Computing power network scheduling method, network domain node, cloud side node, computing power gateway and system

Abstract

The present application provides a computing power network scheduling method, a network domain node, a cloud side node, a computing power gateway and a system. The method comprises: generating a service identifier according to a service registration request (110); determining a computing power measure index according to a resource state of a service instance corresponding to the service identifier (120); issuing the service identifier and SLA information accessed by a corresponding user to a computing power gateway (130); and calculating a computing power route according to the computing power measure index, computing power information notified by the computing power gateway and the SLA information (140).

Description

Computing network scheduling method, network domain node, cloud side node, computing power gateway, and system Technical Field

[0001] The present application relates to the field of communication technology, for example, to a computing network scheduling method, network domain nodes, cloud-side nodes, computing power gateway, and system. Background Art

[0002] The development goal of computing networks is to flexibly meet the integrated needs of various businesses for computing resources and network connectivity. For different demand scenarios, the focus of integrated scheduling is usually different. For example, it may be experience-related, that is, service-level agreement (SLA) indicators related to quality of service (QOS); it may be cost-related, that is, the cost / energy consumption of service resources; it may be resource-related, that is, the usage or status of service resources. Taking edge computing as an example, services are usually scheduled to the nearest node, but not necessarily the optimal node. Especially for extremely sensitive businesses such as latency, service instances are often selected based solely on computing power, and the best option cannot be found. Due to the inability to achieve integrated computing power perception and computing network scheduling, different types of scheduling needs cannot be met, and equipment resource consumption and costs are high.

[0003] Summary of the Invention

[0004] The present application provides a computing network scheduling method, network domain nodes, cloud-side nodes, computing power gateway, system and medium.

[0005] The present application provides a computing network scheduling method, which is applied to a network node, including:

[0006] Generate a service identifier based on the service registration request;

[0007] Determine a computing power measurement indicator based on the resource status of the service instance corresponding to the service identifier;

[0008] Send the service ID and the corresponding user access SLA information to the computing power gateway;

[0009] The computing power route is calculated based on the computing power measurement indicator, the computing power information announced by the computing power gateway, and the SLA information.

[0010] The embodiment of the present application further provides a computing network scheduling method, which is applied to a cloud-side node, including:

[0011] Send service registration requests to the domain node based on the managed service type;

[0012] If the service registration is successful, the resource status of the service instance corresponding to the service identifier is collected;

[0013] The computing power information is notified to the computing power egress gateway (EGW) according to the resource status information, and the computing power information is diffused by the computing power egress gateway to the upstream gateway.

[0014] The present application also provides a computing network scheduling method, which is applied to a computing power ingress gateway (IGW), including:

[0015] Obtain the computing power information announced by the computing power export gateway, as well as the service identifier issued by the network domain node and the service level agreement (SLA) information of the corresponding user access;

[0016] Calculate a computing power route corresponding to a specific service identifier based on the computing power measurement indicator, the computing power information, and the SLA information;

[0017] The service request message is forwarded to the corresponding computing power export gateway according to the computing power routing.

[0018] The present application also provides a computing network scheduling method, which is applied to a computing power egress gateway, including:

[0019] Get the resource status of the service instance;

[0020] Determining a computing power measurement indicator and a local service route based on the resource status;

[0021] Install the virtual link and link attribute information corresponding to the virtual node, where the virtual link and link attribute information are obtained by the computing power egress gateway selecting a service instance for a specific service identifier based on the local service routing surface;

[0022] Spread the computing power information notified by the cloud side to the computing power entry gateway.

[0023] An embodiment of the present application further provides a network domain node, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned network scheduling method when executing the program.

[0024] An embodiment of the present application also provides a cloud-side node, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned computing network scheduling method when executing the program.

[0025] An embodiment of the present application also provides a computing power entry gateway, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned computing network scheduling method when executing the program.

[0026] An embodiment of the present application also provides a computing power export gateway, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned computing network scheduling method when executing the program.

[0027] The embodiment of the present application further provides a computing network scheduling system, comprising: a network domain node and the above-mentioned cloud-side node;

[0028] The network domain node includes the computing network brain, packet network equipment, the above-mentioned computing power entry gateway and the above-mentioned computing power exit gateway.

[0029] An embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the above-mentioned computing network scheduling method is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] FIG1 is a schematic diagram of a computing power network provided by an embodiment;

[0031] FIG2 is a flow chart of a computing network scheduling method provided by an embodiment;

[0032] FIG3 is a schematic diagram of an operation process of a computing network based on a service identifier provided by an embodiment;

[0033] FIG4 is a flow chart of another computing network scheduling method provided by an embodiment;

[0034] FIG5 is a flow chart of another computing network scheduling method provided by an embodiment;

[0035] FIG6 is a flowchart of another computing network scheduling method provided by an embodiment;

[0036] FIG7 is a schematic diagram illustrating an implementation of a computing power perception and computing network integrated scheduling system provided by an embodiment;

[0037] FIG8 is a schematic diagram of an implementation of a control plane provided by an embodiment;

[0038] FIG9 is a schematic diagram of a local service routing table provided by an embodiment;

[0039] FIG10 is a schematic diagram of a global service routing table provided by an embodiment;

[0040] FIG11 is a schematic diagram of an implementation of a forwarding plane provided by an embodiment;

[0041] FIG12 is a schematic diagram of the structure of a computing network scheduling device provided by an embodiment;

[0042] FIG13 is a schematic diagram of the structure of another computing network scheduling device provided by an embodiment;

[0043] FIG14 is a schematic diagram of the structure of another computing network scheduling device provided by an embodiment;

[0044] FIG15 is a schematic diagram of the structure of another computing network scheduling device provided by an embodiment;

[0045] FIG16 is a schematic diagram of the hardware structure of a network domain node provided by an embodiment;

[0046] FIG17 is a schematic diagram of the hardware structure of a cloud-side node provided by an embodiment;

[0047] FIG18 is a schematic diagram of the hardware structure of a computing power entry gateway provided by one embodiment;

[0048] FIG19 is a schematic diagram of the hardware structure of a computing power export provided by one embodiment;

[0049] FIG20 is a schematic diagram of the structure of a computing network scheduling system provided by an embodiment. DETAILED DESCRIPTION

[0050] The present application is described below in conjunction with the accompanying drawings and embodiments. It will be understood that the specific embodiments described herein are merely intended to explain the present application and are not intended to limit the present application. It should be noted that, unless there is a conflict, the embodiments and features within the embodiments of the present application may be combined with each other in any manner. It should also be noted that, for ease of description, the accompanying drawings only illustrate portions relevant to the present application, not all structures.

[0051] FIG1 is a schematic diagram of a computing power network provided by an embodiment. As shown in FIG1 , an operational computing power network generally includes an end side, a network side (also referred to as a network domain), and a cloud side. The network domain nodes mainly include a computing network brain, which can be a software defined network (SDN) controller. In the embodiment of the present application, the computing network brain can also be understood as a centralized management and orchestration node, which can be used for service registration, service-related information distribution, and path calculation in a centralized computing scenario. Nodes within the network domain can also include computing power gateways, specifically IGW and EGW. IGW can be used to maintain a global service routing table and perform path calculation in a distributed computing scenario. EGW can be used to announce computing power information, maintain a local service routing table, and select service instances. The process of sensing computing power can be achieved by the computing network brain directly connecting to the cloud management platform, or by the EGW directly connecting to the perception module deployed in the cloud resource pool to achieve computing power perception. These two perception methods are essentially the same for the integrated computing network scheduling and perception resource model. The subsequent embodiments mainly use the latter as an example to illustrate.

[0052] In order to realize the integrated scheduling of computing and network, the embodiments of the present application can solve the following problems: (1) to support centralized and distributed computing at the same time, and to flexibly switch between the two, it is necessary to design a scheduling method to realize the integration of computing power perception and computing network scheduling, and to complete the routing request and distribution of computing power; (2) the dimensions of computing domain quality (health score, average processing delay, economic cost) and network domain quality (bandwidth, delay, jitter, packet loss) are not unified, and it is often necessary to clarify a preferred factor based on the focus, and other factors are used as constraints. The increase of constraints increases the computing time and consumes the central processing unit (CPU). Unit, CPU) resources and cannot be deployed on a large scale; (3) From the perspective of IGW, different EGWs may be attached to the same type of service instance resources, and the same EGW may also be attached to multiple service instance resources of the same type. Usually, EGW needs to continuously update the computing power information to the upstream IGW. For example, if there are 100 services and each service is attached to 100 service instances, and the computing power information is refreshed every 10 seconds, the IGW needs to process at least 60,000 update messages per minute. In fact, the routing system itself requires a lot of resources to handle the IGP / BGP protocol maintenance and updates. The number and frequency of computing power information updates put huge pressure on the control plane of the network equipment and may even cause it to crash.

[0053] Figure 2 is a flowchart of a computing network scheduling method provided by one embodiment, which is applied to a network domain node. The network domain node mainly includes the computing network brain in the computing network, such as an SDN controller. Some steps can also be implemented by a computing power gateway. As shown in Figure 2, the method provided by this embodiment includes:

[0054] In step 110, a service identifier is generated according to the service registration request.

[0055] In step 120, a computing power measurement indicator is determined according to the resource status of the service instance corresponding to the service identifier.

[0056] In step 130, the service identifier and the SLA information corresponding to the user access are sent to the computing power gateway.

[0057] In step 140, a computing power route is calculated based on the computing power measurement indicator, the computing power information announced by the computing power gateway, and the SLA information.

[0058] In this embodiment, a service registration request can be initiated by a cloud-side node to the computing network brain, and the computing network brain can generate a corresponding service identifier for each type of service based on this, which can also be called a Service Awareness Network Identity (SAN-ID); the computing network brain can issue SAN-ID and corresponding SLA information to each computing power gateway based on the registered service, which is used for the conversion and mapping of service quality requirements within each computing power gateway. The computing power network supports centralized computing. For this application, the computing network brain can calculate the traffic engineering calculation path based on the computing power measurement index, the computing power information announced by the computing power gateway, and the SLA information; it also supports distributed computing, that is, the IGW can calculate the traffic engineering calculation path based on the computing power measurement index, the computing power information announced by the computing power gateway, and the SLA information.

[0059] In this embodiment, different types of services are distinguished by service identifiers, and similar services connected to each EGW can be abstracted into a virtual node, reducing the computing power routing problem to a network traffic engineering problem. When calculating computing power routing, on the one hand, the reachable path from the head node to the virtual node can be determined to clearly determine which EGW to reach, and on the other hand, the EGW can determine which specific service instance to reach. In the scenario of centralized computing, the reachable path (traffic engineering calculation path) can be determined by the computing network brain, and in the scenario of distributed computing, the reachable path can be determined by the IGW. In addition, the specific service instance can be determined by the EGW.

[0060] In this embodiment, the computing power measurement indicators mainly include indicators used to perceive or measure computing network resources, which correspond to the network; while the SLA information mainly corresponds to user access. The SLA information can be flexibly set to meet the multi-factor scheduling requirements of the computing network, thereby meeting the scheduling goals of experience, cost, and resource scenarios. By integrating computing power perception and computing network scheduling, equipment resource consumption and investment can be reduced, and the pressure on the control and forwarding planes brought about by the expansion of computing power information scale can be solved, with good feasibility and scalability. By uniformly sending the SAN-ID and mapped SLA information to the computing power gateway, a basis is provided for realizing computing network joint traffic engineering (Traffic Engineering, TE) and computing power routing optimization.

[0061] Computing power information can be obtained by the computing power perception module connected to the EGW cloud. Relevant information elements of computing power perception include but are not limited to SAN-ID, service instance IP, latency (cloud service latency), physical bandwidth (total cloud capacity), occupied bandwidth (cloud occupied capacity), and cost (cloud cost). Physical bandwidth and occupied bandwidth can be obtained by comprehensively scoring relevant cloud-side indicators (such as CPU utilization and storage resource usage).

[0062] In this embodiment, computing network resources are jointly modeled and perceived based on the SLA information corresponding to the SAN-ID. This transforms the computing network joint TE problem into finding a path from the head node to the virtual node of the service instance that meets the SLA requirements. Depending on the TE configuration, the Path Computation Element Communication Protocol (PCEP) can be used to initiate centralized computing requests (applicable to cross-domain scenarios) or distributed head node IGW computing requests (applicable to single-domain rapid convergence scenarios).

[0063] In one embodiment, the service registration request includes a service domain name, a service IP and port, SLA information, and resource information.

[0064] In one embodiment, the metrics of the SLA information include: delay, cost, bandwidth, jitter and reliability; the computing power metrics include network delay converted from calculation delay, network cost converted from resource cost, interface bandwidth converted from total resource amount, and interface bandwidth occupancy converted from resource occupancy.

[0065] In this embodiment, SLA information is unified into five key metrics: end-to-end access latency, cost, available bandwidth, jitter, and reliability. Depending on the scheduling objectives, SLA metrics can include experience metrics, such as latency, jitter, and reliability; cost metrics, such as cost or energy consumption (i.e., cost); and resource metrics, such as resource usage or status, for example, converting remaining cloud-side resources into available bandwidth. Furthermore, computing latency, resource cost, total resource amount, and resource occupancy are converted into network latency, network cost, interface bandwidth, and interface bandwidth occupancy, respectively, to obtain computing power metrics. The computing power perception results can be converted into network factors, achieving computing-network integration.

[0066] In one embodiment, computing power routing includes global service routing and local service routing; computing power routing is calculated based on computing power measurement indicators, computing power information announced by the computing power gateway, and SLA information, including: determining the traffic engineering calculation path from the computing power entry gateway to the virtual node mapped to the specific service identifier based on the computing power measurement indicators, computing power information announced by the computing power gateway, and SLA information; matching the traffic engineering calculation path with the global service routing table to obtain the global service routing; determining the service instance corresponding to the specific service identifier based on the resource status of the service instance to obtain the local service routing; wherein, the second to last node of the traffic engineering calculation path corresponds to the computing power exit gateway IP, and the end node of the traffic engineering calculation path corresponds to the virtual node mapped to the service identifier.

[0067] In this embodiment, a specific service identifier can be understood as a service identifier for which a computing power route needs to be calculated, such as the service identifier corresponding to the current service request message. Based on the computing power metrics, network TE metrics, and SLA information corresponding to the SAN-ID, the computing network brain or IGW head node can first determine the traffic engineering calculation path from the computing power entry gateway to the virtual node mapped to the service identifier, obtain the reachable path from the IGW to the virtual node S-VIP connected to the EGW, and then match it with the global service routing table to obtain the global service route. The matching principle is that the penultimate node in the traffic engineering calculation path is the EGW gateway selected for the corresponding SAN-ID, and the last node is the virtual node mapped to the SAN-ID. The global service routing table can be generated by the IGW after the computing power information is propagated to the IGW. In addition, the EGW can determine the service instance corresponding to the specific service identifier based on the resource status of the service instance to obtain the local service route. On this basis, the calculation of the service route and the computing network path can be separated, enabling the arbitrary deployment of computing resources on public and private networks.

[0068] In one embodiment, the virtual link and link attribute information corresponding to the virtual node are obtained by the computing power export gateway selecting the service instance to the specific service identifier based on the local service routing surface; wherein, the local service routing table includes the virtual routing forwarding identifier, service identifier, service instance IP and corresponding computing power measurement information; the virtual link and link attribute information are disseminated by the Interior Gateway Protocol (IGP).

[0069] In this embodiment, the EGW can maintain relevant information of a specific service identifier through a local service routing table, and can optimize the relevant service instances, and decide whether to notify or revoke computing power information to the upstream based on the existence of the optimization result.

[0070] In one embodiment, the global service route includes: virtual private network information, service identifier, service color, computing power export gateway IP, computing power export gateway segment identifier, virtual private network segment identifier; the global service route is formed by spreading computing power information from the computing power export gateway to the computing power import gateway; the bearer protocol of the global service route information is based on the BGP protocol family extension.

[0071] In this embodiment, the same type of service can be abstracted into a virtual node in a virtual private network. The information elements of computing power routing include but are not limited to virtual private network (VPN) information, SAN-ID, service color (COLOR), computing power export gateway IP (GW-IP local node IP), computing power export gateway segment identifier (GW-SID local node SID), and the corresponding segment identifier of the virtual private network in virtual routing forwarding (VPN-SID, local node identifier VRF SID). The information carrying protocol of computing power routing is implemented using the Border Gateway Protocol (BGP) family extension, such as the BGP multi-protocol extension (MultiProtocol Border Gateway Protocol, MP-BGP).

[0072] In one embodiment, the computing power information does not include information about the service instance and preferred entry corresponding to the specific service identifier; the computing power information is decided by the computing power export gateway whether to announce and revoke it based on whether the corresponding preferred entry exists locally; the computing power information uses the virtual routing forwarding identifier and service identifier as key indexes; the computing power information is diffused from the computing power export gateway to the computing power import gateway, and is uniformly processed by the computing power import gateway.

[0073] In this embodiment, a local (or partial) service routing table (SRIB) can be constructed in the EGW. As long as there is a preferred entry for the local SRIB facing the SAN-ID, VPN-related computing power information can be formed and disseminated. In principle, the EGW can process the computing power information perceived by the cloud side to obtain the preferred entry for the computing power instance. For example, after receiving the computing power information, the EGW uses the VRF-ID and SAN-ID as key indexes based on the VPN information deployed by the computing power resources, and combines the load calculation latency, available bandwidth, and cost of the service instance IP to maintain the corresponding entry in the SRIB. Local optimization is implemented for each entry based on the requirements of the SLA information corresponding to the SAN-ID. For example, for a certain type of service, if the SLA information indicates a focus on latency, the EGW can locally optimize each entry based on the latency of the service instance. The obtained preferred entry can be sent to the forwarding plane as the service forwarding information base (SFIB) to guide the forwarding of service request messages. By establishing a two-level service routing model and enabling EGW to flexibly select service instances locally, the pressure on the forwarding and control planes can be reduced.

[0074] In one embodiment, the announced computing power information may not be updated as the optimization result changes, thereby effectively reducing the pressure on the control plane. In addition, the announced computing power information is only revoked when the optimization entry corresponding to a specific SAN-ID is zero.

[0075] In one embodiment, each SAN-ID is planned with a public network service virtual node IP (abbreviated as S-VIP), a service virtual node segment identifier (abbreviated as S-NODE-SID) and a service COLOR, which has a globally unique and one-to-one mapping relationship. Based on the local SRIB, the EGW can, for the preferred entry corresponding to a specific SAN-ID and the public network service virtual node IP (S-VIP) corresponding to the SAN-ID, use the latency, bandwidth and cost of the computing power corresponding to the virtual node and the preferred entry as the link attributes connecting the EGW and the S-VIP node, and install them together in the Traffic Engineering Database (TEDB), and flood and spread them within the IGP domain. In other words, TEDB installation and IGP diffusion can be performed on the basis of implementing local optimization, thereby effectively reducing the scale of control plane information diffusion. In some embodiments, the EGW can perform TEDB installation and IGP diffusion directly based on the local SRIB without performing local optimization. By abstracting the computing power information of network resources into a computing power measurement indicator Metric, the virtual nodes and links are installed in the IGP, so that the computing network integrated path calculation is transformed into a network problem. In addition, since an independent service routing layer is established, the EGW does not need to publish routing information containing service instance network segments to the IGW.

[0076] In one embodiment, in a distributed computing scenario, based on the TE configuration of the IGW, the IGW's TE calculation component can calculate the path from the IGW to the virtual node S-VIP connected to the EGW based on the SLA information corresponding to the SAN-ID, using the local TEDB. This calculation generates the segment routing policy (SR-POLICY) information corresponding to the TE, with the key values ​​(KEY) including COLOR and S-VIP. A segment list (SEGMENT LIST) is formed by the SIDs of each node and link along the path, where the second-to-last SID represents the EGW gateway (EGW-SID) selected by the corresponding SAN-ID, and the last SID represents the virtual node.

[0077] In one embodiment, the IGW can receive computing power routes from different EGWs and form SRIBs to different EGWs. The elements of the SRIBs include but are not limited to VRF-ID, SAN-ID, EGW-IP, COLOR, EGW-SID, VPN-SID, etc., where VRF-ID and SAN-ID serve as keys. In the case of the same key, entries related to different EGWs (including EGW-IP / EGW-SID / VPNSID) can be managed as multiple next hops. Based on the COLOR and S-VIP of each entry in the SRIB, each SR-POLICY entry is matched. If the EGW-SID in the SRIB is equal to the second-to-last SID in the SR-POLICY segment list, the SRIB entry can be set as the preferred one, that is, a next hop and a clear reachable path are selected in the global service routing table, and a global SFIB can be generated and sent to the forwarding plane to guide the forwarding of service request messages.

[0078] FIG3 is a schematic diagram of a computing network operation process based on a service identifier according to an embodiment. As shown in FIG3 , the computing network operation process includes:

[0079] 1) Service Registration: The cloud-side operation and management system registers with the Computing Network Brain based on the type of service being managed. Registration elements include the service domain name, service IP address and port, and the corresponding SLA information for the service. SLA information includes, but is not limited to, latency, bandwidth (also known as available bandwidth), reliability (such as packet loss), jitter, cost (also known as expense, COST, or cost), resource utilization (link occupancy or cloud resource pool occupancy), and any combination thereof. This process enables the Computing Network Brain to generate a service identifier (SAN-ID).

[0080] 2) SAN-ID and SLA issuance: The computing network brain issues SAN-ID and corresponding SLA information to each computing power gateway based on the registered services, which is used for conversion and mapping of service quality requirements within the gateway;

[0081] 3) Start service awareness: After receiving the response message of successful service registration, the cloud side can start the service awareness module to collect cloud side resource status information;

[0082] 4) Computing power status information collection and notification: The service perception module notifies the computing power information to the EGW (BGP can be used), the EGW disseminates the computing power information upstream (local optimization can be implemented and the preferred items can be disseminated), and the IGW or the EGW acts as a BGP link state (BGP-LS) announcer to notify the computing power information to the computing network brain;

[0083] 5) Centralized path calculation request: In scenarios involving cross-domain or IGW resource limitations that are not suitable for or support distributed head node calculation, the IGW can initiate PCEP requests for the network brain to centrally calculate the traffic engineering calculation path to the virtual node. For distributed head node calculation scenarios, the IGW can calculate the traffic engineering calculation path, thus unifying centralized and distributed path calculation.

[0084] 6) Path calculation results are sent: For centralized computing scenarios, after the network brain completes the traffic engineering path calculation, it can use the PCEP channel or BGP-SR-POLICY channel to send the calculation results to the IGW; for distributed scenarios, the IGW can locally complete the traffic engineering path calculation and send it.

[0085] 7) Terminal service identification acquisition: The terminal initiates domain name resolution based on the original service domain name to obtain the service identification for service initiation;

[0086] 8) Service initiation: After the terminal obtains the service identifier, the terminal protocol stack can select different encapsulation methods to send the service request message according to the service identifier. For example, if anycast IP is used to represent the service, the service identifier can be directly placed in the destination IP. If a digital service identifier is used, it is often necessary to carry the service identifier through an IP extension header or other extension methods. The destination IP of the service request message is the IGW IP. Regardless of the encapsulation method, the purpose is to allow the service request message to penetrate the access network, be routed to the IGW, and resolve the service identifier to achieve service access.

[0087] 9) Traffic Scheduling: After the service request message reaches the IGW, the IGW parses the service identifier and searches for computing power routing, selects different tunnel strategies, and encapsulates the tunnel header message for hop-by-hop forwarding. After reaching the EGW, the EGW can find the corresponding service instance based on the service identifier and forward the message to the service instance node to provide the service.

[0088] It should be noted that the embodiments of the present application are mainly described based on a network based on source routing technology of IPv6 (Segment Routing Over IPv6, SRv6), and in some embodiments, it is also applicable to a network based on segment routing (SR) multi-protocol label switching (MPLS). For the computing power perception component, in order to simplify the description, although this solution takes the example of docking the cloud-side computing power perception module to obtain computing power, in actual deployment, the computing network brain can dock with the cloud-side operation and management system and notify the computing power gateway or the computing power perception component of each device in the computing power network, and the preferred entry can be notified after the local optimization is implemented to reduce the pressure on the control surface and the scale of information diffusion.

[0089] The computing network scheduling method described in the above embodiment addresses the issue of multiple factors influencing the efficiency of scheduling algorithms by unifying computing power-related factors (such as experience, cost, and resources) into the same dimensional model as network factors, reducing the number of factors. In specific scenarios, such as those focusing on cost, computing network joint calculation is generally not required; only computing power service instances need to be optimized, and network factors are relatively unimportant. By increasing the range of the original modeled computing power COST (so that the COST range exceeds the maximum sum of the computing power network), computing power-dominated optimization can be achieved, resulting in strong adaptability. By abstracting computing power services into virtual nodes and corresponding computing power instances into links connecting the EGW and the virtual nodes, the computing power instance status is used as a virtual link attribute, and computing power instance optimization is performed through the EGW. This significantly reduces the number of links connected to the virtual nodes. Virtual node and link information is directly injected into the TEDB and disseminated using the IGP, avoiding the protocol interaction mechanism between virtual nodes and normal nodes. This normalizes computing network joint calculation to the conventional network TE problem, and the calculation method is fully compatible with the network distributed routing and centralized controller routing of traditional network TE paths. By designing a computing power routing layer independent of the IP routing connection layer and a two-level computing power routing model, VPN deployment is supported. The IGW directly searches the computing power forwarding table based on the computing power ID, encapsulates the tunnel to the EGW, and then selects the computing power instance. This avoids the need to extend other traffic diversion methods based on computing power IDs. The EGW detects the existence of computing power services and instances and decides whether to publish or revoke routes. Computing power routes are published upstream with the EGW as the next hop, significantly reducing the control plane burden. Furthermore, the EGW does not need to publish IP routing information to the computing power instance to the IGW, simplifying routing deployment configuration and enhancing the security and scalability of the entire computing power routing domain. Based on the TE routing results and VPN computing power routes, the IGW uses COLOR and the service virtual node IP address for correlation lookup. Qualified entries are selected by matching the corresponding EGW-SID with the penultimate SID in the Segmentlist in the SR-POLICY. This transforms the joint computing and routing problem of the computing network into a conventional routing domain problem, enabling flexible VPN deployment of computing power resources.

[0090] The above-mentioned computing network scheduling method has strong feasibility, adaptability to different scenarios and scalability, which can accelerate the rapid implementation of computing power networks in existing network environments. Specifically: The computing power information is transparent to the intermediate nodes of the network, and there is no need for equipment upgrades and configuration adjustments, which is easy to deploy on the existing network; the centralized and distributed computing network integrated computing and the centralized and distributed TE path computing of the traditional network are normalized, and the computing network brain and network controller no longer need to add other computing power perception channels, which is highly feasible; the normalized modeling of computing power factors (experience, cost and resources, etc.) and network factors avoids the problems of low efficiency and complexity caused by the joint computing of the computing network, and supports the conversion of complex SLA requirements of the computing network into multiple TE algorithm extensions; the computing power service identification scheme and the control plane are completely decoupled. Different identification schemes only affect the packaging and parsing of the computing power gateway identification, and the control plane software system is highly reusable; while keeping the SR-TE technical framework unchanged, Metric can continue to expand the constraints in the future (this case takes latency as the target), and according to the TE configuration, PCEP can be used to initiate controller centralized computing (cross-domain scenarios) and head node computing (single-domain fast convergence scenarios).

[0091] It should be noted that for distributed computing, the IGW can calculate the traffic engineering calculation path from the IGW to the virtual node. For centralized computing, the computing network brain can centrally generate the corresponding SR-POLICY and send it to the IGW. Other subsequent computing power routing optimization and installation processes are completely decoupled from the centralized and distributed steps.

[0092] FIG4 is a flowchart of a computing network scheduling method provided by an embodiment, which is applied to cloud-side nodes. For technical details not fully described in this embodiment, please refer to any of the above embodiments. As shown in FIG4, the method provided by this embodiment includes:

[0093] In step 210, a service registration request is sent to the domain node according to the managed service type.

[0094] In step 220, if the service registration is successful, the resource status of the service instance corresponding to the service identifier is collected.

[0095] In step 230, computing power information is notified to the computing power egress gateway based on the resource status information, and the computing power information is propagated by the computing power egress gateway to the upstream gateway. In addition, traffic can be directed to the service instance by the computing power ingress gateway and the computing power egress gateway.

[0096] Figure 5 is a flow chart of a computing network scheduling method provided by an embodiment, which is applied to a computing power entry gateway. For technical details not fully described in this embodiment, please refer to any of the above embodiments. As shown in Figure 5, the method provided by this embodiment includes:

[0097] In step 310, the computing power information announced by the computing power export gateway, as well as the service identifier issued by the network domain node and the service level agreement SLA information of the corresponding user access are obtained.

[0098] In step 320, a computing power route corresponding to a specific service identifier is calculated based on the computing power measurement indicator, the computing power information, and the SLA information.

[0099] In step 330, the service request message is forwarded to the corresponding computing power export gateway according to the computing power routing.

[0100] In one embodiment, it further includes:

[0101] Maintaining a global service routing table having multiple next hops for the specific service identifier according to the received computing power information of the multiple computing power export gateways;

[0102] A traffic engineering model is constructed according to the SLA information corresponding to the specific service identifier and the corresponding virtual node, wherein the traffic engineering model includes color information and the virtual node IP mapped to the service identifier.

[0103] In one embodiment, the computing power routing includes global service routing;

[0104] Calculating a computing power route corresponding to a specific service identifier according to the computing power measurement indicator, the computing power information, and the SLA information includes:

[0105] Determine a traffic engineering calculation path from a computing power entry gateway to a virtual node mapped to a service identifier based on the computing power measurement indicator, the computing power information, and the SLA information;

[0106] Matching the traffic engineering calculation path with the global service routing table to obtain the global service routing;

[0107] Among them, the penultimate node of the traffic engineering calculation path corresponds to the computing power exit gateway, and the last node of the traffic engineering calculation path corresponds to the virtual node mapped by the service identifier.

[0108] In one embodiment,

[0109] The computing power information does not include the service instance and preferred item information corresponding to the specific service identifier;

[0110] The computing power export gateway decides whether to announce or revoke the computing power information based on whether there is a corresponding preferred entry locally;

[0111] The computing power information uses the virtual routing forwarding identifier and service identifier as key indexes;

[0112] The computing power information is directly spread from the computing power export gateway to the computing power import gateway and is uniformly processed by the computing power import gateway.

[0113] In one embodiment, for distributed computing, the IGW may execute the process of computing traffic engineering paths and matching with the global service routing table described in any of the above embodiments.

[0114] In one embodiment, the computing power information includes a preferred entry, which is determined by the computing power export gateway based on the computing power information announced by the cloud side, the SLA information corresponding to the service identifier, the virtual private network VPN information, and the load of the service IP, and uses the virtual routing forwarding identifier and the service identifier as key indexes.

[0115] Figure 6 is a flow chart of a computing network scheduling method provided by an embodiment, which is applied to a computing power export gateway. For technical details not fully described in this embodiment, please refer to any of the above embodiments. As shown in Figure 6, the method provided by this embodiment includes:

[0116] In step 410, the resource status of the service instance is obtained;

[0117] In step 420, a computing power measurement indicator and a local service route are determined based on the resource status.

[0118] In step 430, the virtual link and link attribute information corresponding to the virtual node are installed. The virtual link and link attribute information are obtained by the computing power export gateway by selecting the service instance for the specific service identifier according to the local service routing surface.

[0119] In step 440, the computing power information notified by the cloud side is spread to the computing power entry gateway.

[0120] In one embodiment, the computing power measurement index includes network delay converted from computing delay, network cost converted from resource cost, interface bandwidth converted from total resource amount, and interface bandwidth occupancy converted from resource occupancy;

[0121] The method further comprises:

[0122] Generate a local service routing table, wherein the local service routing table includes a virtual routing forwarding (VRF) identifier, a service identifier, a service instance identifier, and corresponding computing power measurement information.

[0123] In one embodiment, installing the virtual link and link attribute information corresponding to the virtual node includes:

[0124] According to the local service routing, the IP corresponding to the service identifier is used as the virtual node, and the computing power measurement information of the corresponding entry is used as the link attribute of the virtual link to be flooded within the domain by the IGP, thereby realizing the diffusion of the virtual link and link attribute information within the domain.

[0125] In one embodiment, the computing power information does not include information about the service instance and preferred entry corresponding to the specific service identifier;

[0126] The computing power export gateway decides whether to announce or revoke the computing power information based on whether there is a corresponding preferred entry locally;

[0127] The computing power information uses the virtual routing forwarding identifier and service identifier as key indexes;

[0128] The computing power information is spread from the computing power export gateway to the computing power gateway entrance and is uniformly processed by the computing power entrance gateway.

[0129] In one embodiment, it further includes:

[0130] The local service routing table is searched according to the service identifier in the service request message to determine the corresponding service instance, and the message is forwarded to the corresponding service instance.

[0131] Figure 7 is a schematic diagram of an implementation of a computing power perception and computing network integrated scheduling system provided by one embodiment. This system can be used to implement computing network joint computing and computing power addressing with normalization of different computing and network factors in a computing power network. As shown in Figure 7, the relevant components and descriptions of the overall technical architecture are as follows:

[0132] Cloud-side computing power perception module: This module can be added to the cloud-side monitoring system. Based on a full understanding of computing power scheduling requirements, it can aggregate and model the original computing power monitoring indicators to form a small number of comprehensive indicators that can stably reflect the computing power quality level. It serves as a speaker for computing power status. For related notification information, see the X1 interface.

[0133] Computing power perception component: This component can be added to the network device (EGW) and connected (X1 interface) to the cloud-side computing power perception module. It can serve as the first consumer of computing power information and simultaneously connect (X2 interface) to the computing power routing management component. By configuring relevant policies, information oscillation suppression and processing can be achieved. After processing, qualified computing power information is then injected into the computing power routing management component.

[0134] Computing power routing management component: This component can be added to the network device and realize the following functions according to the received computing power status information: (1) Generate local computing power routing table SRIB, and perform local optimization according to the optimization strategy (experience, cost and resources, etc.) associated with SAN-ID; (2) Inject the optimized local computing power routing table (X4 interface) into the BGP protocol component for external publication, and at the same time, as EGW, also receive the computing power routing published by the BGP component from the remote BGP node to form a global computing power routing table; (3) Based on the local optimal computing power routing table for each service, form each virtual computing power node and abstract link and link parameters and inject them into TEDB (X3 interface); (4) As an IGW node, receive (interface X6) the SR-TE component based on the northbound configuration (the policy path from this node to the virtual computing power node that meets the COLOR requirements), and generate the final global computing power forwarding table according to the VPN computing power routing iteration; (5) Connect to the message forwarding component and send down (X7 interface) the computing power forwarding table. Among them, points (1)-(4) are the key points of this application;

[0135] Message forwarding component: Expands on network devices, receives and installs the local / global computing power forwarding table issued by the computing power routing management component (X7 interface), and implements traffic scheduling and forwarding of terminal computing power service requests;

[0136] BGP protocol component: Expands on network devices to enable VPN computing power routing local upload / download (X4 interface) and inter-node (X5 node) diffusion;

[0137] SR-TE calculation component: The computing power routing management component subscribes to the relevant calculation result table items and does not modify the external interface.

[0138] The key interfaces and description examples included in the overall technical architecture are as follows:

[0139] X1 and X2 interfaces: In the interface model, X1 and X2 differ only in the notification frequency, bearer protocol, and channel. Information types include add, update, and delete. Model elements include but are not limited to service ID, service instance IP, latency (cloud instance processing latency), cost (cloud instance usage cost), physical bandwidth (cloud instance available resources), and occupied bandwidth (cloud instance occupied resources).

[0140] X3 interface: Virtual node elements include but are not limited to service ID, virtual node IP, and virtual node NODE SID. Virtual link attributes include but are not limited to information such as latency, cost, link bandwidth, and link occupied bandwidth. These link attributes represent the status of the preferred local computing instance and are refreshed as the preferred local computing instance changes or its status is updated.

[0141] X4 interface: includes EGW node upstream and IGW node downstream, and its elements include but are not limited to VRF-ID, SAN-ID, COLOR, EGW-IP, and EGW-SID;

[0142] X5 interface: uses MP-BGP extensions, including but not limited to VPN information (Route Target / Route Distinguisher, RT / RD), SAN-ID, COLOR, EGW-IP, EGW-SID, and VPN-SID.

[0143] X6 interface: Delivers SR-POLICY information. Indexes typically include Color and S-VIP. Attributes include at least the segment list {SID1, SID2...EGW-SID, S-NODE-SID}, the directly connected outbound interface, and the gateway IP address.

[0144] X7 interface: Local computing power routing table elements include but are not limited to: VRF-ID, SAN-ID, INS-IP (INSTANCE-IP). Global computing power routing table elements include but are not limited to: VRF-ID, SAN-ID, VPN-SID, OIF:SR-POLICY (COLOR, EGW-IP), Segment List {SID1, SID2...EGW-SID}, directly connected L3 outbound interface, and corresponding gateway IP.

[0145] The following briefly describes the workflow of traffic steering in the computing power network:

[0146] (1) The client initiates a computing service request. The service request message carries the SID. There are multiple ways to carry the SID (ANYCAST address as SID, IP extension header, etc.). Regardless of the carrying method, the goal is to ensure that the service request message reaches the IGW and the IGW perceives the SID.

[0147] (2) The IGW receives the service request message from the client, identifies the corresponding SID and selects the corresponding EGW, and issues a specific network path to meet the network quality requirements required for service access.

[0148] (3) The EGW receives the service request forwarded by the IGW, identifies the corresponding SID and selects the appropriate service instance IP. It also modifies the destination address of the service request message to the service instance IP, checks the IP routing table, and forwards the service request message to the service instance to achieve service connection.

[0149] (4) The service instance responds to the service request message and modifies the source IP address of the message to the destination IP address corresponding to the service request message on the EGW. The subsequent steps are part of the normal network service process.

[0150] For the control plane, to achieve compute-aware traffic steering, the conventional design approach for network equipment control planes is to use IGP to flood network link attributes and IGP / BGP extensions to implement compute awareness and notify upstream nodes, forming a traffic engineering database (TEDB) and a compute-aware database (CADB). This requires cross-layer association of the CADB and TEDB, and joint computation (centralized or distributed) according to service access SLAs to obtain service instances and network paths that meet the requirements. This leads to two problems: 1) Due to different computing speed requirements, centralized and distributed computing must be supported simultaneously, resulting in the repeated design of an SDN architecture similar to that based on PCEP. 2) The dimensions of computing domain quality (health score, average processing delay, economic cost, and resource utilization) and network domain quality (bandwidth, latency, jitter, and packet loss) are not unified. Computation time increases with increasing constraints, consuming a large amount of CPU resources and preventing large-scale deployment. In addition, computing instances of the same service type can be flexibly deployed to the same EGW and / or different EGWs. If the EGW computing resource status is continuously updated to the upstream IGW, the massive computing status information update will increase huge pressure on the network device control plane and even cause the system to crash.

[0151] FIG8 is a schematic diagram of a control plane implementation provided by an embodiment. In this embodiment, traffic can be redirected to SLA targets to unify the five key metrics of end-to-end access delay, cost, bandwidth, jitter, and packet loss, depending on the target: 1) Experience category, which often focuses on delay, jitter, and packet loss; 2) Cost category, which often focuses on cost / energy consumption, that is, cost; 3) Resource category, which focuses on resource usage / status. If the remaining cloud resources are converted into available bandwidth, then the focus is on available bandwidth. In actual business deployment, one of the five metrics is often selected according to different targets, and the other metrics are selected together as constraints. Exemplarily, as shown in FIG8 , taking distributed computing as an example, the control plane process mainly includes:

[0152] S1. Register, upgrade, or cancel services based on the service type and the service instance to which each service is attached. Furthermore, computing power information may be obtained from a computing resource pool or an operation management system. This embodiment does not limit the method for obtaining computing power information.

[0153] S2. EGW computing load updates. Registered service instances can be represented by IP addresses. By recording computing metrics, the acquired computing power information is converted into a dimension consistent with the network.

[0154] Based on the computing power information obtained from S2, S3 and EGW generate a local service routing table containing the VRF identifier, service identifier, service instance IP address, and computing metric. This table can then be further optimized locally, for example based on latency. The optimized entries are then propagated upstream to the head node or the computing network brain for computing power routing.

[0155] S4. The EGW forms virtual nodes (representing the preferred services currently connected to the EGW) and links based on the preferred service instances (the preferred links, whose metrics serve as the metrics for the preferred computing resources). This significantly reduces the number of links and improves network computing efficiency. It should be noted that it is not necessary to perform optimization; the number of links is determined by the number of service instances connected to the EGW. In this step, the computing power information of the computing resources can be converted into the TE attributes of the links.

[0156] S5. Based on IGP, the TE link attributes and virtual node information are spread and flooded to upstream nodes.

[0157] S6. After diffusion, the IGW has global link information.

[0158] S7. The computing network joint scheduling problem is transformed into a network path problem. The IGW can calculate a path from the IGW to the virtual node corresponding to the service identifier.

[0159] S8, EGW01, and EGW02 publish MP-BGP VPN service routes. It should be noted that S9 can be performed simultaneously with S5 / S6.

[0160] S9. Generate a global service routing table to form a global service route with multiple EGW next hops. For example, for each service, if each EGW has two preferred entries, a total of four global service routes can be formed.

[0161] S10: Combine the result of S7 (EGW IP of the penultimate hop) and the routing result of S9 (whether the penultimate hop is EGW01 or EGW02) to determine whether to forward to EGW01 or EGW02, and obtain the final forwarding table, i.e., the global service route.

[0162] On this basis, EGW01 or EGW02 can further determine the service instance.

[0163] Figure 9 is a schematic diagram of a local service routing table provided by one embodiment. The EGW perceives the computing instance status from the Edge Manager and, based on the VPN information deployed on the computing resource, uses the VRF-ID and SID as indexes, combined with the service instance IP address and the latency, bandwidth, and cost factors of the computing resource, maintains the corresponding service instance entry in the SRT (see Figure 9). Local processing is performed based on the SLA constraints corresponding to the SID (e.g., if the service SLA focuses on latency, local priority is based on instance latency), forms a local SRT entry preference, and sends the preferred entry as an SFT (see Figure 10) to the forwarding plane for computing service request processing and forwarding.

[0164] When a preferred service instance exists for a specific SID in the local SRT, the EGW issues a routing update message to the IGW. Once the resource status deteriorates and the preferred service instance becomes zero, a routing withdrawal message is issued to the IGW. The bearer protocol is usually implemented using the BGP protocol family extension (MP-BGP), carrying elements such as message type, SID, EGW-IP, VPN-SID, RT / RD, etc. Service routes are issued to the IGW instead of specific service instance information, thus forming a service routing layer independent of VPN IP routing, reducing the pressure on the control plane.

[0165] The EGW selects the entry for each SID based on the local SRT and installs the virtual node and virtual link corresponding to the SID into the IGP. The virtual node is connected to the EGW through a virtual link. The latency, bandwidth, and cost of the optimized service instance are used as the link attributes of this virtual link. The network metrics are flooded within the IGP domain, greatly reducing the scale of control plane information dissemination.

[0166] Figure 10 is a schematic diagram of a global service routing table provided by an embodiment. As shown in Figure 10, the requirement for optimal traffic redirection for accessing the computing service SID on the IGW is transformed into a traditional network traffic engineering problem: path calculation is performed between the virtual nodes corresponding to the SIDs connected to the IGW and EGW according to the SLA constraints corresponding to the service SIDs, generating the corresponding SR-POLICY-1 path (COLO, Endpoint: SID). The penultimate Segment ID (NODE) ​​in the Segment List represents the preferred EGW for service access under the current conditions. This SR-POLICY-1 is then converted into a SR-POLICY-2 path (COLOR, Endpoint: EGW-IP) that meets the requirements.

[0167] The IGW receives VPN service routing information published by each EGW and forms a global SRT to multiple EGWs, using the VRF ID and SID as the key and different EGW IP addresses as multiple next hops. Based on the color and EGW IP address of each SRT entry, it matches the SR-POLICY-2 entry, obtains the optimal global SRT entry, and generates a global SFT (see Figure 7), which is then sent to the IGW forwarding plane for traffic steering requests.

[0168] Figure 11 is a schematic diagram of an implementation of a forwarding plane provided by an embodiment. As shown in Figure 11, the main process of the forwarding plane is as follows:

[0169] S1: The terminal initiates a computing power service request. The SAN-ID can be carried in various ways (ANYCAST address as SAN-ID, carried in an extension header, etc.). This embodiment of the application does not limit this. Regardless of the carrying method, the purpose is to ensure that the request message reaches the IGW and the IGW perceives the SAN-ID. In this example, the UE initiates a computing power service request with SAN-ID = SID1.

[0170] S2: IGW receives the service request message from S1 and searches the VPN computing power SFIB according to the SAN-ID SID1 carried in the message: {vidx, SID1, vidx-EGW01-SID, py-SID1-EGW01};

[0171] S3: The IGW encapsulates the outer tunnel header and Segment Routing Header (SRH) (including the VPN SID: vidx-EGW01-SID) based on the SFIB content. The inner packet encapsulation remains unchanged and the packet is sent to the outbound interface. The packet is eventually forwarded to EGW01 via the intermediate P1 node.

[0172] S4: EGW01 receives the tunnel-encapsulated service request message from the IGW, removes the tunnel encapsulation, and retrieves SID1 based on the vidx-EGW01-SID and the SAN-ID feature. It searches the VPN SFIB to obtain instance IP1 and converts the destination address (DA) in the message to IP1, forming a Network Address Translation (NAT) entry.

[0173] S5: EGW01 searches the local IP routing table for the decapsulated message and forwards it to the service instance node.

[0174] S6: The service instance node responds to the service request message, with the source IP being IP1 and the destination IP being UE_IP.

[0175] S7: EGW01 receives the response message and searches the NAT table in S4 based on the message SRC = IP1. It modifies the source IP address of the message to SID1 and searches the corresponding VPN routing table based on the destination IP address of the message. This VPN routing table is derived from the routes advertised by the IGW to the EGW and is usually related to network planning and deployment. A specific SR-TE path can also be planned for the return message, but this embodiment is not limited to this.

[0176] S8: EGW01 encapsulates the outer tunnel header and SRH (including the VPN SID) based on the table lookup result, leaving the inner packet encapsulation unchanged. It then sends the packet to the outbound interface, which is then forwarded to the IGW via the intermediate node.

[0177] S9: The IGW performs normal packet processing based on the packet received in S8. This is a standard L3VPN packet processing flow at a PE (Provider Edge) node and is not detailed here.

[0178] S10: The IGW searches the local VPN routing table based on the service message decapsulated in S9 and finally sends it to the UE. This completes the round-trip processing of the UE service request and service instance response message. Subsequent UE-initiated messages repeat the S1-S10 process.

[0179] The present application also provides a computing network scheduling device. FIG12 is a schematic diagram of the structure of a computing network scheduling device provided by an embodiment. As shown in FIG12, the computing network scheduling device includes:

[0180] A service registration module 510 is configured to generate a service identifier according to a service registration request;

[0181] An indicator determination module 520 is configured to determine a computing power indicator according to a resource state of a service instance corresponding to the service identifier;

[0182] The sending module 530 is configured to send the service identifier and the service level agreement (SLA) information of the corresponding user access to the computing power gateway;

[0183] The calculation module 540 is configured to calculate the computing power route based on the computing power measurement indicator, the computing power information announced by the computing power gateway, and the SLA information.

[0184] The computing network scheduling device of this embodiment distinguishes different types of services by service identifiers. The SLA information corresponding to each service can be flexibly set to meet the multi-factor scheduling requirements of the computing network and meet the scheduling goals of experience, cost, and resource scenarios. By integrating computing power perception and computing network scheduling, it can reduce equipment resource consumption and investment, solve the pressure on the control and forwarding planes caused by the expansion of computing power information scale, and has good feasibility and scalability. By uniformly sending the SAN-ID and mapped SLA information to the computing power gateway, it provides a basis for realizing the joint TE of the computing network and the optimization of computing power routing.

[0185] In one embodiment, the service registration request includes a service domain name, a service IP and port, SLA information, and resource information.

[0186] In one embodiment, the metrics of the SLA information include: latency, cost, bandwidth, jitter, and reliability;

[0187] The computing power measurement indicators include network delay converted from computing delay, network cost converted from resource cost, interface bandwidth converted from total resource amount, and interface bandwidth occupancy converted from resource occupancy.

[0188] In one embodiment, the calculation module 530 includes:

[0189] A path determination unit, configured to determine a traffic engineering calculation path from the computing power entry gateway to the virtual node mapped to the service identifier based on the computing power measurement indicator, the computing power information announced by the computing power gateway, and the SLA information;

[0190] The matching unit is configured to match the traffic engineering calculation path with the global service routing table to obtain a global service route; wherein;

[0191] an instance determining unit, configured to determine the service instance corresponding to the specific service identifier according to the resource state of the service instance, and obtain the local service route;

[0192] Among them, the penultimate node of the traffic engineering calculation path corresponds to the computing power exit gateway IP, and the last node of the traffic engineering calculation path corresponds to the virtual node mapped by the service identifier.

[0193] In one embodiment, the virtual link and link attribute information corresponding to the virtual node are obtained by the computing power egress gateway selecting the service instance for the specific service identifier according to the local service routing table; wherein the local service routing table includes the virtual routing forwarding identifier, the service identifier, the service instance IP address, and the corresponding computing power measurement information;

[0194] The virtual link and link attribute information are spread by the Interior Gateway Protocol (IGP).

[0195] In one embodiment, the global service route includes: virtual private network information, service identifier, service color, computing power export gateway IP, computing power export gateway segment identifier, virtual private network segment identifier;

[0196] The global service routing is formed by spreading computing power information from the computing power egress gateway to the computing power ingress gateway;

[0197] The global service routing information bearer protocol is based on an extension of the BGP protocol suite.

[0198] In one embodiment, the computing power information does not include information about the service instance and preferred entry corresponding to the specific service identifier;

[0199] The computing power export gateway decides whether to announce or revoke the computing power information based on whether there is a corresponding preferred entry locally;

[0200] The computing power information uses the virtual routing forwarding identifier and service identifier as key indexes;

[0201] The computing power information is spread from the computing power export gateway to the computing power import gateway and is uniformly processed by the computing power import gateway.

[0202] The network scheduling device proposed in this embodiment and the network scheduling method proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to any of the above embodiments, and this embodiment has the same beneficial effects as executing the network scheduling method.

[0203] The present application also provides a computing network scheduling device. FIG13 is a schematic diagram of the structure of a computing network scheduling device provided by an embodiment. As shown in FIG13, the computing network scheduling device includes:

[0204] The request module 610 is configured to send a service registration request to the network domain node according to the managed service type;

[0205] The collection module 620 is configured to collect the resource status of the service instance corresponding to the service identifier when the service registration is successful;

[0206] The notification module 630 is configured to notify the computing power information to the computing power export gateway according to the resource status. The computing power information is spread from the computing power export gateway to the upstream gateway, and the computing power import gateway and the computing power export gateway guide the traffic to the service instance.

[0207] The network scheduling device proposed in this embodiment and the network scheduling method proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to any of the above embodiments, and this embodiment has the same beneficial effects as executing the network scheduling method.

[0208] The present application also provides a computing network scheduling device. FIG14 is a schematic diagram of the structure of a computing network scheduling device provided by an embodiment. As shown in FIG14, the computing network scheduling device includes:

[0209] An acquisition module 710 is configured to obtain computing power information notified by a computing power export gateway, as well as a service identifier issued by a network domain node and service level agreement (SLA) information corresponding to user access;

[0210] A routing calculation module 720 is configured to calculate a computing power route corresponding to a specific service identifier based on the computing power measurement indicator, the computing power information, and the SLA information;

[0211] The forwarding module 730 is configured to forward the service request message to the corresponding computing power export gateway according to the computing power routing.

[0212] In one embodiment, it further includes:

[0213] A maintenance module, configured to maintain a global service routing table having multiple next hops for the specific service identifier based on the received computing power information of the multiple computing power export gateways;

[0214] The construction module is configured to construct a traffic engineering model according to the SLA information corresponding to the specific service identifier and the corresponding virtual node, wherein the traffic engineering model includes color information and the virtual node IP mapped to the service identifier.

[0215] In one embodiment, computing power routing includes global service routing;

[0216] The routing calculation module 720 includes:

[0217] a path determination unit, configured to determine a traffic engineering calculation path from a computing power entry gateway to a virtual node mapped to a service identifier based on the computing power measurement indicator, the computing power information, and the SLA information;

[0218] a matching unit, configured to match the traffic engineering calculation path with a global service routing table to obtain the global service route;

[0219] Among them, the penultimate node of the traffic engineering calculation path corresponds to the computing power exit gateway, and the last node of the traffic engineering calculation path corresponds to the virtual node mapped by the service identifier.

[0220] In one embodiment, the computing power information does not include the service instance and preferred item information corresponding to the specific service identifier;

[0221] The computing power export gateway decides whether to announce or revoke the computing power information based on whether there is a corresponding preferred entry locally;

[0222] The computing power information uses the virtual routing forwarding identifier and service identifier as key indexes;

[0223] The computing power information is spread from the computing power export gateway to the direct computing power import gateway and is uniformly processed by the computing power import gateway.

[0224] The network scheduling device proposed in this embodiment and the network scheduling method proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to any of the above embodiments, and this embodiment has the same beneficial effects as executing the network scheduling method.

[0225] The present application also provides a computing network scheduling device. FIG15 is a schematic diagram of the structure of a computing network scheduling device provided by an embodiment. As shown in FIG15, the computing network scheduling device includes:

[0226] A status acquisition module 810 is configured to acquire the resource status of a service instance;

[0227] A determination module 820 is configured to determine a computing power measurement indicator and a local service route according to the resource status;

[0228] An installation module 830 is configured to install the virtual link and link attribute information corresponding to the virtual node, wherein the virtual link and link attribute information are obtained by the computing power egress gateway selecting a service instance for a specific service identifier based on the local service routing surface;

[0229] The diffusion module 840 is configured to diffuse the computing power information notified by the cloud side to the computing power entry gateway.

[0230] In one embodiment, the computing power measurement index includes network delay converted from computing delay, network cost converted from resource cost, interface bandwidth converted from total resource amount, and interface bandwidth occupancy converted from resource occupancy;

[0231] The device also includes: a local generation module, which is configured to generate a local service routing table, wherein the local service routing table includes a virtual routing forwarding VRF identifier, a service identifier, a service instance identifier and corresponding computing power measurement information.

[0232] In one embodiment, the installation module is configured to: based on the local service routing, use the IP corresponding to the service identifier as a virtual node, use the computing power measurement information of the corresponding entry as the link attribute of the virtual link to be flooded within the domain by the IGP, so as to realize the diffusion of the virtual link and link attribute information within the domain.

[0233] In one embodiment, the computing power information does not include information about the service instance and preferred entry corresponding to the specific service identifier;

[0234] The computing power export gateway decides whether to announce or revoke the computing power information based on whether there is a corresponding preferred entry locally;

[0235] The computing power information uses the virtual routing forwarding identifier and service identifier as key indexes;

[0236] The computing power information is spread from the computing power export gateway to the computing power gateway entrance and is uniformly processed by the computing power entrance gateway.

[0237] In one embodiment, the device further includes: a forwarding module configured to search a local service routing table according to a service identifier in the service request message to determine a corresponding service instance, and forward the message to the corresponding service instance.

[0238] The network scheduling device proposed in this embodiment and the network scheduling method proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to any of the above embodiments, and this embodiment has the same beneficial effects as executing the network scheduling method.

[0239] An embodiment of the present application also provides a computing network brain. Figure 16 is a schematic diagram of the hardware structure of a computing network brain provided by an embodiment. As shown in Figure 16, the computing network brain provided by the present application includes a processor 11 and a memory 12; the processor 11 in the computing network brain can be one or more, and Figure 16 takes one processor 11 as an example; the memory 12 is configured to store one or more programs; the one or more programs are executed by the one or more processors 11, so that the one or more processors 11 implement the computing network scheduling method as described in the embodiment of the present application.

[0240] The computing network brain also includes: a communication device 13, an input device 14 and an output device 15.

[0241] The processor 11, memory 12, communication device 13, input device 14 and output device 15 in the computing network brain can be connected through a bus or other means. Figure 16 takes the bus connection as an example.

[0242] The input device 14 can be used to receive input digital or character information and generate key signal input related to user settings and function control of the computing network brain. The output device 15 can include a display device such as a display screen.

[0243] The communication device 13 may include a receiver and a transmitter. The communication device 13 is configured to perform information transmission and reception communication according to the control of the processor 11.

[0244] The memory 12, as a computer-readable storage medium, can be configured to store software programs, computer executable programs, and modules, such as the program instructions / modules corresponding to the computing network scheduling method described in the embodiments of the present application (for example, the service registration module 510, the issuing module 520, and the calculation module 530 in the computing network scheduling device). The memory 12 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and applications required for at least one function; the data storage area may store data created based on the use of the computing network brain, etc. In addition, the memory 12 may include a high-speed random access memory and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 12 may further include a memory remotely located relative to the processor 11, and these remote memories may be connected to the computing network brain via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0245] An embodiment of the present application also provides a cloud-side node. Figure 17 is a schematic diagram of the hardware structure of a cloud-side node provided by an embodiment. As shown in Figure 17, the cloud-side node provided by the present application includes a processor 21 and a memory 22; the processor 21 in the cloud-side node can be one or more, and Figure 17 takes one processor 21 as an example; the memory 22 is configured to store one or more programs; the one or more programs are executed by the one or more processors 21, so that the one or more processors 21 implement the computing network scheduling method as described in the embodiment of the present application.

[0246] The cloud-side node further includes: a communication device 23 , an input device 24 and an output device 25 .

[0247] The processor 21, memory 22, communication device 23, input device 24 and output device 25 in the cloud-side node can be connected via a bus or other means. FIG17 takes the bus connection as an example.

[0248] The input device 24 may be used to receive input digital or character information and generate key signal input related to user settings and function control of the cloud-side node. The output device 25 may include a display device such as a display screen.

[0249] The communication device 23 may include a receiver and a transmitter. The communication device 23 is configured to perform information transmission and reception communication according to the control of the processor 21.

[0250] The memory 22, as a computer-readable storage medium, can be configured to store software programs, computer executable programs, and modules, such as the program instructions / modules corresponding to the computing network scheduling method described in the embodiment of the present application (for example, the request module 610, the acquisition module 620, and the notification module 630 in the computing network scheduling device). The memory 22 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and applications required for at least one function; the data storage area may store data created based on the use of the cloud side node, etc. In addition, the memory 22 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 22 may further include a memory remotely located relative to the processor 21, and these remote memories may be connected to the cloud side node via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0251] An embodiment of the present application also provides a computing power entry gateway. Figure 18 is a schematic diagram of the hardware structure of a computing power entry gateway provided by an embodiment. As shown in Figure 18, the computing power entry gateway provided by the present application includes a processor 31 and a memory 32; the processor 31 in the computing power entry gateway can be one or more, and Figure 18 takes one processor 31 as an example; the memory 32 is configured to store one or more programs; the one or more programs are executed by the one or more processors 31, so that the one or more processors 31 implement the computing network scheduling method as described in the embodiment of the present application.

[0252] The computing power entry gateway also includes: a communication device 33, an input device 34 and an output device 35.

[0253] The processor 31, memory 32, communication device 33, input device 34 and output device 35 in the computing power entry gateway can be connected through a bus or other means. Figure 18 takes the bus connection as an example.

[0254] The input device 34 can be used to receive input digital or character information and generate key signal input related to user settings and function control of the computing power entry gateway. The output device 35 can include a display device such as a display screen.

[0255] The communication device 33 may include a receiver and a transmitter. The communication device 33 is configured to perform information transmission and reception communication according to the control of the processor 31.

[0256] The memory 32, as a computer-readable storage medium, can be configured to store software programs, computer executable programs, and modules, such as the program instructions / modules corresponding to the computing network scheduling method described in the embodiments of the present application (for example, the acquisition module 710, determination module 720, and forwarding module 730 in the computing network scheduling device). The memory 32 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created based on the use of the computing power entry gateway, etc. In addition, the memory 32 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 32 may further include a memory remotely located relative to the processor 31, and these remote memories may be connected to the computing power entry gateway via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0257] An embodiment of the present application also provides a computing power export gateway. Figure 19 is a schematic diagram of the hardware structure of a computing power export gateway provided by an embodiment. As shown in Figure 19, the computing power export gateway provided by the present application includes a processor 41 and a memory 42; the processor 41 in the computing power export gateway can be one or more, and Figure 19 takes one processor 41 as an example; the memory 42 is configured to store one or more programs; the one or more programs are executed by the one or more processors 41, so that the one or more processors 41 implement the computing network scheduling method as described in the embodiment of the present application.

[0258] The computing power export gateway also includes: a communication device 43, an input device 44 and an output device 45.

[0259] The processor 41, memory 42, communication device 43, input device 44 and output device 45 in the computing power export gateway can be connected through a bus or other means. Figure 19 takes the bus connection as an example.

[0260] The input device 44 can be used to receive input digital or character information and generate key signal input related to user settings and function control of the computing power export gateway. The output device 45 can include a display device such as a display screen.

[0261] The communication device 43 may include a receiver and a transmitter. The communication device 43 is configured to perform information transmission and reception communication according to the control of the processor 41.

[0262] Memory 42, as a computer-readable storage medium, can be configured to store software programs, computer executable programs, and modules, such as program instructions / modules corresponding to the computing network scheduling method described in the embodiments of the present application (for example, the diffusion module 810, acquisition module 820, determination module 830, and forwarding module 840 in the computing network scheduling device). Memory 42 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created based on the use of the computing power export gateway, etc. In addition, memory 42 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some instances, memory 42 may further include memory remotely located relative to processor 41, and these remote memories may be connected to the computing power export gateway via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0263] Figure 20 is a schematic diagram of the structure of a computing network scheduling system provided by one embodiment. This embodiment of the application also provides a computing network scheduling system, comprising: a network domain node 910 and a cloud-side node 920 as described in any of the above embodiments; the network domain node 910 includes a computing network brain 911, a packet network device 912, a computing power entry gateway 913 as described in any of the above embodiments, and a computing power exit gateway 914 as described in any of the above embodiments.

[0264] The computing network brain 911 generates a service identifier based on the service registration request from the cloud-side node 920 and sends the service identifier and corresponding service level agreement (SLA) information to the computing power gateway. In centralized computing scenarios, it can also calculate a traffic engineering calculation path based on the computing power metrics, computing power information notified by the computing power gateway, and SLA information, match the traffic engineering calculation path with the global service routing table to obtain a global service route, and send the global service route to the computing power ingress gateway 913. The cloud-side node 920 sends a service registration request to the network domain node 910 based on the managed service type. If the service registration is successful, it collects cloud-side resource status information and, based on the resource status information, notifies the computing power egress gateway 914 of the computing power information. The computing power information is then propagated upstream by the computing power egress gateway 914 through the packet network device 912. The computing power ingress gateway 913 and the computing power egress gateway 914 then direct traffic to the service instance.

[0265] The computing power entry gateway 913 obtains the computing power information announced by the computing power exit gateway, as well as the service identifier and corresponding service level agreement (SLA) information issued by the computing network brain 910. In a distributed computing scenario, the computing power entry gateway 913 can locally calculate the traffic engineering calculation path based on the computing power measurement indicators, the computing power information announced by the computing power gateway, and the SLA information, match the traffic engineering calculation path with the global service routing table, obtain the global service route, and complete the issuance. The service request message is forwarded to the corresponding computing power exit gateway 914 based on the global service route.

[0266] The computing power export gateway 914 obtains the resource status of the service instance; determines the computing power measurement index and the local service route based on the resource status; installs the virtual link and link attribute information corresponding to the virtual node, and the virtual link and link attribute information are obtained by the computing power export gateway by selecting the service instance to the specific service identifier according to the local service routing surface; and spreads the computing power information notified by the cloud side to the computing power entry gateway 913.

[0267] In this embodiment, the computing network brain 911 can perform centralized computing, or the computing power entry gateway 913 can perform distributed head node computing. Different types of services are distinguished by service identifiers, and the SLA information corresponding to each service can be flexibly set to meet the multi-factor scheduling requirements of the computing network as a whole, and meet the scheduling goals of experience, cost, and resource scenarios; by integrating computing power perception and computing network scheduling, it can reduce equipment resource consumption and investment, solve the pressure on the control plane and forwarding plane brought about by the expansion of computing power information scale, and have good feasibility and scalability. By uniformly sending the SAN-ID and mapped SLA information to the computing power gateway, it provides a basis for realizing the joint TE of the computing network and the optimization of computing power routing.

[0268] The embodiment of the present application also provides a storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, it implements the computing network scheduling method described in any of the embodiments of the present application. The method includes: generating a service identifier according to a service registration request; determining a computing power measurement index according to the resource status of the service instance corresponding to the service identifier; sending the service identifier and the service level agreement (SLA) information of the corresponding user access to the computing power gateway; and calculating the computing power routing according to the computing power measurement index, the computing power information notified by the computing power gateway, and the SLA information.

[0269] Alternatively, the method includes: sending a service registration request to a network domain node according to the type of service to be managed; collecting the resource status of the service instance corresponding to the service identifier when the service registration is successful; notifying the computing power information to the computing power export gateway according to the resource status, and the computing power information is spread by the computing power export gateway to the upstream gateway.

[0270] Alternatively, the method includes: obtaining computing power information announced by the computing power export gateway, as well as the service identifier issued by the network domain node and the service level agreement SLA information accessed by the corresponding user; calculating the computing power route corresponding to the specific service identifier based on the computing power measurement indicator, the computing power information and the SLA information; and forwarding the service request message to the corresponding computing power export gateway according to the computing power route.

[0271] Alternatively, the method includes: obtaining the resource status of the service instance; determining the computing power measurement index and the local service route based on the resource status; installing the virtual link and link attribute information corresponding to the virtual node, and the virtual link and link attribute information are obtained by the computing power export gateway by selecting the service instance to the specific service identifier according to the local service routing surface; and spreading the computing power information notified by the cloud side to the computing power entry gateway.

[0272] The computer storage medium of the embodiment of the present application can adopt any combination of one or more computer-readable media.Computer-readable media can be computer-readable signal media or computer-readable storage media.Computer-readable storage media can be, for example, but not limited to: electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or devices, or any combination of the above.More specific examples (non-exhaustive list) of computer-readable storage media include: electrical connections with one or more wires, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM), flash memories, optical fibers, portable CD-ROMs, optical storage devices, magnetic storage devices, or any suitable combination of the above.Computer-readable storage media can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it.

[0273] A computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such a propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0274] The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wire, optical cable, radio frequency (RF), etc., or any suitable combination of the foregoing.

[0275] The computer program code for performing the operations of the present application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and also conventional procedural programming languages ​​such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., using an Internet service provider to connect via the Internet).

[0276] The above description is merely an exemplary embodiment of the present application and is not intended to limit the scope of protection of the present application.

[0277] It will be appreciated by those skilled in the art that the term user terminal covers any suitable type of wireless user equipment, such as a mobile phone, a portable data processing device, a portable web browser or a vehicle-mounted mobile station.

[0278] In general, various embodiments of the present application may be implemented in hardware or dedicated circuits, software, logic, or any combination thereof. For example, some aspects may be implemented in hardware, while other aspects may be implemented in firmware or software that can be executed by a controller, microprocessor, or other computing device, although the present application is not limited thereto.

[0279] Embodiments of the present application may be implemented by executing computer program instructions by a data processor of a mobile device, for example, in a processor entity, or by hardware, or by a combination of software and hardware. The computer program instructions may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages.

[0280] The block diagram of any logical flow in the drawings of this application may represent program steps, or may represent interconnected logical circuits, modules and functions, or may represent a combination of program steps and logical circuits, modules and functions. A computer program may be stored on a memory. The memory may be of any type suitable for the local technical environment and may be implemented using any suitable data storage technology, such as but not limited to read-only memory (ROM), random access memory (RAM), optical storage devices and systems (digital versatile discs (DVD) or compact disks (CD), etc.). Computer-readable media may include non-transitory storage media. The data processor may be of any type suitable for the local technical environment, such as but not limited to a general-purpose computer, a special-purpose computer, a microprocessor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) and a processor based on a multi-core processor architecture.

Claims

1. A computing network scheduling method, applied to a network domain node, comprising: Generate a service identifier according to a service registration request; Determine a computing power measurement indicator according to the resource status of the service instance corresponding to the service identifier; Send the service identifier and the service level agreement (SLA) information of the corresponding user access to the computing power gateway; The computing power route is calculated according to the computing power measurement indicator, the computing power information announced by the computing power gateway, and the SLA information.

2. The method according to claim 1, wherein: The service registration request includes a service domain name, a service Internet Protocol IP and port, the SLA information, and resource information.

3. The method according to claim 1, wherein: The metrics of the SLA information include: latency, cost, bandwidth, jitter, and reliability; The computing power measurement indicators include: network delay obtained by converting computing delay, network cost obtained by converting resource cost, interface bandwidth obtained by converting total resource amount, and interface bandwidth occupancy obtained by converting resource occupancy.

4. The method according to claim 1, wherein: The computing power routing includes global service routing and local service routing; The calculating the computing power route according to the computing power measurement indicator, the computing power information notified by the computing power gateway, and the SLA information includes: Determine a traffic engineering calculation path from the computing power entry gateway to the virtual node mapped to the specific service identifier according to the computing power measurement indicator, the computing power information notified by the computing power gateway, and the SLA information; Matching the traffic engineering calculation path with the global service routing table to obtain the global service routing; Determine the service instance corresponding to the specific service identifier according to the resource state of the service instance, and obtain the local service route; Among them, the penultimate node of the traffic engineering calculation path corresponds to the computing power exit gateway, and the last node of the traffic engineering calculation path corresponds to the virtual node mapped by the specific service identifier.

5. The method according to claim 4, wherein: The virtual link and link attribute information corresponding to the virtual node are obtained by the computing power export gateway selecting the service instance to the specific service identifier according to the local service routing surface; wherein the local service routing table includes the virtual routing forwarding identifier, the service identifier, the service instance IP and the corresponding computing power measurement information; The virtual link and link attribute information are diffused by the Internal Gateway Protocol (IGP).

6. The method according to claim 4, wherein: The global service route includes: virtual private network information, service identifier, service color, computing power export gateway IP, computing power export gateway segment identifier, virtual private network segment identifier; The global service route is formed by the computing power egress gateway diffusing the computing power information to the computing power ingress gateway; The protocol for bearing the global service routing information is based on an extension of the Border Gateway Protocol BGP family.

7. The method according to claim 6, wherein: The computing power information does not include information about the service instance and preferred item corresponding to the specific service identifier; The computing power information is decided by the computing power export gateway whether to announce or revoke it according to whether there is a corresponding preferred entry locally; The computing power information uses the virtual routing forwarding identifier and the service identifier as key indexes; The computing power information is spread from the computing power export gateway to the computing power import gateway and is uniformly processed by the computing power import gateway.

8. A computing network scheduling method, applied to a cloud-side node, comprising: Send a service registration request to the domain node according to the managed service type; If the service registration is successful, the resource status of the service instance corresponding to the service identifier is collected; The computing power information is notified to the computing power export gateway according to the resource status, and the computing power information is diffused by the computing power export gateway to the upstream gateway.

9. A computing network scheduling method, applied to a computing power entry gateway, comprising: Obtain the computing power information announced by the computing power export gateway, as well as the service identifier issued by the network domain node and the service level agreement SLA information accessed by the corresponding user; Calculate the computing power route corresponding to the specific service identifier according to the computing power measurement indicator, the computing power information and the SLA information; The service request message is forwarded to the corresponding computing power export gateway according to the computing power routing.

10. The method according to claim 9, further comprising: Maintaining a global service routing table having multiple next hops for the specific service identifier according to the received computing power information of the multiple computing power export gateways; A traffic engineering model is constructed according to the SLA information corresponding to the specific service identifier and the corresponding virtual node, wherein the traffic engineering model includes color information and the virtual node IP mapped to the service identifier.

11. The method according to claim 10, wherein: The computing power routing includes global service routing; The calculating the computing power route corresponding to the specific service identifier according to the computing power measurement indicator, the computing power information and the SLA information includes: Determine a traffic engineering calculation path from the computing power entry gateway to the virtual node mapped to the service identifier according to the computing power measurement indicator, the computing power information and the SLA information; Matching the traffic engineering calculation path with the global service routing table to obtain the global service routing; The penultimate node of the traffic engineering calculation path corresponds to the computing power export gateway, and the traffic engineering calculation path The last node of corresponds to the virtual node mapped by the service identifier.

12. The method according to claim 10, wherein: The computing power information does not include the service instance and preferred item information corresponding to the specific service identifier; The computing power information is decided by the computing power export gateway whether to announce or revoke it according to whether there is a corresponding preferred entry locally; The computing power information uses the virtual routing forwarding identifier and the service identifier as key indexes; The computing power information is directly spread from the computing power export gateway to the computing power import gateway and is uniformly processed by the computing power import gateway.

13. A computing network scheduling method, applied to a computing power egress gateway, comprising: Get the resource status of the service instance; Determine a computing power measurement indicator and a local service route according to the resource status; Install the virtual link and link attribute information corresponding to the virtual node, wherein the virtual link and link attribute information are obtained by the computing power export gateway selecting a service instance from a specific service identifier according to the local service routing surface; Spread the computing power information notified by the cloud side to the computing power entry gateway.

14. The method according to claim 13, wherein: The computing power measurement indicators include: network delay converted from computing delay, network cost converted from resource cost, interface bandwidth converted from total resource amount, and interface bandwidth occupancy converted from resource occupancy; The method further comprises: Generate the local service routing table, which includes a virtual routing forwarding VRF identifier, a service identifier, a service instance identifier, and corresponding computing power measurement information.

15. The method according to claim 13, wherein: The virtual link and link attribute information corresponding to the virtual node installation includes: According to the local service routing, the IP corresponding to the service identifier is used as a virtual node, and the computing power measurement information of the corresponding entry is used as the link attribute of the virtual link to be flooded within the domain by the IGP, so as to realize the diffusion of the virtual link and link attribute information within the domain.

16. The method according to claim 13, wherein: The computing power information does not include information about the service instance and preferred item corresponding to the specific service identifier; The computing power information is decided by the computing power export gateway whether to announce or revoke it according to whether there is a corresponding preferred entry locally; The computing power information uses the virtual routing forwarding identifier and service identifier as key indexes; The computing power information is diffused from the computing power export gateway to the computing power gateway import gateway and is uniformly processed by the computing power import gateway.

17. The method according to claim 13, further comprising: The local service routing table is searched according to the service identifier in the service request message to determine the corresponding service instance, and the message is forwarded to the corresponding service instance.

18. A network domain node, comprising: memory, and one or more processors; The memory is configured to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the computing network scheduling method as described in any one of claims 1 to 7.

19. A cloud-side node, comprising: memory, and one or more processors; The memory is configured to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the computing network scheduling method as described in claim 8.

20. A computing power entry gateway, comprising: memory, and one or more processors; The memory is configured to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the computing network scheduling method as described in any one of claims 9 to 12.

21. A computing power export gateway, comprising: memory, and one or more processors; The memory is configured to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the computing network scheduling method as described in any one of claims 13-17.

22. A computing network scheduling system, comprising: A network domain node and a cloud side node as claimed in claim 19; The network domain node includes a computing network brain, a packet network device, a computing power entry gateway as described in claim 20, and a computing power exit gateway as described in claim 21.

23. A computer-readable storage medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the computing network scheduling method as described in any one of claims 1 to 17.

Images