A resource coordination and time synchronization method and system for a large-scale computing power cluster

Abstract

The application provides a resource coordination and time synchronization method and system for a large-scale computing power cluster, and relates to the technical field of computing power network architecture. First, the computing power nodes are registered and maintained by an authoritative DNS. After receiving a task, the system intelligently selects the optimal target computing power node and plans the data transmission path. To achieve precise coordination, the system dynamically creates a clock synchronization domain according to the task, elects a master clock and plans the optimal clock synchronization path. Finally, the system encapsulates and packages the optimal task data transmission path, the master clock timestamp and the optimal clock synchronization path identifier, and distributes them to the related nodes, thereby realizing efficient resource coordination and high-precision time synchronization. The method not only realizes the optimal matching of tasks and resources, guarantees the efficiency of data transmission, but also guarantees the accuracy and reliability of time synchronization from the physical link, and finally realizes efficient resource coordination and high-precision time synchronization in the computing power cluster.

Description

Technical Field

[0001] This invention relates to the technical field of computing network architecture, and more specifically, to a method and system for resource coordination and time synchronization for large-scale computing clusters. Background Technology

[0002] A computing power cluster refers to a centralized, large-scale, and collaborative supercomputing resource pool formed by interconnecting a large number of computing nodes, such as GPUs, through a high-speed network. Large-scale computing power cluster networks involve massive amounts of resources, dispersed across different nodes and systems. To coordinate these massive resources, it is necessary to break down resource silos, integrate and share resources, and provide a unified resource view for tasks. However, existing heterogeneous computing power cluster networks suffer from complex device types, large scale, and fragmented management of network and computing resources. Resource registration relies on the Kubernetes platform or proprietary protocols, lacking a unified service-oriented registration and scheduling mechanism, making cross-domain resource retrieval and load balancing difficult.

[0003] In cloud computing scenarios, Segment Routing IPv6 (SRv6), based on the IPv6 forwarding plane, serves as a core protocol for next-generation network technology. It supports flexible traffic scheduling across data centers and dynamic network orchestration in edge computing environments. Computing networks emphasize the deep integration of computing power and the network. The network can be programmed on demand according to business needs, flexibly scheduling ubiquitous resources and coordinating computing and network resources across the entire network to achieve computing power routing and improve overall system performance and efficiency. However, while the existing SRv6 protocol supports path programming, it lacks deep collaboration with the overall network resource management system. Its resource scheduling fails to fully utilize the flexible resource encoding capabilities of SRv6 and Security Identifiers (SIDs), and SID configurations are often static, making it difficult to dynamically respond to changes in resource status. This results in insufficient collaboration between the control and data planes and a lack of deep integration with the global resource status monitoring and scheduling system, making it unable to dynamically adapt to resource changes and meet the needs of low-latency, high-computing-power scenarios.

[0004] Furthermore, time synchronization is a core foundation of computing cluster networks, and its accuracy and real-time performance directly determine the efficiency of distributed computing, the reliability of real-time services, and the utilization rate of global resources. Existing high-precision synchronization solutions using the Precise Time Protocol (PTP) require dedicated hardware, and some devices lack hardware support. Moreover, existing master clock selection cannot adapt to dynamic scenarios and has not been integrated as a core element into the network control plane or resource scheduling decisions, creating optimization blind spots. Summary of the Invention

[0005] To address the problem that existing computing power clusters cannot simultaneously achieve resource coordination and time synchronization, this invention proposes a method and system for resource coordination and time synchronization for large-scale computing power clusters, enabling efficient resource coordination and high-precision time synchronization within the computing power cluster.

[0006] To achieve the above-mentioned technical effects, the technical solution of the present invention is as follows:

[0007] Firstly, this application proposes a resource coordination and time synchronization method for large-scale computing power clusters:

[0008] S1. Manage and initialize the computing nodes in the computing power cluster, register resources with the authoritative DNS using the computing nodes, and synchronize and maintain the status of the computing nodes.

[0009] S2. Receive task requirements, divide the task area based on the task requirements, determine the optimal target computing power node within the task area to respond to the task requirements, and plan the optimal task data transmission path.

[0010] S3. Based on the task area range, the computing power cluster is divided into several corresponding clock synchronization domains. A computing power node in each clock synchronization domain is selected as the master clock node, and its clock status is used as the master clock to generate the master clock timestamp.

[0011] S4. Starting from the master clock node, plan the optimal clock synchronization path and generate the optimal clock synchronization path identifier;

[0012] S5. Encapsulate the optimal task data transmission path, master clock timestamp, and optimal clock synchronization path identifier, and use the master clock node to forward them to all other computing power nodes in the clock synchronization domain to achieve resource collaboration and time synchronization.

[0013] In this technical solution, the computing nodes are first registered and their status maintained using an authoritative DNS. Upon receiving a task, the system intelligently selects the optimal target computing node and plans the data transmission path. To achieve precise collaboration, the system dynamically creates a clock synchronization domain based on the task, elects a master clock, and plans the optimal clock synchronization path. Finally, the system encapsulates the optimal task data transmission path, the master clock timestamp, and the optimal clock synchronization path identifier, and distributes them to the relevant nodes, thereby achieving efficient resource collaboration and high-precision time synchronization. This method not only achieves optimal matching between tasks and resources, ensuring efficient data transmission, but also guarantees the accuracy and reliability of time synchronization from the physical link perspective, ultimately realizing efficient resource collaboration and high-precision time synchronization in the computing cluster.

[0014] Preferably, step S1, which involves managing and initializing the computing nodes of the computing power cluster, comprises the following steps:

[0015] After the computing node starts up, it establishes a secure channel with SDN based on the Secure Shell protocol and the transport layer security protocol, and communicates through the NETCONF network configuration protocol.

[0016] SDN verifies the identity of computing power nodes and determines the device type of computing power nodes based on certificates or pre-shared keys, and records the device address of computing power nodes;

[0017] SDN uses the NETCONF network configuration protocol to send the master clock address to the computing power node and receive the time synchronization confirmation information from the computing power node.

[0018] SDN distributes DNS server addresses and resource registration templates to computing nodes, which are used to configure DNS server resolution rules for the computing nodes and complete the management and initialization of the computing nodes.

[0019] Preferably, the process of registering resources with the authoritative DNS using computing power nodes in step S1 is as follows:

[0020] The computing nodes generate DNS resource records based on the DNS server address and resource registration template issued by SDN, and use transaction signature TSIG authentication to ensure the legitimacy of the request. The DNS resource records include SID-INFO records and SRV records. The SID-INFO records are used to record resource status, including: the IP address of the computing node, the device type of the computing node, the real-time load rate, and the geographical location of the computing node. The SRV records are used to locate the standard record type of the service and point to the service port.

[0021] The computing node sends a DNS dynamic update protocol request to the authoritative DNS server. After the request is approved, the resource status is registered with the DNS server in real time using the DNS dynamic update protocol.

[0022] The authoritative DNS server verifies whether the IP address of the computing node sending the DNS dynamic update protocol request is in the SDN pre-authorized list and verifies the validity of the transaction signature TSIG. If both verifications pass, the registration is successful. If either verification fails, the registration fails, and the computing node re-registers according to the exponential backoff strategy until SDN intervenes to repair the issue.

[0023] Preferably, the state synchronization and maintenance of the computing power node in step S1 includes constructing a state synchronization and maintenance mechanism for the computing power node to synchronize the resource status of the computing power node;

[0024] The state synchronization and maintenance mechanism of the computing nodes includes an SDN periodic monitoring mechanism and a short-cycle monitoring mechanism;

[0025] The process of the SDN periodic monitoring mechanism is as follows:

[0026] The computing nodes periodically report their load and health status to the SDN. The SDN determines whether the load of the computing nodes exceeds a pre-set threshold. If it does, the SDN triggers a DNS record weight adjustment, reducing the priority of SRV records. If it does not exceed the threshold, the SDN continues to monitor the load and health status of the computing nodes.

[0027] The short-cycle monitoring mechanism includes: setting the TTL of the SID-INFO record to a short cycle; in each short cycle, the computing node sends a DNS UPDATE request to SDN at least once to refresh its SID-INFO record; otherwise, SDN considers the computing node offline and uses NETCONF to delete the node configuration and calls the DNS API to delete the SID-INFO record.

[0028] Preferably, the process of step S2 is as follows:

[0029] S21.SDN divides the task area based on task requirements, which include the type and number of computing power nodes, and generates a task area identifier (Task-ID) and a list of computing power node members.

[0030] S22. Determine whether the task requires specific computing resources. If so, the computing node sends a resource request to the SDN, the SDN sends a resource query request to the local DNS, the local DNS returns a list of matching candidate resource nodes, and S23 is executed. If the task does not require specific computing resources, S23 is executed directly.

[0031] S23. Select the optimal computing node to execute the task, use a multi-factor weighting algorithm to calculate the node selection weight, select the node with the highest weight as the target computing node, and generate an explicit path containing the IP address of the target computing node and the SID sequence of intermediate forwarding nodes.

[0032] S24.SDN explicitly sends the path containing the target compute node's IP address and the SID sequence of intermediate forwarding nodes to the switch of the compute node, thus fulfilling the task requirements.

[0033] Preferably, the process of dividing the computing cluster into several clock synchronization domains based on the task area range in step S3 is as follows:

[0034] SDN uses the computing power node located at the center of the physical topology within each task area as the core of the clock synchronization domain. It jumps outward from the core of the time synchronization domain until the preset maximum number of node jumps is reached. The computing power node where the preset maximum number of node jumps is reached is defined as the boundary of the time synchronization domain, thus completing the division of the time synchronization domain.

[0035] Preferably, the master clock selection process in step S3 is as follows:

[0036] The operating parameters of all computing nodes within the same time synchronization domain are collected as the scoring criteria. The operating parameters include in-degree, throughput, clock precision level and link latency. The weights of the scoring criteria are initialized based on the network type.

[0037] Based on the scoring criteria, the score of each computing node is calculated, and based on the network status, the score of the computing node is updated according to a preset time period.

[0038] The clock state of the highest-scoring computing power node is set as the master clock. The highest-scoring computing power node is the master clock node, and the other computing power nodes are slave clock nodes. If two or more computing power nodes have the same score, they are compared level by level based on the preset scoring criteria priority until the unique and optimal computing power node is selected. The priority of the scoring criteria is as follows: in-degree, clock accuracy level, throughput, and link latency.

[0039] Preferably, the process of generating the optimal clock synchronization path identifier is as follows:

[0040] A minimum spanning tree is constructed using Dijkstra's algorithm. The root node of the minimum spanning tree is the master clock node, which is used to plan the shortest path from the master clock computing power node to all other nodes in the same time synchronization domain.

[0041] Generate the optimal clock synchronization path identifier based on the shortest path.

[0042] Secondly, this application also proposes a resource coordination and time synchronization system for large-scale computing power clusters, the system comprising:

[0043] The computing node initialization unit is used to manage and initialize computing nodes in the computing cluster, register resources with the authoritative DNS using computing nodes, and synchronize and maintain the status of computing nodes.

[0044] The data transmission path planning unit is used to receive task requirements, divide the task area based on the task requirements, determine the optimal target computing power node that responds to the task requirements within the task area, and plan the optimal task data transmission path.

[0045] The master clock timestamp generation unit is used to divide the computing power cluster into several corresponding clock synchronization domains based on the task area range, select a computing power node in each clock synchronization domain as the master clock node, take its clock status as the master clock, and generate the master clock timestamp.

[0046] The clock synchronization path generation unit is used to plan the optimal clock synchronization path starting from the master clock node and generate the optimal clock synchronization path identifier.

[0047] The resource coordination and time calibration unit is used to encapsulate the optimal task data transmission path, the master clock timestamp, and the optimal clock synchronization path identifier, and forward them to all other computing power nodes in the clock synchronization domain using the master clock node, thereby achieving resource coordination and time synchronization.

[0048] Thirdly, this application also proposes a computer device, the computer device comprising: a memory, a processor, and a computer program stored in the memory that can be run by the processor, the processor executing the computer program.

[0049] Compared with the prior art, the beneficial effects of the present invention are:

[0050] This invention proposes a resource coordination and time synchronization method and system for large-scale computing power clusters. It utilizes an authoritative DNS for resource registration and status maintenance of computing power nodes. Upon receiving a task, it intelligently selects the optimal target computing power node and plans the data transmission path. To achieve precise coordination, it dynamically creates a clock synchronization domain based on the task, elects a master clock, and plans the optimal clock synchronization path. Finally, it encapsulates the optimal task data transmission path, the master clock timestamp, and the optimal clock synchronization path identifier, and distributes them to relevant nodes, thereby achieving efficient resource coordination and high-precision time synchronization. This method not only achieves optimal matching between tasks and resources, ensuring efficient data transmission, but also guarantees the accuracy and reliability of time synchronization from the physical link perspective, ultimately realizing efficient resource coordination and high-precision time synchronization in computing power clusters. Attached Figure Description

[0051] Figure 1 This is a flowchart illustrating a resource coordination and time synchronization method for large-scale computing power clusters proposed in Embodiment 1 of the present invention.

[0052] Figure 2 This diagram illustrates the system architecture for resource collaboration and time synchronization for large-scale computing power clusters proposed in Embodiment 1 of the present invention.

[0053] Figure 3 This is a schematic diagram illustrating the structure of a resource coordination and time synchronization system for large-scale computing power clusters proposed in Embodiment 3 of the present invention.

[0054] Figure 4 This is a schematic diagram of the structure of the computer device proposed in Embodiment 4 of the present invention. Detailed Implementation

[0055] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this patent.

[0056] To better illustrate this embodiment, some parts of the accompanying drawings may be omitted, enlarged, or reduced, and do not represent the actual dimensions;

[0057] It is understandable to those skilled in the art that some well-known details may be omitted from the accompanying drawings.

[0058] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0059] The positional relationships depicted in the accompanying drawings are for illustrative purposes only and should not be construed as limiting this patent.

[0060] Example 1

[0061] This embodiment proposes a resource coordination and time synchronization method for large-scale computing clusters. A flowchart illustrating this method can be found here. Figure 1 This includes the following steps:

[0062] S1. Manage and initialize the computing nodes in the computing power cluster, register resources with the authoritative DNS using the computing nodes, and synchronize and maintain the status of the computing nodes.

[0063] S2. Receive task requirements, divide the task area based on the task requirements, determine the optimal target computing power node within the task area to respond to the task requirements, and plan the optimal task data transmission path.

[0064] S3. Based on the task area range, the computing power cluster is divided into several corresponding clock synchronization domains. A computing power node in each clock synchronization domain is selected as the master clock node, and its clock status is used as the master clock to generate the master clock timestamp.

[0065] S4. Starting from the master clock node, plan the optimal clock synchronization path and generate the optimal clock synchronization path identifier;

[0066] S5. Encapsulate the optimal task data transmission path, master clock timestamp, and optimal clock synchronization path identifier, and use the master clock node to forward them to all other computing power nodes in the clock synchronization domain to achieve resource collaboration and time synchronization.

[0067] In this embodiment, the computing nodes are first registered and their status maintained using an authoritative DNS. Upon receiving a task, the system intelligently selects the optimal target computing node and plans the data transmission path. To achieve precise collaboration, the system dynamically creates a clock synchronization domain based on the task, elects a master clock, and plans the optimal clock synchronization path. Finally, the system encapsulates the optimal task data transmission path, the master clock timestamp, and the optimal clock synchronization path identifier, and distributes them to the relevant nodes, thereby achieving efficient resource collaboration and high-precision time synchronization. This method not only achieves optimal matching between tasks and resources, ensuring efficient data transmission, but also guarantees the accuracy and reliability of time synchronization from the physical link perspective, ultimately realizing efficient resource collaboration and high-precision time synchronization in the computing cluster.

[0068] Specifically, the system architecture for implementing this method is divided into a collaborative application layer 1, a controller layer 2, and an infrastructure layer 3; see the structural diagram of the system architecture for details. Figure 2 The collaborative application layer 1 includes time synchronization service, DNS service and application (computing tasks), the controller layer 2 mainly includes SDN controller, and the infrastructure layer 3 includes intelligent computing cluster nodes such as switches, routers, servers, and virtual machines. The collaborative application layer 1 and the controller layer 2 are extended based on the SDN northbound interface 4, and the controller layer 2 and the infrastructure layer 3 are extended based on the SDN southbound interface 5.

[0069] Example 2

[0070] In this embodiment, step S1, which involves managing and initializing the computing nodes of the computing power cluster, is as follows:

[0071] After the computing node starts up, it establishes a secure channel with SDN based on the Secure Shell protocol and the transport layer security protocol, and communicates through the NETCONF network configuration protocol.

[0072] SDN verifies the identity of computing power nodes and determines the device type of computing power nodes based on certificates or pre-shared keys, and records the device address of computing power nodes;

[0073] SDN uses the NETCONF network configuration protocol to send the master clock address to the computing power node and receive the time synchronization confirmation information from the computing power node.

[0074] SDN distributes DNS server addresses and resource registration templates to computing nodes, which are used to configure DNS server resolution rules for the computing nodes and complete the management and initialization of the computing nodes.

[0075] In this embodiment, step S1, which involves registering resources with an authoritative DNS using a computing power node, is as follows:

[0076] The computing nodes generate DNS resource records based on the DNS server address and resource registration template issued by SDN, and use transaction signature TSIG authentication to ensure the legitimacy of the request. The DNS resource records include SID-INFO records and SRV records. The SID-INFO records are used to record resource status, including: the IP address of the computing node, the device type of the computing node, the real-time load rate, and the geographical location of the computing node. The SRV records are used to locate the standard record type of the service and point to the service port.

[0077] The computing node sends a DNS dynamic update protocol request to the authoritative DNS server. After the request is approved, the resource status is registered with the DNS server in real time using the DNS dynamic update protocol.

[0078] The authoritative DNS server verifies whether the IP address of the computing node sending the DNS dynamic update protocol request is in the SDN pre-authorized list and verifies the validity of the transaction signature TSIG. If both verifications pass, the registration is successful. If either verification fails, the registration fails, and the computing node re-registers according to the exponential backoff strategy until SDN intervenes to repair the issue.

[0079] Specifically, after the computing node starts up, it can establish a secure channel with the SDN controller via SSH / TLS and communicate using the NETCONF protocol. The SDN controller verifies the node's identity based on the certificate or pre-shared key, confirms the device role, such as "server node", and synchronizes the device address to the controller.

[0080] Specifically, after the master clock is determined, the SDN controller uses the NETCONF protocol. <edit-config>The operation involves sending the master clock address, for example, time.example.com, to the computing nodes; after the nodes synchronize their time, they return a response to the controller. <rpc-reply>Confirm that the configuration is effective; after it is effective, the SDN controller will issue the DNS server address, such as dns.example.com, and the resource registration template, and configure the node DNS resolution rules.

[0081] Specifically, the computing nodes generate DNS resource records based on the resource registration template issued by the SDN controller, including: SID-INFO record: a custom record type containing node IP, computing power type, such as GPU-NVIDIA-A100, real-time load rate, geographical location and other metadata; SRV record: pointing to the service port, such as _srv6._tcp.example.com 86400 INSRV 0 5 8080 node1.example.com.

[0082] In this application, three new types of resource records are added to the DNS server, including:

[0083] SRV-COMPUTE: Registers computing resource attributes (CPU / GPU model, video memory / storage capacity, real-time load); SRV-NETWORK: Registers network resources (bandwidth margin, latency, topology location); SRV-MODEL: Registers AI model parameter version, storage location, and API call.

[0084] In this embodiment, the state synchronization and maintenance of computing power nodes in step S1 includes constructing a state synchronization and maintenance mechanism for computing power nodes to synchronize the resource status of computing power nodes;

[0085] The state synchronization and maintenance mechanism of the computing nodes includes an SDN periodic monitoring mechanism and a short-cycle monitoring mechanism;

[0086] The process of the SDN periodic monitoring mechanism is as follows:

[0087] The computing nodes periodically report their load and health status to the SDN. The SDN determines whether the load of the computing nodes exceeds a pre-set threshold. If it does, the SDN triggers a DNS record weight adjustment, reducing the priority of SRV records. If it does not exceed the threshold, the SDN continues to monitor the load and health status of the computing nodes.

[0088] The short-cycle monitoring mechanism includes: setting the TTL of the SID-INFO record to a short cycle; in each short cycle, the computing node sends a DNS UPDATE request to SDN at least once to refresh its SID-INFO record; otherwise, SDN considers the computing node offline and uses NETCONF to delete the node configuration and calls the DNS API to delete the SID-INFO record.

[0089] Specifically, the computing nodes are connected via NETCONF. <notification>The system periodically reports load and health status to the SDN controller. The threshold for the computing node is 90%. If the node load exceeds the threshold, the SDN controller triggers a DNS record weight adjustment, reducing the priority of the SRV record.

[0090] Specifically, one cycle of the survival period TTL is 60 seconds.

[0091] Specifically, the state synchronization and maintenance of the computing nodes also includes anomaly handling and security mechanisms, including:

[0092] The NETCONF configuration rollback mechanism is as follows: if time synchronization or DNS address delivery fails, the controller rolls back to the previous version configuration and records the alarm; the DNS registration retry mechanism is as follows: when a node fails to register resources, it retryes according to the exponential backoff strategy until the SDN controller intervenes to repair it.

[0093] In this embodiment, step S2 is as follows:

[0094] S21.SDN divides the task area based on task requirements, which include the type and number of computing power nodes, and generates a task area identifier (Task-ID) and a list of computing power node members.

[0095] S22. Determine whether the task requires specific computing resources. If so, the computing node sends a resource request to the SDN, the SDN sends a resource query request to the local DNS, the local DNS returns a list of matching candidate resource nodes, and S23 is executed. If the task does not require specific computing resources, S23 is executed directly.

[0096] S23. Select the optimal computing node to execute the task, use a multi-factor weighting algorithm to calculate the node selection weight, select the node with the highest weight as the target computing node, and generate an explicit path containing the IP address of the target computing node and the SID sequence of intermediate forwarding nodes.

[0097] S24.SDN explicitly sends the path containing the target compute node's IP address and the SID sequence of intermediate forwarding nodes to the switch of the compute node, thus fulfilling the task requirements.

[0098] Specifically, the rules for dividing the task area are as follows:

[0099] First, the SDN controller calls the DNS global resource view to filter nodes that meet the task conditions. Then, based on the k-means clustering algorithm, the task regions are divided according to the node's geographical location, resource type, and network latency. Finally, the number of nodes in the region is adjusted considering the region size limit. Specifically, the number of nodes in a single region is ≤128 (adjusted according to the task size), and the maximum number of hops is ≤8 (to avoid time synchronization accuracy degradation).

[0100] Specifically, when a task requires specific computing resources or the load on a computing node reaches a threshold, the node sends a resource request to the SDN controller via a southbound interface (such as NETCONF or gRPC). The request includes metadata such as resource type (GPU computing power, storage space), priority of the request, and service QoS requirements (latency, bandwidth). After receiving the resource request, the SDN controller sends a query to the local DNS via a northbound interface (RESTful API), carrying the resource type and service constraints (such as "GPU type = Tesla V100, latency ≤ 50ms"). After sending the resource request, the local DNS returns a list of matching candidate resource nodes, or queries the authoritative DNS and returns the list, including the IP address of each node, resource type, current load rate, and SID encoding template (such as the Locator prefix).

[0101] Filtering invalid nodes involves the controller excluding unreachable nodes based on the network topology (link status collected by the LLDP protocol).

[0102] The weights of the multi-factor weighting algorithm include: resource load (e.g., a high score is given for GPU utilization ≤ 80%).

[0103] Path cost (hop count and latency of explicit SRv6 paths) and service priority (high-priority services are forced to select low-load nodes).

[0104] The SID structure design includes: a Locator field, used to identify the resource pool domain with a fixed prefix (e.g., 2001:db8:1:: / 64); and a Function field, used to dynamically write the last 64 bits of the target resource IP and embed the resource type identifier (e.g., 0xA1 indicates GPU resource).

[0105] An example of SID encoding is as follows: If the target resource IP is 2001:db8:2:1::10 and the resource type is GPU, then the SID encoding is 2001:db8:1::A1:2001:db8:2:1::10.

[0106] Specifically, the process of explicitly distributing SID sequences includes: OpenFlow flow table injection, where the controller distributes SRv6 SID encoding and forwarding rules to the switches of the initial or intermediate nodes through the southbound interface, specifying the End.X behavior (explicit next hop); and BGP-LS synchronization, where, in cross-domain scenarios, the controller broadcasts SID path information to neighboring domains through the BGP-LS protocol to ensure global route reachability.

[0107] The distribution process also includes bidirectional forwarding detection (BFD): the controller and nodes monitor path connectivity through the BFD protocol, and if the packet loss rate is greater than 1%, rerouting is triggered.

[0108] In this embodiment, the process of dividing the computing cluster into several clock synchronization domains based on the task area range in step S3 is as follows:

[0109] SDN uses the computing power node located at the center of the physical topology within each task area as the core of the clock synchronization domain. It jumps outward from the core of the time synchronization domain until the preset maximum number of node jumps is reached. The computing power node where the preset maximum number of node jumps is reached is defined as the boundary of the time synchronization domain, thus completing the division of the time synchronization domain.

[0110] Specifically, the maximum number of node hops is preset to 8 (refer to the default value of the MSTP protocol) to avoid a decrease in synchronization accuracy and accumulation of delays due to excessive hop count.

[0111] Specifically, the scope of the time synchronization domain is dynamically adjusted. The SDN controller collects link status in real time through southbound interfaces (such as OpenFlow) and combines federated learning to predict network congestion and topology changes to dynamically adjust the domain scope. When the number of hops in a domain exceeds a threshold, a split is triggered, and a new domain elects a local master clock.

[0112] In this embodiment, the master clock selection process in step S3 is as follows:

[0113] The operating parameters of all computing nodes within the same time synchronization domain are collected as the scoring criteria. The operating parameters include in-degree, throughput, clock precision level and link latency. The weights of the scoring criteria are initialized based on the network type.

[0114] Based on the scoring criteria, the score of each computing node is calculated, and based on the network status, the score of the computing node is updated according to a preset time period.

[0115] The clock state of the highest-scoring computing power node is set as the master clock. The highest-scoring computing power node is the master clock node, and the other computing power nodes are slave clock nodes. If two or more computing power nodes have the same score, they are compared level by level based on the preset scoring criteria priority until the unique and optimal computing power node is selected. The priority of the scoring criteria is as follows: in-degree, clock accuracy level, throughput, and link latency.

[0116] Specifically, the weights of the scoring criteria include: in-degree weight. This is used to measure how frequently a node is referenced by other nodes, reflecting its reliability as a time reference. The expression is:

[0117]

[0118] in, Rate the in-degree score. As the in-degree weight, In the Industrial Internet, ;

[0119] Out-degree weight This is used to measure a node's ability to verify other nodes, reflecting its contribution to global time synchronization. The expression is:

[0120]

[0121] in, For output rating, For out-degree weight, In the Industrial Internet, ;

[0122] Overall In / Out Score The expression is:

[0123]

[0124] Throughput weight This is used to represent a node's ability to process PTP synchronization messages; throughput score. The expression is:

[0125]

[0126] in, Scoring based on throughput As throughput weight, In the Industrial Internet, ;

[0127] Clock accuracy rating Based on clock quality level weighting With clock stability weight The calculation is performed based on the clock quality level and stability of IEEE 1588, and the expression is as follows:

[0128]

[0129] in, Indicates the clock quality level. Indicates stability. In the Industrial Internet, , ;

[0130] Link delay score Based on time delay weight With jitter weight The calculation is performed, and the expression is:

[0131]

[0132] in, This represents the actual transmission delay. Indicates the actual jitter. In the Industrial Internet, ,

[0133] The combined in-degree / out-of-degree score, throughput score, clock accuracy score, and link delay score are expressed as follows:

[0134]

[0135] in, The weights of the scores are dynamically adjusted in different application scenarios. For example, when the network topology is stable, the weights are increased. (In-degree and out-degree weights); In high-concurrency scenarios, improve... (Throughput weight); clock precision sensitive scenarios (such as industrial control), improve ;

[0136] Specifically, the scoring of the operating parameters of the Industrial Internet of Things (IIoT) is based on the following weights: in-degree weight. Out-degree weight Throughput weight Clock quality level weight Clock stability weight Delay weight Dithering weight In satellite networks, increasing latency weighting. Reduce throughput weight to 0.5. Up to 0.1.

[0137] Specifically, the preset time period is one hour.

[0138] In this embodiment, the process of generating the optimal clock synchronization path identifier is as follows:

[0139] A minimum spanning tree is constructed using Dijkstra's algorithm. The root node of the minimum spanning tree is the master clock node, which is used to plan the shortest path from the master clock computing power node to all other nodes in the same time synchronization domain.

[0140] Generate the optimal clock synchronization path identifier based on the shortest path.

[0141] Specifically, the weight of the path planning is the time delay. For link utilization, prioritize low-load, low-jitter paths.

[0142] Specifically, when a node performs clock synchronization forwarding based on the clock synchronization path identifier, it also needs to consider the latency during path delay forwarding and embed the Time-Sync TLV field in the SRv6 extension header, carrying the master clock timestamp and path delay compensation value.

[0143] Intermediate nodes parse TLV and correct local clocks to achieve hop-by-hop synchronization (accuracy ≤ 1μs).

[0144] Specifically, after selecting the master clock, the following steps are also included: periodic detection, collecting throughput and latency every hour; verifying clock accuracy every 24 hours via NETCONF rpc / rpc-reply messages; if the master clock triggers a reselection threshold, then a new master clock is selected. The reselection thresholds include: absolute threshold: forced reselection when the master clock score is <0.7 (out of 1.0); relative threshold: the score of the suboptimal node exceeds 85% of the current master clock and continues for 3 monitoring cycles (default cycle is 1 hour); and a fast switching mechanism: when the master clock fails, a node is selected from the rotating queue to take over.

[0145] Example 3

[0146] This embodiment proposes a resource coordination and time synchronization system for large-scale computing power clusters. In this embodiment, the system is used to implement a resource coordination and time synchronization method for large-scale computing power clusters. The structural diagram is shown below. Figure 3 As shown, it includes:

[0147] The computing node initialization unit is used to manage and initialize computing nodes in the computing cluster, register resources with the authoritative DNS using computing nodes, and synchronize and maintain the status of computing nodes.

[0148] The data transmission path planning unit is used to receive task requirements, divide the task area based on the task requirements, determine the optimal target computing power node that responds to the task requirements within the task area, and plan the optimal task data transmission path.

[0149] The master clock timestamp generation unit is used to divide the computing power cluster into several corresponding clock synchronization domains based on the task area range, select a computing power node in each clock synchronization domain as the master clock node, take its clock status as the master clock, and generate the master clock timestamp.

[0150] The clock synchronization path generation unit is used to plan the optimal clock synchronization path starting from the master clock node and generate the optimal clock synchronization path identifier.

[0151] The resource coordination and time calibration unit is used to encapsulate the optimal task data transmission path, the master clock timestamp, and the optimal clock synchronization path identifier, and forward them to all other computing power nodes in the clock synchronization domain using the master clock node, thereby achieving resource coordination and time synchronization.

[0152] Example 4

[0153] In this embodiment, a computer device is proposed, comprising a memory 101, a processor 102, and a computer program stored in the memory 101 that can be executed by the processor. The processor 102 executes the computer program to implement a resource coordination and time synchronization method for large-scale computing power clusters. A schematic diagram of the device is shown below. Figure 4 As shown.

[0154] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.< / notification>

Claims

1. A method for resource coordination and time synchronization for large-scale computing power clusters, characterized in that, Includes the following steps: S1. Manage and initialize the computing nodes in the computing power cluster, register resources with the authoritative DNS using the computing nodes, and synchronize and maintain the status of the computing nodes. S2. Receive task requirements, divide the task area based on the task requirements, determine the optimal target computing power node within the task area to respond to the task requirements, and plan the optimal task data transmission path. S3. Based on the task area range, the computing power cluster is divided into several corresponding clock synchronization domains. A computing power node in each clock synchronization domain is selected as the master clock node, and its clock status is used as the master clock to generate the master clock timestamp. S4. Starting from the master clock node, plan the optimal clock synchronization path and generate the optimal clock synchronization path identifier; S5. Encapsulate the optimal task data transmission path, master clock timestamp, and optimal clock synchronization path identifier, and use the master clock node to forward them to all other computing power nodes in the clock synchronization domain to achieve resource collaboration and time synchronization.

2. The resource coordination and time synchronization method for large-scale computing power clusters according to claim 1, characterized in that, Step S1 describes the process of managing and initializing the computing nodes of the computing power cluster. After the computing node starts up, it establishes a secure channel with SDN based on the Secure Shell protocol and the transport layer security protocol, and communicates through the NETCONF network configuration protocol. SDN verifies the identity of computing power nodes and determines the device type of computing power nodes based on certificates or pre-shared keys, and records the device address of computing power nodes; SDN uses the NETCONF network configuration protocol to send the master clock address to the computing power node and receive the time synchronization confirmation information from the computing power node. SDN distributes DNS server addresses and resource registration templates to computing nodes, which are used to configure DNS server resolution rules for the computing nodes and complete the management and initialization of the computing nodes.

3. The resource coordination and time synchronization method for large-scale computing power clusters according to claim 2, characterized in that, Step S1, which involves registering resources with an authoritative DNS using computing power nodes, is as follows: The computing nodes generate DNS resource records based on the DNS server address and resource registration template issued by SDN, and use transaction signature TSIG authentication to ensure the legitimacy of the request. The DNS resource records include SID-INFO records and SRV records. The SID-INFO records are used to record resource status, including: the IP address of the computing node, the device type of the computing node, the real-time load rate, and the geographical location of the computing node. The SRV records are used to locate the standard record type of the service and point to the service port. The computing node sends a DNS dynamic update protocol request to the authoritative DNS server. After the request is approved, the resource status is registered with the DNS server in real time using the DNS dynamic update protocol. The authoritative DNS server verifies whether the IP address of the computing node sending the DNS dynamic update protocol request is in the SDN pre-authorized list and verifies the validity of the transaction signature TSIG. If both verifications pass, the registration is successful. If either verification fails, the registration fails, and the computing node re-registers according to the exponential backoff strategy until SDN intervenes to repair the issue.

4. The resource coordination and time synchronization method for large-scale computing power clusters according to claim 3, characterized in that, The state synchronization and maintenance of computing power nodes in step S1 includes constructing a state synchronization and maintenance mechanism for computing power nodes to synchronize the resource status of computing power nodes; The state synchronization and maintenance mechanism of the computing nodes includes an SDN periodic monitoring mechanism and a short-cycle monitoring mechanism; The process of the SDN periodic monitoring mechanism is as follows: The computing nodes periodically report their load and health status to the SDN. The SDN determines whether the load of the computing nodes exceeds a pre-set threshold. If it does, the SDN triggers a DNS record weight adjustment, reducing the priority of the SRV record. If the limit is not exceeded, SDN continues to monitor the load and health status of the computing power nodes; The short-cycle monitoring mechanism includes: setting the TTL of the SID-INFO record to a short cycle; in each short cycle, the computing node sends a DNS UPDATE request to SDN at least once to refresh its SID-INFO record; otherwise, SDN considers the computing node offline and uses NETCONF to delete the node configuration and calls the DNS API to delete the SID-INFO record.

5. The resource coordination and time synchronization method for large-scale computing power clusters according to claim 4, characterized in that, The process of step S2 is as follows: S21.SDN divides the task area based on task requirements, which include the type and number of computing power nodes, and generates a task area identifier (Task-ID) and a list of computing power node members. S22. Determine whether the task requires specific computing resources. If so, the computing node sends a resource request to the SDN, the SDN sends a resource query request to the local DNS, the local DNS returns a list of matching candidate resource nodes, and S23 is executed. If the task does not require specific computing resources, S23 is executed directly. S23. Select the optimal computing node to execute the task, use a multi-factor weighting algorithm to calculate the node selection weight, select the node with the highest weight as the target computing node, and generate an explicit path containing the IP address of the target computing node and the SID sequence of intermediate forwarding nodes. S24.SDN explicitly sends the path containing the target compute node's IP address and the SID sequence of intermediate forwarding nodes to the switch of the compute node, thus fulfilling the task requirements.

6. The resource coordination and time synchronization method for large-scale computing power clusters according to claim 5, characterized in that, Step S3 describes the process of SDN dividing the computing cluster into several clock synchronization domains based on the task area range: SDN uses the computing power node located at the center of the physical topology within each task area as the core of the clock synchronization domain. It jumps outward from the core of the time synchronization domain until the preset maximum number of node jumps is reached. The computing power node where the preset maximum number of node jumps is reached is defined as the boundary of the time synchronization domain, thus completing the division of the time synchronization domain.

7. The resource coordination and time synchronization method for large-scale computing power clusters according to claim 6, characterized in that, The master clock selection process described in step S3 is as follows: The operating parameters of all computing nodes within the same time synchronization domain are collected as the scoring criteria. The operating parameters include in-degree, throughput, clock precision level and link latency. The weights of the scoring criteria are initialized based on the network type. Based on the scoring criteria, the score of each computing node is calculated, and based on the network status, the score of the computing node is updated according to a preset time period. The clock state of the highest-scoring computing power node is set as the master clock. The highest-scoring computing power node is the master clock node, and the other computing power nodes are slave clock nodes. If two or more computing power nodes have the same score, they are compared level by level based on the preset scoring criteria priority until the unique and optimal computing power node is selected. The priority of the scoring criteria is as follows: in-degree, clock accuracy level, throughput, and link latency.

8. The resource coordination and time synchronization method for large-scale computing power clusters according to claim 7, characterized in that, The process of generating the optimal clock synchronization path identifier is as follows: A minimum spanning tree is constructed using Dijkstra's algorithm. The root node of the minimum spanning tree is the master clock node, which is used to plan the shortest path from the master clock computing power node to all other nodes in the same time synchronization domain. Generate the optimal clock synchronization path identifier based on the shortest path.

9. A resource coordination and time synchronization system for large-scale computing power clusters, characterized in that, The system is used to implement the method according to any one of claims 1 to 8, comprising: The computing node initialization unit is used to manage and initialize computing nodes in the computing cluster, register resources with the authoritative DNS using computing nodes, and synchronize and maintain the status of computing nodes. The data transmission path planning unit is used to receive task requirements, divide the task area based on the task requirements, determine the optimal target computing power node that responds to the task requirements within the task area, and plan the optimal task data transmission path. The master clock timestamp generation unit is used to divide the computing power cluster into several corresponding clock synchronization domains based on the task area range, select a computing power node in each clock synchronization domain as the master clock node, take its clock status as the master clock, and generate the master clock timestamp. The clock synchronization path generation unit is used to plan the optimal clock synchronization path starting from the master clock node and generate the optimal clock synchronization path identifier. The resource coordination and time calibration unit is used to encapsulate the optimal task data transmission path, the master clock timestamp, and the optimal clock synchronization path identifier, and forward them to all other computing power nodes in the clock synchronization domain using the master clock node, thereby achieving resource coordination and time synchronization.

10. A computer device, characterized in that, The computer device includes: a memory, a processor, and a computer program stored in the memory that can be run on the processor, wherein the processor executes the computer program to implement the method according to any one of claims 1 to 8.

Images