words, the policy controller 23 can use the session records of the business statistics of workloads 37 to identify the "expected" business relationship between workloads 37 within the development environment, and use this "expected" relationship to generate application firewall policies for workloads 37 for use in the production environment to allow such "expected" business and block other forms of business. Therefore, the technology of the present disclosure can allow the creation of firewall policies that are more comprehensive, accurate, flexible, and effectively generated than traditional methods.

[0112] Figure 4 FIG. 1 is a block diagram of an example computing device 100 that executes a virtual router 120 of a virtual network according to one or more aspects of the present disclosure. The computing device 100 may represent Figures 1 to 3Any server 12 or other device in, such as any TOR switch 16. In this example, the computing device 100 includes a system bus 142 that couples the hardware components of the computing device 100 of the hardware environment. The system bus 142 couples a memory 144, a network interface card (NIC) 106A-106B (collectively referred to as "NIC 106"), a storage disk 107, and a multi-core computing environment 102 having multiple processing cores 108A-108J (collectively referred to as "processing core 108"). The network interface card 106 includes an interface configured to use a link exchange packet of the underlying physical network. The multi-core computing environment 102 can include any number of processors and any number of hardware cores, for example, from 4 to thousands. Each processing core 108 each includes an independent execution unit to execute instructions that conform to the instruction set architecture of the core. The processing cores 108 can each be implemented as a separate integrated circuit (IC), or can be combined in one or more multi-core processors (or "multi-core" processors), each of which is implemented using a single IC (i.e., a single-chip multiprocessor).

[0113] Disk 107 represents a computer-readable storage medium, which includes volatile and / or nonvolatile, removable and / or non-removable media implemented in any method or technology for storing information such as processor-readable instructions, data structures, program modules or other data. Computer-readable storage media include, but are not limited to, random access memory (RAM), read-only memory (ROM), EEPROM, flash memory, CD-ROM, digital versatile disk (DVD) or other optical storage devices, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by core 108.

[0114] The main memory 144 includes one or more computer-readable storage media, which may include random access memory (RAM), such as various forms of dynamic RAM (DRAM) (e.g., DDR2 / DDR3 SDRAM), or static RAM (SRAM), flash memory, or any other form of fixed or removable storage media that can be used to carry or store the required program code and program data in the form of instructions or data structures and can be accessed by the computer. The main memory 144 provides a physical address space consisting of addressable memory locations.

[0115] In some examples, memory 144 may present a non-uniform memory access (NUMA) architecture to multi-core computing environment 102. That is, cores 108 may not have the same memory access times as the various storage media that make up memory 144. In some cases, cores 108 may be configured to use the portion of memory 144 that provides the lowest memory latency to the core to reduce overall memory latency.

[0116] In some instances, the physical address space for computer-readable storage media can be shared between one or more cores 108 (i.e., shared memory). For example, cores 108A, 108B can be connected to one or more DRAM packages, modules, and / or chips (also not shown) via a memory bus (not shown), which presents a physical address space accessible by cores 108A, 108B. Although the physical address space can provide the lowest memory access time to cores 108A, 108B for any portion of memory 144, at least some of the remaining portions of memory 144 can be directly accessible to cores 108A, 108B. One or more of the cores 108 may also include L1 / L2 / L3 caches or a combination thereof. The corresponding caches of cores 108 provide the lowest latency memory access to any storage medium for cores 108.

[0117] Memory 144, network interface cards (NICs) 106A-106B (collectively referred to as "NICs 106"), storage disks 107, and multi-core computing environment 102 provide an operating environment for executing a software stack of virtual router 120 and one or more workloads 37. Workload 37 may represent Figure 2 The computing device 100 partitions the virtual and / or physical address space provided by the main memory 144, and in the case of virtual memory, by the disk 107, into a user space 111 allocated for running user processes, and a kernel space 112 that is protected and generally inaccessible by user processes. The operating system kernel ( Figure 4 37 (not shown in FIG. 3 ) may be executed in kernel space and may include, for example, Linux, Berkeley Software Distribution (BSD), another Unix variant kernel, or a Windows server operating system kernel available from Microsoft. In some cases, computing device 100 may execute a hypervisor to manage workload 37 (also not shown in FIG. 3 ). Figure 4 ). Figure 2 1. An example hypervisor 31 is shown in FIG. Example hypervisors include Kernel-based Virtual Machine (KVM) for the Linux kernel, Xen, ESXi available from VMware, Windows Hyper-V available from Microsoft, and other open source and proprietary hypervisors. In some examples, dedicated hardware programmed with routing information such as FIB 124 may implement virtual router 120. Although depicted as executing within kernel space 112 for ease of illustration, in some examples, workloads 37A-37K may execute in user space 111.

[0118] Eth0 114A and Eth1 114B represent devices according to a software device model and provide device driver software routines for processing packets received / sent by the corresponding NIC 106. Packets received by the NIC 106 from the underlying physical network structure for the virtual network may include an outer header to allow the physical network structure to tunnel the payload or "inner packet" to the physical network address of one of the NICs 106. The outer header may include not only the physical network address, but also a virtual network identifier, such as a VxLAN tag or a multi-protocol label switching (MPLS) label, which identifies a virtual network and a corresponding routing instance 122. The inner packet includes an inner header with a destination network address that conforms to the virtual network addressing space of the virtual network identified by the virtual network identifier. For example, the virtual router forwarding plane 128 may receive a packet with an outer header from the NIC 106 via Eth1, the outer header including the VxLAN associated with the routing instance 122A in the virtual router forwarding plane 128. The packet may have an inner header with a destination network address that is the destination address of WL37A, which is routed into routing instance 122A via junction interface 146A.

[0119] The virtual router 120 in this example includes a kernel space 112 module: a virtual router forwarding plane 128, and a user space 111 module: a VN agent 35. The virtual router forwarding plane 128 performs the "forwarding plane" or packet forwarding function of the virtual router 120, and the VN agent 35 performs the "control plane" function of the virtual router 120. The VN agent 35 may represent Figure 2 An example instance of any VN Proxy 35.

[0120] The virtual router forwarding plane 128 includes a plurality of routing instances 122A-122C (collectively referred to as "routing instances 122") for corresponding virtual networks. Each routing instance 122 includes a corresponding one of a forwarding information base (FIB) 124A-124C (collectively referred to as "FIB 124") and a flow table 126A-126C (collectively referred to as "flow table 126"). Although shown as separate data structures, the flow table 126 may in some cases be a logical table implemented as a single table or other associated data structure, wherein the entries of each flow table 126 may be identified by a virtual network identifier (e.g., a VRF identifier such as a VxLAN tag or an MPLS label). The FIB 124 includes a lookup table that maps a destination address to a destination next hop. The destination address may include a Layer 3 network prefix or a Layer 2 MAC address. The flow table 126 enables a forwarding policy to be applied to a flow. Each flow table 126 includes flow table entries, each of which matches one or more flows that can traverse the virtual router forwarding plane 128, and includes a forwarding policy for applying to the matching flows. For example, the virtual router forwarding plane 128 attempts to match a packet processed by routing instance 122A with one of the flow table entries of flow table 126A. If a matching flow table entry exists for a given packet, the virtual router forwarding plane 128 applies the flow action specified in the policy to the packet. This can be referred to as "fast path" packet processing. If a matching flow table entry does not exist for the packet, the packet can represent the initial packet of a new packet flow, and the virtual router forwarding plane 128 can request the VN agent 35 to install a flow table entry in the flow table for the new packet flow. This can be referred to as "slow path" packet processing for the initial packet of a packet flow, and in the example shown in FIG. 1 , the VN agent 35 may install a flow table entry in the flow table for the new packet flow. Figure 4 denoted by slow path 140.

[0121] In this example, the VN agent 35 may be a user space 111 process executed by the computing device 100. The VN agent 35 includes configuration data 134, virtual router and forwarding instance configuration 136 ("VRF 136"), and policy table 138 ("policy 138"). The VN agent 35 communicates with one or more virtual network controllers (e.g., Figure 1 to Figure 2 The VN agent 35 may also report analysis status, install forwarding status to the FIB 124 of the virtual router forwarding plane 128, discover WL 37 and its attributes. As described above, the VN agent 35 also applies slow path packet processing to the first (initial) packet of each new flow that traverses the virtual router forwarding plane 128, and installs the corresponding flow entry to the flow table 126 for fast path processing of the virtual router forwarding plane 128 for subsequent packets of the flow.

[0122] In some examples, the VN agent 35 also includes a tag database 141. As described above, the policy controller 23 of the SDN controller 132 assigns one or more tags of one or more levels and / or one or more categories of a specified object. The policy controller 23 can apply tags to objects across multiple levels or categories, and apply tags to objects across multiple data centers. The policy controller 23 distributes these tags to the VN agent 35, and the VN agent 35 stores the assigned tags in the tag database 141.

[0123] exist Figure 4 In the example of FIG. 1 , the policy agent 139 is a subcomponent of the VN agent 35. The policy agent 139 can be used with Figure 1 The policy agent 35 of the embodiment of the present invention functions in a substantially similar manner. The policy agent 139 receives one or more policies from the policy controller 23 and stores such policies in the policy 138. Each of the policies 138 may include one or more policy rules, which specify whether a specific network business should be allowed or blocked. In addition, each policy rule includes one or more policy tags, which specify which objects (such as applications executed on WL 37) apply the policy rule. For example, the policy agent 139 can cross-reference one or more tags of the policy rule with one or more tags in the tag database 141 to determine one or more objects to which the policy rule should be applied. When determining one or more objects to which the policy rule should be applied, the policy agent 139 allows or blocks network business originating from or sent to one or more objects according to the policy rule. In some examples, the policy agent 139 allows or blocks network business at the interface level of WL 37. In this example, the interface 146 inherits the tag from the parent object (such as WL 37, virtual network, parent project or global policy set 10). Thus, policy agent 139 allows or blocks network traffic to or from interface 146 of WL 37 by matching the markings of one or more policy rules with the markings applied to interface 146 .

[0124] For example, policy agent 35A receives a first application firewall policy for web process 304A, which includes a first application firewall policy rule that allows network traffic originating from web process 304A and destined for API process 306A. Policy agent 35A may then allow network traffic flowing from an interface of a virtual machine executing web process 304A to an interface of a virtual machine executing API process 306A, and block other traffic flowing into or out of an interface of a virtual machine executing web process 304A that is not destined for API process 306A.

[0125] As another example, policy agent 35A receives a second application firewall policy for API process 306A, which includes a first application firewall policy rule that allows network traffic originating from web process 304A and destined for API process 306A, and a second application firewall policy rule that allows network traffic originating from API process 306A and destined for database 308A. Policy agent 35A may then allow network traffic flowing from an interface of a virtual machine executing web process 304A to an interface of a virtual machine executing API process 306A, or from an interface of a virtual machine executing API process 306A to an interface of database 308A, and block other traffic flowing into or out of an interface of a virtual machine executing API process 306A.

[0126] In order to update the policy for a specific object or group of objects, the administrator can instruct the policy controller 23 to distribute the new policy to the policy agents 139 of the various VN agents 35 of multiple data centers 10. In some examples, the policy controller 23 distributes the new policy via BGP messages. In this way, the administrator can manage a scalable multi-dimensional policy framework for multiple data centers. In addition, each policy agent 139 can automatically apply relevant policies without the administrator having to individually manage or update the policies of each VN agent 35. Therefore, the technology disclosed in the present invention can greatly reduce the complexity and simplify the management of such policies across multiple data centers 10.

[0127] Figure 5 FIG. 5 is a block diagram of an example network 500 in which a policy controller 23 manages servers or computing nodes and provides visualization of various aspects related to workflows and policy objects associated with a computer network through a dashboard 203 according to one or more aspects of the present disclosure. Figure 5 In the example shown, policy controller 23 is shown as part of cluster 502. However, in other examples, policy controller 23 is not part of cluster 502, and in such examples, "cluster 502" refers to the computing nodes rather than the controllers. Figure 5 The cluster 502 may represent a cloud-based computing network and / or a computing domain or project, or other types of computing clusters. Figure 5 In the case where the cluster 502 represents a cloud environment, such a cloud environment may be an OpenStack cloud environment or a Kubernetes cloud environment. The cluster 502 may be distributed in multiple environments, such as different data centers.

[0128] The policy controller 23 may be combined with Figures 1 to 4For example, the policy controller 23 may collect and publish information related to the tag name or tag value assigned to each tag that has been assigned to a workload executed as part of a project and / or a workload executed by a virtual network included in the computer system.

[0129] The dashboard 203 may be combined with Figures 1 to 4 The method described and shown is implemented. Figure 1 As described above, dashboard 203 may be created, updated, and / or maintained primarily by policy controller 23 and / or a dashboard module executed on policy controller 23. In some examples, a computing device separate from policy controller 23 may generate dashboard 203. Figure 5 As shown in FIG. 5 , the policy controller 23 may generate a dashboard 203, which may represent a collection of graphical user interfaces (e.g., including user interface 510) that provide information about topology, structure, hierarchy, utilization, and / or performance status and metrics associated with infrastructure elements and workloads executed within the monitored computer network. The policy controller 23 generates and outputs a graphical user interface screen (e.g., user interface 510) for display by the user interface device 129. Figure 5 In the example of FIG. 5 , user interface 510 includes a set of graphical elements arranged within a graphical user interface according to the examples and techniques described below and any equivalents thereof. Policy controller 23 may use tag information and tag values ​​assigned to tags to tag graphical elements being displayed as part of a graphical user interface displayed by user interface device 129. Such tag information may identify each server 12 as hosting one or more instances of a policy object and may be arranged in a manner showing the categories and subcategories assigned to those policy objects.

[0130] Each graphical element associated with a particular policy object being displayed may be presented using a different color to help provide visual distinction between separate policy objects within the graphical user interface. Additionally, an interconnect element representing communication between policy objects may extend, for example, between a pair of graphical elements representing two respective policy objects. A color indicator may be used to present the interconnect element, the color being used to indicate a particular state of a communication assignment or definition described by a particular interconnect element.

[0131] In some examples, the policy controller 23 may generate or update a graphical user interface within the dashboard 203 in response to input from a user (e.g., an interaction within a display area of ​​the user interface 510) such that infrastructure elements are filtered by device type, interconnect, health, risk, aggregation, project, network (virtual or physical), type, and / or in other ways, such as by tagging and / or category / subcategory grouping. In such examples, filters may cause some graphical elements to be hidden within the dashboard 203 or the user interface of the dashboard 203, while presenting other graphical elements within the dashboard 203 based on the filter input. Filters may be applied using a functional paradigm. For example, each filter may correspond to a function, so that for a given set of "x" resources, elements, or other items to be filtered, and filter functions "f" and "g", application of the filter may calculate f(g(x)). Where filters follow a functional paradigm, applying the filters in a different order (e.g., g(f(x))) will have the same result. In some examples, some or all of the filter functions are pure, so as not to change state information outside the context of the function. Each time a filter value (or other user interface component) is changed, the policy controller 23 can apply the appropriate filter to all resources in the resource hash table and then reallocate the resource array to the result array. Filters can be applied and / or reapplied when an interaction with the user interface 510 is detected, when one or more infrastructure element representations within the main display area 513 are selected or changed, when the policy controller 23 receives metric data, and / or in other circumstances.

[0132] In some examples, as part of generating user interface 510, policy controller 23 provides a visual display of a circular ring, which may be referred to as a "tag view" and may be displayed as a default version of the graphical user interface provided by dashboard 203. The "tag view" of the graphical user interface may also be displayed based on input from dashboard 203 receiving a selection of a "tag" option included in a view selection menu, as described below with reference to FIG. Figure 8 In other examples, policy controller 23 provides an arrangement of information in a two-dimensional grid of one or more columns and one or more rows. Various user inputs may be provided to the displayed graphical user interface via dashboard 203 at any given time to allow more or less detail related to a particular policy object, project, particular communication interconnect, particular agent, virtual network, or to a particular port or virtual address to be presented as a user interface displayed by dashboard 203.

[0133] In some examples, the policy controller 23 presents one or more application firewall policies for review to a user or administrator via the dashboard 203, the application firewall policies being based on the above information. Figures 1 to 4As described above, in response to receiving input from a user or administrator approving one or more application firewall policy rules, the policy controller 23 distributes the one or more application firewall policy rules to the VR agent 35 for application to network traffic.

[0134] The systems, apparatus, and methods described in the present disclosure allow for the generation of these graphical user interfaces that depict various graphical depictions associated with the operation of projects and virtual networks, and associated communications occurring in the monitored computer network, in real time or near real time. In addition, the graphical user interface allows for the display of these projects and virtual networks in a manner that organizes these graphical elements that depict policy objects and associated communications and communication interconnections based on tags and categories / subcategories assigned to these policy objects.

[0135] Although the graphical elements described throughout this disclosure may relate to projects and virtual networks operating within a computer network modeled by a graphical user interface, the representations included in the examples of the graphical user interface are not limited to projects and / or virtual networks. The graphical elements described in this disclosure may include graphical symbols representing virtual entities (such as virtual servers, virtual routers) and / or virtual instances of other computer / software entities that may have connections to one or more physical devices and / or to other virtual entities that may be graphically presented and visually represented in a graphical user interface according to any of the examples described in this disclosure.

[0136] The graphical user interfaces described in the present disclosure and any equivalents thereof that can be generated and displayed using the systems, apparatus, and methods described herein provide visualization tools that can allow users (such as information technology (IT) professionals or technicians) to quickly and effectively model, visualize, and locate problems related to computer network performance in real time or near real time. The user interface as described in the present disclosure allows users to view various relationships between policy objects (including communication interconnections and policy rules associated with these communication interconnections of multiple policy objects operating within the computer system depicted by the graphical user interface) and quickly move to a more detailed graphical user interface (which provides a visual representation of more specific information, such as graphical and / or textual information associated with various policy objects and / or communications occurring within the computer network based on simple inputs to the system generating the graphical user interface (such as inputs to the user interface 129 coupled to the policy controller 23).

[0137] Figure 66 is a table showing multiple example session records 600 for traffic statistics of traffic flows according to the techniques of this disclosure. In some examples, as described above, the traffic collector 38 can collect multiple session records 600. In some examples, the session record 600 specifies multiple information and metadata describing the flow for each traffic flow. As an example, the session record 600 specifies 5-tuple information (e.g., source address of the flow, source port of the flow, destination address of the flow, destination port of the flow, and protocol of the flow) for each traffic flow.

[0138] In some examples, session record 600 specifies other types of information, such as tags applied to one or more objects associated with the flow. Figure 6 In the example of , session record 600 specifies a tag of a workload associated with the flow (eg, a web application, a financial application, an email application, etc.).

[0139] Figure 7 This is to show that the technology according to the present disclosure will Figure 6 A plurality of example session records 600 of traffic statistics for a traffic flow are converted into a graphical representation of a plurality of firewall policies.

[0140] Based on the tags of session record 600 applied to one or more objects associated with the flow, policy controller 23 associates session record 600 with session records of business statistics for one or more workloads associated with the tags. For example, policy controller 23 may associate flows 1, 2, 3, and 6 with session records of business statistics for web applications 702A, flows 4, 5, 7, and 8 with session records of business statistics for financial applications 702B, and flows 9 and 10 with session records of business statistics for email applications 702C (collectively referred to as “session records of business statistics for workloads 702”) based on the tags specified in session record 600.

[0141] The policy controller 23 generates one or more application firewall policies 704A-704C (collectively referred to as "application firewall policies 704") based on the session record 702 of the business statistics of the workload. For example, based on the session record 702A of the web application, the policy controller 23 generates the web application policy 704A, which includes web application policy rules that allow corresponding business flows 1, 2, 3, and 6, respectively. Similarly, based on the session record 702B of the financial application, the policy controller 23 generates the financial application policy 704B, which includes financial application policy rules that allow corresponding business flows 4, 5, 7, and 8, respectively. In addition, based on the session record 702C of the email application, the policy controller 23 generates the email application policy 704C, which includes email application policy rules that allow corresponding business flows 9 and 10, respectively. In some examples, each of the application firewall policies 704 blocks all other businesses associated with the corresponding application that are not explicitly whitelisted by the application firewall policy rules of the application firewall policy 704.

[0142] Figure 8 Flowchart showing an example operation for generating multiple firewall policies according to the technology of the present disclosure. Figure 8 is relative to Figure 1 described, although Figure 8 The operation can also be applied to Figure 2 and Figure 3 .

[0143] In one example, Figure 2 The virtual routers 30 process ingress and egress flows of network services generated by one or more applications and forwarded by each virtual router 30 to generate service flow statistics. Each virtual router 30 transmits service flow statistics information to a corresponding VR agent 35. Each VR agent 35 processes the service flow statistics of the ingress and egress flows to identify pairs of ingress and egress flows corresponding to a common communication session between workflows 37. For each identified pair, the VR agent 35 generates a session record of service statistics of network services forwarded by the virtual router 30. The VR agent 35 uploads the session record together with the tag information of the given flow to the service collector 38 of the SDN controller 32 for the cloud data center 10. The service collector 38 receives the session records from multiple VR agents 35 and forwards the session records to the policy controller 23.

[0144] The policy controller 23 receives session records of traffic statistics of multiple traffic flows from the traffic collector 38 (802). The policy controller 23 associates the session records of the multiple traffic flows to the session records of traffic statistics of multiple application workloads 37 based on the tags specified by the session records of the multiple traffic flows (804). For example, the policy controller 23 matches the application workload 37 with a specific session record of the traffic flow using the tags specified by the session records of the multiple traffic flows. Based on the tags applied to the workload 37 that match the tags specified by the session records of the multiple traffic flows, the policy controller 23 aggregates the session records of the traffic flows into the session records of the application workload 37.

[0145] The policy controller 23 generates one or more application firewall policies for multiple application workloads 37 based on the session records of the traffic statistics of the multiple application workloads 37 (806). Each of the multiple application firewall policies includes one or more application firewall policy rules, and each of the one or more application firewall policy rules defines whether one of the VR agents 35 should allow or deny the traffic flow of the corresponding workload 37. For example, the policy controller 23 can use the 5-tuple of the session records for the workload 37 to define the 5-tuple of the application firewall policy rules. In addition, the policy controller 23 can define the rules of the application firewall policy rules based on the session records (e.g., allow, block, log, or report the traffic corresponding to the rules). In some examples, each application firewall policy rule specifies a default rule that allows the corresponding network traffic and blocks all other network traffic that does not match the generated application firewall policy rule.

[0146] The policy controller 23 distributes one or more application firewall policies to the virtual router agent 35 for application to multiple business flows (808). The VR agent 35 can then apply the application firewall policy to the business flows within the network 8. In some examples, before distributing the one or more application firewall policies, the policy controller 23 presents the one or more application firewall policies to the user or administrator via the UI device 129 for review. In response to receiving input from the user or administrator approving one or more application firewall policy rules, the policy controller 23 distributes the one or more application firewall policy rules to the VR agent 35 for application to network business. The VR agent 35 that collects business statistics and relays the business statistics to the policy controller 23 can be different from the VR agent that receives and applies one or more application firewall policies applied to subsequent business between workloads 37.

[0147] Therefore, the technology of the present disclosure can allow the policy controller 23 to automatically generate intent-based application firewall policies for workloads 37. In other words, the policy controller 23 can use the session records of the business statistics of workloads 37 to identify the "expected" business relationship between workloads 37 within the development environment, and use this "expected" relationship to generate application firewall policies for workloads 37 for use in the production environment to allow such "expected" business and block other forms of business. Therefore, the technology of the present disclosure can allow the creation of firewall policies that are more comprehensive, accurate, flexible, and effectively generated than traditional methods.

[0148] Fig. 9 Flowchart showing an example operation for generating multiple firewall policies according to the technology of the present disclosure. Fig. 9 is relative to Figure 1 described, although Fig. 9 The operation can also be applied to Figure 2 and Figure 3 .

[0149] Policy controller 23 receives session records of traffic statistics for multiple traffic flows from traffic collector 38 (902). Operation 902 may be similar to the above description regarding Figure 8 Operation 802 occurs in a substantially similar manner as described above.

[0150] The policy controller 23 associates the session records of the plurality of service flows with the session records of the service statistics of the plurality of application workloads 37 (904). Figure 8 In contrast to the operation of , the policy controller 23 associates the session records of the multiple business flows to the session records of the business statistics of the multiple application workloads 37 without using a matching tag. As an example, in order to associate the session records of the business flows to the session records of the workload 37, the policy controller 23 presents the multiple session records of the business statistics of the multiple business flows to the user or administrator via the UI device 129. In addition, the policy controller 23 receives input via the UI device 129 specifying which business flows among the multiple business flows correspond to the corresponding application workloads 37 among the multiple application workloads 37, and based on the input, aggregates the session records of the business flows into the session records of the workload 37.

[0151] In another example, the policy controller 23 associates the session records of the service flow to the session records of the workload 37 by applying a clustering algorithm to the session records of the service flow. For example, the policy controller 23 may use a clustering algorithm to identify the relationship between the session records of the service flow and a particular workload 37, and based on these relationships, aggregate the session records of the service flow into the session records of the workload 37.

[0152] The policy controller 23 records the session statistics based on the business statistics of multiple application workloads 37 to Figure 8 In a manner substantially similar to that described in operation 806 of FIG. 8 , the policy controller 23 generates one or more application firewall policies for the plurality of application workloads 37 ( 806 ). In addition, the policy controller 23 distributes the one or more application firewall policies to the virtual router agents 35 so as to be executed in the same manner as described above with respect to FIG. Figure 8 A substantially similar manner as described in operation 808 of is applied to multiple business flows (808). The VR agent 35 can then apply the application firewall policy to the business flows within the network 8. Therefore, the technology of the present disclosure can allow the policy controller 23 to automatically generate intent-based application firewall policies for workloads 37. In other words, the policy controller 23 can use session records of business statistics of workloads 37 to identify "expected" business relationships between workloads 37 within the development environment, and use this "expected" relationship to generate application firewall policies for workloads 37 for use in the production environment to allow such "expected" business and block other forms of business. Therefore, the technology of the present disclosure can allow the creation of firewall policies that are more comprehensive, accurate, flexible, and effectively generated than traditional methods.

[0153] The techniques described herein may be implemented in hardware, software, firmware, or any combination thereof. The various features described as modules, engines, units, or components may be implemented together in an integrated logic device, or separately as discrete but interoperable logic devices or other hardware devices. In some cases, various features of electronic circuits may be implemented as one or more integrated circuit devices, such as an integrated circuit chip or chipset.

[0154] If implemented in hardware, the present disclosure may be directed to an apparatus such as a processor or an integrated circuit device (such as an integrated circuit chip or chipset). Alternatively or additionally, if implemented in software or firmware, the technology may be implemented at least in part by a computer-readable data storage medium including instructions that, when executed, cause a processor to perform one or more of the above methods. For example, a computer-readable data storage medium may store such instructions for execution by a processor.

[0155] The computer-readable medium may form part of a computer program product, which may include packaging materials. The computer-readable medium may include computer data storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile random access memory (NVRAM), electrically erasable programmable read-only memory (EEPROM), flash memory, magnetic or optical data storage media, etc. In some examples, the article of manufacture may include one or more computer-readable storage media.

[0156] In some examples, computer-readable storage media may include non-transitory media. The term "non-transitory" may indicate that the storage medium is not implemented in a carrier wave or propagating signal. In some examples, non-transitory storage media may store data that changes over time (e.g., in RAM or cache).

[0157] The code or instructions may be software and / or firmware executed by a processing circuit, which may include one or more processors, such as one or more digital signal processors (DSPs), general-purpose microprocessors, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other equivalent integrated or discrete logic circuits. Thus, the term "processor" as used herein may refer to any of the foregoing structures or any other structure suitable for implementing the techniques described herein. Additionally, in some aspects, the functionality described in the present disclosure may be provided within a software module or a hardware module.

[0158] In addition to or as an alternative to the above, the following examples are described. Features described in any of the following examples may be used with any other examples described herein.

[0159] Example 1. A method, comprising: receiving, by a policy controller of a computer network, business statistics of multiple business flows between multiple application workloads, the multiple application workloads being executed by a first group of one or more computing devices of the computer network; associating, by the policy controller, the business statistics of the multiple business flows to session records of the business statistics of the multiple application workloads; generating, by the policy controller, one or more application firewall policies for the multiple application workloads based on the session records of the business statistics of the multiple application workloads, wherein the one or more application firewall policies define whether business flows between the application workloads of the multiple application workloads are allowed or denied; and distributing, by the policy controller, the one or more application firewall policies to a second group of one or more computing devices for application to business flows between instances of the application workloads.

[0160] Example 2. A method according to Example 1, wherein receiving traffic statistics for multiple traffic flows between multiple application workloads includes receiving traffic statistics for multiple traffic flows in the multiple application workloads from a first group of virtual router agents of a first group of one or more computing devices, and wherein distributing one or more application firewall policies to a second group of one or more computing devices for application to traffic flows between instances of the application workloads includes distributing one or more application firewall policies to a second group of virtual router agents of the second group of one or more computing devices for application to traffic flows between instances of the application workloads.

[0161] Example 3. The method according to Example 2 further includes applying, by a second set of virtual router agents, one or more application firewall policies to traffic flows between instances of the application workload to at least allow or deny at least one traffic flow.

[0162] Example 4. The method of any one of Examples 1 to 3 further includes presenting, by the policy controller, one or more application firewall policies for display to a user.

[0163] Example 5. A method according to any one of Examples 1 to 4, wherein associating the business statistics of multiple business flows to the session record of the business statistics of multiple application workloads includes: receiving, by a policy controller, input from a user specifying which business flows of the multiple business flows correspond to the application workloads of the multiple application workloads; and associating, by the policy controller based on the input, the business statistics of the multiple business flows to the session record of the business statistics of the multiple application workloads.

[0164] Example 6. A method according to any one of Examples 1 to 5, wherein associating the business statistics of multiple business flows to the session records of the business statistics of multiple application workloads includes: a policy controller associating the business statistics of multiple business flows to the session records of the business statistics of multiple application workloads based on one or more tags specified by the business statistics of the multiple business flows.

[0165] Example 7. A method according to any one of Examples 1 to 6, wherein associating the business statistics of multiple business flows to the session records of the business statistics of multiple application workloads includes: applying a clustering algorithm by a policy controller to identify which business flows among the multiple business flows correspond to the application workloads of the multiple application workloads; and associating the business statistics of the multiple business flows to the session records of the business statistics of the multiple application workloads based on the clustering algorithm.

[0166] Example 8. A method according to any one of Examples 1 to 7, wherein, based on session records of business statistics of multiple application workloads, generating one or more application firewall policies for multiple application workloads includes: for each application workload in the multiple application workloads: identifying one or more business flows of multiple business flows associated with the application workload; generating one or more application firewall policy rules for the application workload, wherein each of the one or more application firewall policy rules includes a 5-tuple and a rule, the 5-tuple being the same as the 5-tuple of the corresponding business flow of the one or more business flows associated with the application workload, the rule specifying an action to be performed on the corresponding business flow; and generating an application firewall policy for the application workload, the application firewall policy including the one or more application firewall policy rules for the application workload.

[0167] Example 9. The method according to Example 8, wherein the action performed on the corresponding business flow is one or more of the following: allowing the corresponding business flow and denying the corresponding business flow; recording the corresponding business flow; and reporting the corresponding business flow.

[0168] Example 10. A method according to any combination of Examples 1 to 9, wherein the session record of business statistics for each business flow of multiple business flows includes one or more of the following: one or more tags for the source of the business flow; one or more tags for the destination of the business flow.

[0169] Example 11. A policy controller for a computer network, the policy controller executing on a processing circuit and configured to: receive business statistics of multiple business flows between multiple application workloads, the multiple application workloads being executed by a first group of one or more computing devices of the computer network; associate the business statistics of the multiple business flows with session records of the business statistics of the multiple application workloads; generate one or more application firewall policies for the multiple application workloads based on the session records of the business statistics of the multiple application workloads, wherein the one or more application firewall policies define whether business flows between the application workloads of the multiple application workloads are allowed or denied; and distribute the one or more application firewall policies to a second group of one or more computing devices for application to business flows between instances of the application workloads.

[0170] Example 12. A policy controller according to Example 11, wherein, in order to receive business statistics of multiple business flows between multiple application workloads, the policy controller is also configured to receive business statistics of multiple business flows between the multiple application workloads from a first group of virtual router agents of a first group of one or more computing devices, and wherein, in order to distribute one or more application firewall policies to a second group of one or more computing devices for application to business flows between instances of the application workloads, the policy controller is also configured to distribute one or more application firewall policies to a second group of virtual router agents of the second group of one or more computing devices for application to business flows between instances of the application workloads.

[0171] Example 13. The policy controller of any of Examples 11 to 12, wherein the policy controller is further configured to present one or more application firewall policies for display to a user.

[0172] Example 14. A policy controller according to any one of Examples 11 to 13, wherein, in order to associate business statistics of multiple business flows with session records of business statistics of multiple application workloads, the policy controller is further configured to: receive input from a user specifying which business flows of the multiple business flows correspond to application workloads of the multiple application workloads; and associate the business statistics of the multiple business flows with the session records of the business statistics of the multiple application workloads based on the input.

[0173] Example 15. A policy controller according to any one of Examples 11 to 14, wherein, in order to associate the business statistics of multiple business flows with the session records of the business statistics of multiple application workloads, the policy controller is also configured to: associate the business statistics of the multiple business flows with the session records of the business statistics of multiple application workloads based on one or more tags specified by the business statistics of the multiple business flows.

[0174] Example 16. A policy controller according to any one of Examples 11 to 15, wherein, in order to associate the business statistics of multiple business flows to the session records of the business statistics of multiple application workloads, the policy controller is further configured to: apply a clustering algorithm to identify which business flows of the multiple business flows correspond to the application workloads of the multiple application workloads; and based on the clustering algorithm, associate the business statistics of the multiple business flows to the session records of the business statistics of the multiple application workloads.

[0175] Example 17. A policy controller according to any one of Examples 11 to 16, wherein, in order to generate one or more application firewall policies for multiple application workloads based on session records of business statistics of multiple application workloads, the policy controller is further configured to: for each application workload in the multiple application workloads: identify one or more business flows of the multiple business flows associated with the application workload; generate one or more application firewall policy rules for the application workload, wherein each of the one or more application firewall policy rules includes a 5-tuple and a rule, the 5-tuple is the same as the 5-tuple of the corresponding business flow of the one or more business flows associated with the application workload, and the rule specifies an action to be performed on the corresponding business flow; and generate an application firewall policy for the application workload, the application firewall policy including the one or more application firewall policy rules for the application workload.

[0176] Example 18. A policy controller according to Example 17, wherein the action performed on the corresponding business flow is one or more of the following: allowing the corresponding business flow and denying the corresponding business flow; recording the corresponding business flow; and reporting the corresponding business flow.

[0177] Example 19. A policy controller according to any combination of Examples 11 to 18, wherein a session record of traffic statistics for each of a plurality of traffic flows comprises one or more of: one or more tags for a source of the traffic flow; one or more tags for a destination of the traffic flow.

[0178] Example 20. A non-temporary computer-readable medium comprising instructions that, when executed, cause a processing circuit to execute a policy controller for a computer network to: receive business statistics for multiple business flows between multiple application workloads, the multiple application workloads being executed by a first group of one or more computing devices of the computer network; associate the business statistics for the multiple business flows with session records of the business statistics for the multiple application workloads; generate one or more application firewall policies for the multiple application workloads based on the session records of the business statistics for the multiple application workloads, wherein the one or more application firewall policies define whether business flows between the application workloads of the multiple application workloads are allowed or denied; and distribute the one or more application firewall policies to a second group of one or more computing devices for application to business flows between instances of the application workloads.

[0179] Furthermore, any specific features set forth in any of the above examples may be combined into beneficial examples of the described technology. That is, any specific features are generally applicable to all examples of the present invention. Various examples of the present invention have been described.

Claims

1. A method for a computer network, comprising: receiving, by a policy controller of a computer network, traffic statistics for a plurality of traffic flows between first instances of a plurality of application workloads, the first instances of the plurality of application workloads being executed by a first set of one or more computing devices of the computer network; applying, by the policy controller, a clustering algorithm to associate the traffic statistics of the plurality of traffic flows with session records of the traffic statistics of the plurality of application workloads; generating, by the policy controller and from a session record of traffic statistics of the plurality of application workloads, one or more indicia of the plurality of application workloads; generating, by the policy controller and based on the one or more tags of the multiple application workloads, one or more application firewall policies for the multiple application workloads, wherein the one or more application firewall policies define whether traffic flows between application workloads of the multiple application workloads are allowed or denied; and The one or more application firewall policies are distributed by the policy controller to a second group of one or more computing devices for application to business flows between second instances of the multiple application workloads, and the second instances of the multiple application workloads are executed by the second group of one or more computing devices, wherein the second group of one or more computing devices are different from the first group of one or more computing devices.

2. The method according to claim 1, wherein: Applying the clustering algorithm to associate the traffic statistics of the plurality of traffic flows with the session records of the traffic statistics of the plurality of application workloads includes: applying the clustering algorithm to identify at least one relationship between traffic statistics of the plurality of traffic flows; and Based on the at least one identified relationship between the traffic statistics of the plurality of traffic flows, the traffic statistics of the plurality of traffic flows are aggregated into a session record of the traffic statistics of the plurality of application workloads.

3. The method according to claim 1, wherein: The clustering algorithm includes one of the following: K-means clustering algorithm; Mean-shift clustering algorithm; Density-based noise application spatial clustering DBSCAN clustering algorithm; Expectation-Maximization (EM) clustering algorithm using Gaussian Mixture Models (GMM); or Agglomerative hierarchical clustering algorithm.

4. The method according to claim 1, wherein: The one or more tags specify a plurality of different object levels for the plurality of application workloads.

5. The method according to claim 4, wherein: The one or more tags specify one or more of the following: a global environment level of the plurality of application workloads; a project level of the plurality of application workloads; a virtual network level for the plurality of application workloads; a virtual machine level of the plurality of application workloads; or The interface levels of the plurality of application workloads.

6. The method according to claim 1, wherein: Generating one or more tags for the plurality of application workloads from the session recording of traffic statistics for the plurality of application workloads includes generating one or more tags for the plurality of application workloads from metadata of the session recording of traffic statistics for the plurality of application workloads.

7. The method according to any one of claims 1 to 6, wherein: The one or more application firewall policies define whether traffic flows between interfaces of the plurality of application workloads marked with one or more markings of the plurality of application workloads are allowed or denied.

8. The method according to any one of claims 1 to 6, in, receiving traffic statistics for a plurality of traffic flows between first instances of the plurality of application workloads comprises receiving traffic statistics for a plurality of traffic flows between first instances of the plurality of application workloads from a first set of virtual router agents of the first set of one or more computing devices, and Wherein, distributing the one or more application firewall policies to the second group of one or more computing devices for application to business flows between second instances of the multiple application workloads includes distributing the one or more application firewall policies to a second group of virtual router agents of the second group of one or more computing devices for application to business flows between second instances of the multiple application workloads.

9. The method of claim 8, further comprising applying, by the second set of virtual router agents, the one or more application firewall policies to traffic flows between second instances of the plurality of application workloads to at least allow or deny at least one traffic flow.

10. The method of any one of claims 1 to 6, further comprising presenting, by the policy controller, the one or more application firewall policies for display to a user.

11. A policy controller for a computer network, the policy controller comprising processing circuitry and configured to: receiving traffic statistics for a plurality of traffic flows between first instances of a plurality of application workloads executed by a first set of one or more computing devices of a computer network; applying a clustering algorithm to correlate the traffic statistics of the plurality of traffic flows to session records of the traffic statistics of the plurality of application workloads; generating one or more indicia of the plurality of application workloads from a session record of traffic statistics of the plurality of application workloads; generating one or more application firewall policies for the multiple application workloads based on the one or more tags of the multiple application workloads, wherein the one or more application firewall policies define whether traffic flows between the application workloads of the multiple application workloads are allowed or denied; and Distribute the one or more application firewall policies to a second group of one or more computing devices for application to business flows between second instances of the multiple application workloads, where the second instances of the multiple application workloads are executed by the second group of one or more computing devices, wherein the second group of one or more computing devices are different from the first group of one or more computing devices.

12. The policy controller according to claim 11, wherein: In order to apply the clustering algorithm to associate the traffic statistics of the plurality of traffic flows into the session records of the traffic statistics of the plurality of application workloads, the policy controller is further configured to: applying the clustering algorithm to identify at least one relationship between the traffic statistics of the plurality of traffic flows; as well as Based on the at least one identified relationship between the traffic statistics of the plurality of traffic flows, the traffic statistics of the plurality of traffic flows are aggregated into a session record of the traffic statistics of the plurality of application workloads.

13. The policy controller according to claim 11, wherein: The clustering algorithm includes one of the following: K-means clustering algorithm; Mean-shift clustering algorithm; Density-based noise application spatial clustering DBSCAN clustering algorithm; Expectation-Maximization (EM) clustering algorithm using Gaussian Mixture Models (GMM); or Agglomerative hierarchical clustering algorithm.

14. The policy controller according to claim 11, wherein: The one or more tags specify a plurality of different object levels for the plurality of application workloads.

15. The policy controller according to claim 14, wherein: The one or more tags specify one or more of the following: a global environment level of the plurality of application workloads; a project level of the plurality of application workloads; a virtual network level for the plurality of application workloads; a virtual machine level of the plurality of application workloads; or The interface levels of the plurality of application workloads.

16. The policy controller according to claim 11, wherein: In order to generate one or more tags of the multiple application workloads from the session records of the traffic statistics of the multiple application workloads, the policy controller is configured to generate the one or more tags of the multiple application workloads from metadata of the session records of the traffic statistics of the multiple application workloads.

17. The policy controller according to any one of claims 11 to 16, wherein: The one or more application firewall policies define whether traffic flows between interfaces of the plurality of application workloads marked with one or more markings of the plurality of application workloads are allowed or denied.

18. A policy controller according to any one of claims 11 to 16, in, To receive traffic statistics of a plurality of traffic flows between first instances of a plurality of application workloads, the policy controller is further configured to receive, from a first set of virtual router agents of the first set of one or more computing devices, statistics of a plurality of traffic flows between first instances of the plurality of application workloads, and In which, in order to distribute the one or more application firewall policies to the second group of one or more computing devices for application to business flows between second instances of the multiple application workloads, the policy controller is also configured to distribute the one or more application firewall policies to a second group of virtual router agents of the second group of one or more computing devices for application to business flows between second instances of the multiple application workloads.

19. The policy controller according to any one of claims 11 to 16, further configured to present the one or more application firewall policies for display to a user.

20. A computer-readable storage medium storing a computer program, wherein the computer program can be executed by a processor to implement the method for a computer network according to any one of claims 1 to 10.