Virtual-real interconnection method of multiple access modes

Abstract

The invention belongs to the technical field of cloud platforms, and particularly relates to a virtual-real interconnection method for multiple access modes, which comprises the following steps of: 1, establishing mapping between a virtual network identifier and a virtual local area network (VLAN); 2, constructing a unified physical transmission channel and scheduling traffic; 3, establishing and dynamically maintaining a logic mapping function; 4, generating and optimizing an abstract configuration strategy; 5, compilation and optimization of multi-manufacturer equipment instructions are carried out; and step 6, performing atomization issuing and consistency verification of the configuration. According to the method, performance loss and protocol compatibility limitation caused by traditional three-layer NAT forwarding are effectively avoided, and high-performance communication of a native two-layer network is provided; and special VXLAN gateway equipment is replaced by general VLAN switch networking, so that the hardware cost and the technical threshold are greatly reduced.

Description

Technical Field

[0001] This invention belongs to the field of cloud platform technology, specifically relating to a virtual-physical interconnection method with multiple access methods. Background Technology

[0002] In current OpenStack-based cloud platforms, achieving efficient interconnection between virtual machines and physical devices is crucial for improving overall resource efficiency. While traditional Layer 3 NAT gateways can achieve interconnection, they suffer from limitations such as performance degradation and protocol compatibility issues. Direct Layer 2 bridging, on the other hand, requires frequent manual configuration of physical switch ports, resulting in low automation and cumbersome maintenance. Therefore, there is an urgent need for a solution that can be deeply integrated into the cloud platform, providing high-performance Layer 2 network access while centrally managing physical network configurations through software-defined methods. This would enable a single physical device connection with flexible logical control, significantly improving the automation level and operational efficiency of virtual-physical interconnection.

[0003] Defects of the Background Technology Existing cloud virtual machines typically communicate with external physical devices in two ways: one is through Layer 3 NAT forwarding via OpenStack virtual routers, but this method incurs additional performance overhead and cannot support critical business protocols that rely on Layer 2 broadcast or multicast; the other is through Layer 2 interconnection via overlay technologies such as VXLAN, but this method requires physical switches to support VXLAN gateway functionality, and because different vendors' switch configuration instructions and implementation mechanisms vary, complex manual adjustments are required for each device, resulting in high deployment and maintenance costs, as well as low automation and poor scalability in heterogeneous network environments. Summary of the Invention

[0004] (a) Technical problems to be solved The technical problem to be solved by this invention is: how to provide a virtual-physical interconnection method with multiple access methods.

[0005] (II) Technical Solution To address the aforementioned technical problems, this invention provides a virtual-physical interconnection method with multiple access methods. The method is developed based on OpenStack and improved upon it to achieve virtual-physical interconnection functionality. Specifically, the Layer 3 access uses the external network NAT scheme provided by OpenStack, while the Layer 2 access is designed and implemented independently. When each virtual machine's network interface card (NIC) is connected to the br-int virtual bridge, it is equivalent to being connected to a Layer 2 switch and configured with a VLAN. The br-int is then connected to a physical NIC, which is used solely for virtual-physical integration. The physical device is connected to the access switch, and this NIC is also connected to the access switch. The NIC is configured to communicate in Trunk mode. The port on the access switch corresponding to the physical device is then configured with the same VLAN as the virtual subnet, enabling communication. This approach, using a one-time physical device connection and logical management, avoids frequent adjustments to the physical connections of the devices, improving operability.

[0006] The method includes: Step 1: Establish virtual network identifier and VLAN mapping; Step 2: Construction and traffic scheduling of a unified physical transmission channel; Step 3: Establishment and dynamic maintenance of logical mapping functions; Step 4: Generation and optimization of abstract configuration strategies; Step 5: Compiling and optimizing instructions for multi-vendor devices; Step 6: Atomic deployment and consistency verification of configuration.

[0007] Step 1: Establishing the virtual network identifier and VLAN mapping; At the OpenStack compute node level, when a virtual machine's network data packet enters the data plane of the virtualization platform, a conversion mechanism between the virtual network identifier and the physical VLAN tag must first be established. This step designs an efficient tag mapping module at the br-int bridge, which maintains a mapping table between virtual network IDs and local VLAN IDs. The specific implementation process is as follows: when a raw data packet from a virtual machine is detected, the system queries the virtual network ID to which the virtual machine belongs, and then quickly locates the corresponding local VLAN tag through a hash algorithm.

[0008] A unique identifier for a virtual network, typically a UUID assigned by OpenStack Neutron; Lightweight hash function to ensure even distribution of virtual networks to VLAN IDs; Modulo operation ensures that the generated VLAN ID is within the standard VLAN range (1-4094); The calculated local VLAN ID is used for subsequent physical network transmission.

[0009] Step 2: Construction and traffic scheduling of a unified physical transmission channel; To achieve efficient transmission of virtual network traffic over physical infrastructure, this invention designs an intelligent traffic scheduling mechanism; after the dedicated physical network card is configured in Trunk mode, a refined traffic classification and scheduling strategy needs to be established; the system will monitor the traffic scheduling of each VLAN in real time to ensure the network quality of critical services.

[0010] : Queue bandwidth allocated to the i-th VLAN; Total available bandwidth of the physical network interface card; The weight coefficient of the i-th VLAN is dynamically adjusted according to the service priority. : .

[0011] Step 3: Establishment and dynamic maintenance of the logical mapping function; The logical mapping function is the core component for implementing software-defined network management in this method. This function not only establishes the initial mapping relationship, but also needs to maintain the consistency of the mapping state in real time. The system adopts a distributed transaction mechanism to ensure strong consistency of the mapping relationship, ensuring that the mapping relationship between physical ports and virtual networks is accurate and reliable at any time.

[0012] t: timestamp, ensuring time consistency of the mapping relationship; Physical switch port identifier, including switch IP and port number; The 48-bit MAC address of the physical device; : The VLAN ID corresponding to this mapping relationship at time t.

[0013] Step 4: Generation and optimization of abstract configuration strategies; The abstract configuration strategy generator adopts an intent-based network programming paradigm. The system first parses the user's business intent, and then generates the optimal configuration strategy through a strategy optimization algorithm. The optimization process considers multiple constraints such as network topology, device performance, and security policies to generate the optimal solution that satisfies all constraints.

[0014] Optimal configuration strategy; The execution cost function of strategy S, including time complexity and resource overhead; The set of constraints that the strategy must satisfy; Find the strategy parameters that minimize the cost function.

[0015] Step 5: Compilation and optimization of instructions for multi-vendor devices; The instruction compiler adopts a multi-layer intermediate representation design architecture; first, it converts the abstract configuration strategy into a vendor-independent intermediate representation, then it passes through multiple rounds of optimization, and finally generates optimized device instructions based on the characteristics of the target device; the optimization process includes key technologies such as instruction merging, redundancy elimination, and parallelization.

[0016] Reduce high-level strategies to intermediate representations; Legalizing the target equipment ensures that instructions comply with equipment specifications; : Peephole optimization, performing local instruction optimization; The final generated sequence of instructions for the target device.

[0017] Step 6: Configured atomic distribution and consistency verification; The configuration distribution process employs a two-phase commit protocol to ensure atomicity; firstly, the configuration is pre-committed to all relevant devices, and then uniformly committed to take effect after all devices have successfully confirmed it; the verification phase uses formal verification methods, and model checks are used to ensure the correctness of the configuration.

[0018] AG: Globally always true, meaning it is always satisfied in all execution paths; AF: Eventually true, meaning it will definitely be satisfied at some point in the future; State: The actual operating status of the equipment; Expected: The expected state of the device.

[0019] The method involves the following: (1) Dual-mode network access integration: Based on the native Layer 3 NAT gateway access method of OpenStack, a Layer 2 L2 direct access solution was independently designed and implemented, providing users with the ability to flexibly choose high-performance Layer 2 interconnection or general Layer 3 interconnection according to business needs; (2) Hybrid bridging based on standard VLANs: The traffic of the virtual network is tagged with the local VLAN by the integrated bridge br-int on the computing node and is handled by a dedicated physical network card in Trunk mode, which transparently forwards the VLAN information of multiple virtual networks to the physical access switch; (3) Logical centralized management and control platform: Construct a management subsystem to abstract and input information on physical network resources and physical devices, and digitize and visualize physical connection relationships; the physical network resources include access switch IP and port number; (4) Software-defined automated configuration: The management platform automatically sends the VLAN configuration of the virtual network to the specified physical access switch port through standard protocols or SSH, replacing manual command line configuration and realizing the automation and intelligence of physical network configuration.

[0020] The standard protocols mentioned include SNMP and NETCONF.

[0021] (III) Beneficial Effects Compared with existing technologies, this invention is developed based on OpenStack and improved upon it to achieve virtual-physical interconnection functionality. Specifically, Layer 3 access uses the external network NAT solution provided by OpenStack, while Layer 2 access is designed and implemented independently. The purpose of this invention is to improve virtual-physical interconnection technology based on OpenStack and various access methods. By optimizing the Layer 2 network access architecture, while retaining the platform's original Layer 3 NAT solution, it innovatively adopts a physical network card trunk mode and logical management method based on standard VLANs. This allows virtual machines to directly communicate with physical devices at Layer 2 through general-purpose access switches, thereby eliminating dependence on dedicated VXLAN switches, avoiding the operational complexity caused by differences in configurations of devices from multiple vendors, and significantly improving operational efficiency and networking flexibility through "one-time deployment of physical connections and flexible configuration of logical VLANs."

[0022] Key points of the invention Key point 1: It implements a hybrid traffic carrying and automated configuration method based on standard VLAN Trunk and Access modes, replacing the mandatory reliance on dedicated VXLAN switches in existing technologies.

[0023] Key point two: A logical mapping database and an automatic command generator adapted to multiple vendors were built, which solved the problem of manual configuration caused by the heterogeneity of configuration instructions for different switches.

[0024] Key Point 3: A method of "one-time physical connection deployment and flexible logical policy management" was proposed. Policies are generated through a graphical interface and automatically distributed to physical devices, realizing software-defined network management and greatly improving operational efficiency.

[0025] Effects of the present invention This invention achieves the following significant effects through a hybrid networking architecture based on standard VLANs and software-defined centralized logical management: It effectively avoids the performance loss and protocol compatibility limitations caused by traditional Layer 3 NAT forwarding, providing high-performance communication in native Layer 2 networks; it significantly reduces hardware costs and technical barriers by replacing dedicated VXLAN gateway devices with general-purpose VLAN switches; it completely eliminates the tedious manual configuration caused by command-line differences in heterogeneous devices by automatically adapting to the configuration generation and distribution of multi-vendor switches, achieving automation and intelligence in physical network configuration; and finally, through the innovative model of "one-time deployment of physical connections and flexible binding of logical policies," it greatly improves the deployment efficiency, maintenance convenience, and system scalability of virtual-physical interconnection. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of a Layer 2 access circuit. Figure 2 The overall process is the control flow that combines virtual and real elements. Figure 3 For the virtual-real integrated control process 2: offline operation; Figure 4 For the virtual-real integrated control process 3: Data entry; Figure 5 For the virtual-real integration control process 4: Draw the topology; Figure 6 For the virtual-real integration control process 5: Configure virtual-real integration. Detailed Implementation

[0027] To make the objectives, contents, and advantages of the present invention clearer, the specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples.

[0028] Example 1 This embodiment includes: Step 1: Establish Virtual Network Identifier and VLAN Mapping At the OpenStack compute node level, when a virtual machine's network data packet enters the data plane of the virtualization platform, a conversion mechanism between the virtual network identifier and the physical VLAN tag must first be established. This step designs an efficient tag mapping module at the br-int bridge, which maintains a mapping table between virtual network IDs and local VLAN IDs. The specific implementation process is as follows: when a raw data packet from a virtual machine is detected, the system queries the virtual network ID to which the virtual machine belongs, and then quickly locates the corresponding local VLAN tag through a hash algorithm.

[0029] A unique identifier for a virtual network, typically a UUID assigned by OpenStack Neutron; Lightweight hash function to ensure even distribution of virtual networks to VLAN IDs; Modulo operation ensures that the generated VLAN ID is within the standard VLAN range (1-4094); The calculated local VLAN ID is used for subsequent physical network transmission; Step 2: Construction and traffic scheduling of a unified physical transmission channel; To achieve efficient transmission of virtual network traffic over physical infrastructure, this invention designs an intelligent traffic scheduling mechanism; after the dedicated physical network card is configured in Trunk mode, a refined traffic classification and scheduling strategy needs to be established; the system will monitor the traffic scheduling of each VLAN in real time to ensure the network quality of critical services.

[0030] : Queue bandwidth allocated to the i-th VLAN; Total available bandwidth of the physical network interface card; The weight coefficient of the i-th VLAN is dynamically adjusted according to the service priority. : ; Step 3: Establishment and dynamic maintenance of logical mapping functions; The logical mapping function is the core component for implementing software-defined network management in this method. This function not only establishes the initial mapping relationship, but also needs to maintain the consistency of the mapping state in real time. The system adopts a distributed transaction mechanism to ensure strong consistency of the mapping relationship, ensuring that the mapping relationship between physical ports and virtual networks is accurate and reliable at any time.

[0031] t: timestamp, ensuring time consistency of the mapping relationship; Physical switch port identifier, including switch IP and port number; The 48-bit MAC address of the physical device; : The VLAN ID corresponding to this mapping relationship at time t; Step 4: Generation and optimization of abstract configuration strategies; The abstract configuration strategy generator adopts an intent-based network programming paradigm. The system first parses the user's business intent, and then generates the optimal configuration strategy through a strategy optimization algorithm. The optimization process considers multiple constraints such as network topology, device performance, and security policies to generate the optimal solution that satisfies all constraints.

[0032] Optimal configuration strategy; The execution cost function of strategy S, including time complexity and resource overhead; The set of constraints that the strategy must satisfy; Find the strategy parameters that minimize the cost function; Step 5: Compiling and optimizing instructions for multi-vendor devices; The instruction compiler adopts a multi-layer intermediate representation design architecture; first, it converts the abstract configuration strategy into a vendor-independent intermediate representation, then it passes through multiple rounds of optimization, and finally generates optimized device instructions based on the characteristics of the target device; the optimization process includes key technologies such as instruction merging, redundancy elimination, and parallelization.

[0033] Reduce high-level strategies to intermediate representations; Legalizing the target equipment ensures that instructions comply with equipment specifications; : Peephole optimization, performing local instruction optimization; The final generated sequence of instructions for the target device; Step 6: Atomic deployment and consistency verification of the configuration; The configuration distribution process employs a two-phase commit protocol to ensure atomicity; firstly, the configuration is pre-committed to all relevant devices, and then uniformly committed to take effect after all devices have successfully confirmed it; the verification phase uses formal verification methods, and model checks are used to ensure the correctness of the configuration.

[0034] AG: Globally always true, meaning it is always satisfied in all execution paths; AF: Eventually true, meaning it will definitely be satisfied at some point in the future; State: The actual operating status of the equipment; Expected: The expected state of the device.

[0035] In summary, the method involves the following: (1) Dual-mode network access integration: Based on the native Layer 3 NAT gateway access method of OpenStack, a Layer 2 L2 direct access solution was independently designed and implemented, providing users with the ability to flexibly choose high-performance Layer 2 interconnection or general Layer 3 interconnection according to business needs; (2) Hybrid bridging based on standard VLANs: The traffic of the virtual network is tagged with the local VLAN by the integrated bridge br-int on the computing node and is handled by a dedicated physical network card in Trunk mode, which transparently forwards the VLAN information of multiple virtual networks to the physical access switch; (3) Logical centralized management and control platform: Construct a management subsystem to abstract and input information on physical network resources and physical devices, and digitize and visualize physical connection relationships; the physical network resources include access switch IP and port number; (4) Software-defined automated configuration: The management platform automatically sends the VLAN configuration of the virtual network to the specified physical access switch port through standard protocols or SSH, replacing manual command line configuration and realizing the automation and intelligence of physical network configuration.

[0036] The standard protocols include SNMP and NETCONF.

[0037] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for interconnecting virtual and physical networks using multiple access methods, characterized in that, The method is developed based on OpenStack and improved upon it to achieve virtual-physical interconnection. Among them, the Layer 3 access uses the external network NAT scheme provided by OpenStack, while the Layer 2 access is designed and implemented independently. When each virtual machine's network interface card (NIC) is connected to the br-int virtual bridge, it is equivalent to being connected to a Layer 2 switch and configured with a VLAN. The br-int is then connected to a physical NIC, which is used solely for virtual-physical integration. The physical device is connected to the access switch, and this NIC is also connected to the access switch. The NIC is configured to communicate in Trunk mode. The port on the access switch corresponding to the physical device is then configured with the same VLAN as the virtual subnet, enabling communication. This approach, using a one-time physical device connection and logical management, avoids frequent adjustments to the physical connections of the devices, improving operability.

2. The virtual-physical interconnection method with multiple access methods as described in claim 1, characterized in that, The method includes: Step 1: Establish virtual network identifier and VLAN mapping; Step 2: Construction and traffic scheduling of a unified physical transmission channel; Step 3: Establishment and dynamic maintenance of logical mapping functions; Step 4: Generation and optimization of abstract configuration strategies; Step 5: Compiling and optimizing instructions for multi-vendor devices; Step 6: Atomic deployment and consistency verification of configuration.

3. The virtual-physical interconnection method with multiple access methods as described in claim 2, characterized in that, Step 1: Establishing virtual network identifier and VLAN mapping; At the OpenStack compute node level, when a virtual machine's network data packet enters the data plane of the virtualization platform, a conversion mechanism between the virtual network identifier and the physical VLAN tag must first be established. This step designs an efficient tag mapping module at the br-int bridge, which maintains a mapping table between virtual network IDs and local VLAN IDs. The specific implementation process is as follows: when a raw data packet from a virtual machine is detected, the system queries the virtual network ID to which the virtual machine belongs, and then quickly locates the corresponding local VLAN tag through a hash algorithm. A unique identifier for a virtual network, typically a UUID assigned by OpenStack Neutron; Lightweight hash function to ensure even distribution of virtual networks to VLAN IDs; Modulo operation ensures that the generated VLAN ID is within the standard VLAN range (1-4094); The calculated local VLAN ID is used for subsequent physical network transmission.

4. The virtual-physical interconnection method with multiple access methods as described in claim 3, characterized in that, Step 2: Construction and traffic scheduling of a unified physical transmission channel; To achieve efficient transmission of virtual network traffic over physical infrastructure, this invention designs an intelligent traffic scheduling mechanism; after the dedicated physical network card is configured in Trunk mode, a refined traffic classification and scheduling strategy needs to be established; the system will monitor the traffic scheduling of each VLAN in real time to ensure the network quality of critical services. : Queue bandwidth allocated to the i-th VLAN; Total available bandwidth of the physical network interface card; The weight coefficient of the i-th VLAN is dynamically adjusted according to the service priority. ： 。 5. The virtual-physical interconnection method with multiple access methods as described in claim 4, characterized in that, Step 3: Establishment and dynamic maintenance of logical mapping functions; The logical mapping function is the core component for implementing software-defined network management in this method. This function not only establishes the initial mapping relationship, but also needs to maintain the consistency of the mapping state in real time. The system adopts a distributed transaction mechanism to ensure strong consistency of the mapping relationship, ensuring that the mapping relationship between physical ports and virtual networks is accurate and reliable at any time. t: timestamp, ensuring time consistency of the mapping relationship; Physical switch port identifier, including switch IP and port number; The 48-bit MAC address of the physical device; : The VLAN ID corresponding to this mapping relationship at time t.

6. The virtual-physical interconnection method with multiple access methods as described in claim 5, characterized in that, Step 4: Generation and optimization of abstract configuration strategies; The abstract configuration strategy generator adopts an intent-based network programming paradigm; The system first parses the user's business intent, and then generates the optimal configuration policy through a policy optimization algorithm. The optimization process considers multiple constraints such as network topology, device performance, and security policies to generate the optimal solution that satisfies all constraints. Optimal configuration strategy; The execution cost function of strategy S, including time complexity and resource overhead; The set of constraints that the strategy must satisfy; Find the strategy parameters that minimize the cost function.

7. The virtual-physical interconnection method with multiple access methods as described in claim 6, characterized in that, Step 5: Compilation and optimization of instructions for multi-vendor devices; The instruction compiler adopts a multi-layer intermediate representation design architecture; first, it converts the abstract configuration strategy into a vendor-independent intermediate representation, then it passes through multiple rounds of optimization, and finally generates optimized device instructions based on the characteristics of the target device; the optimization process includes key technologies such as instruction merging, redundancy elimination, and parallelization. Reduce high-level strategies to intermediate representations; Legalizing the target equipment ensures that instructions comply with equipment specifications; : Peephole optimization, performing local instruction optimization; The final generated sequence of instructions for the target device.

8. The virtual-physical interconnection method with multiple access methods as described in claim 7, characterized in that, Step 6: Configure atomic distribution and consistency verification; The configuration distribution process employs a two-phase commit protocol to ensure atomicity; firstly, the configuration is pre-committed to all relevant devices, and then uniformly committed to take effect after all devices have successfully confirmed it; the verification phase uses formal verification methods, and model checks are used to ensure the correctness of the configuration. AG: Globally always true, meaning it is always satisfied in all execution paths; AF: Eventually true, meaning it will definitely be satisfied at some point in the future; State: The actual operating status of the equipment; Expected: The expected state of the device.

9. The virtual-physical interconnection method with multiple access methods as described in claim 8, characterized in that, The method involves the following: (1) Dual-mode network access integration: Based on the native Layer 3 NAT gateway access method of OpenStack, a Layer 2 L2 direct access solution was independently designed and implemented, providing users with the ability to flexibly choose high-performance Layer 2 interconnection or general Layer 3 interconnection according to business needs; (2) Hybrid bridging based on standard VLANs: The traffic of the virtual network is tagged with the local VLAN by the integrated bridge br-int on the computing node and is handled by a dedicated physical network card in Trunk mode, which transparently forwards the VLAN information of multiple virtual networks to the physical access switch; (3) Logical centralized management and control platform: Construct a management subsystem to abstract and input information on physical network resources and physical devices, and digitize and visualize physical connection relationships; the physical network resources include access switch IP and port number; (4) Software-defined automated configuration: The management platform automatically sends the VLAN configuration of the virtual network to the specified physical access switch port through standard protocols or SSH, replacing manual command line configuration and realizing the automation and intelligence of physical network configuration.

10. The virtual-physical interconnection method with multiple access methods as described in claim 9, characterized in that, The standard protocols include SNMP and NETCONF.

Images