Multi-VRRP group and ECMP coordinated load adjustment method and system
By coordinating multiple VRRP groups with ECMP, a ring primary-backup topology is constructed to achieve cross-VRRP group state synchronization and dynamic weight adjustment. This solves the problems of VRRP resource idleness and ECMP load imbalance, thereby improving network resource utilization and the service quality of critical services.
Patent Information
- Application Number
- CN202511747516.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-26
- Publication Date
- 2026-02-24
AI Technical Summary
In the traditional VRRP protocol, the Backup router resources are idle, ECMP cannot detect changes in link status, resulting in load imbalance, which cannot guarantee the SLA of critical services and cannot dynamically adjust path weights.
By collaborating with multiple VRRP groups and ECMP, a ring-shaped primary-backup topology is constructed to achieve cross-VRRP group state synchronization. Service types are classified based on the DSCP field, and the weights of ECMP equivalent paths are dynamically adjusted through a dynamic weight model and differentiated traffic scheduling strategies.
This improved resource utilization, ensured high-quality network services for critical businesses, reduced packet loss and latency jitter, and doubled network capacity while minimizing the impact of faults.
Smart Images
Figure CN121567642A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of network communication technology, and in particular relates to a load adjustment method and system for multi-VRRP groups and ECMP coordination. Background Technology
[0002] In traditional network architectures, there is a fundamental contradiction between VRRP primary / backup redundancy and ECMP load balancing. VRRP backup paths are idle for a long time, resulting in insufficient bandwidth utilization. ECMP cannot perceive service priorities, and critical services (such as industrial control) are mixed with ordinary traffic. Link failure switching relies on ARP aging, which cannot meet the sub-second switching requirements of scenarios such as finance and healthcare. New scenarios such as 5G and industrial internet require networks to have both ultra-high reliability and intelligent load balancing capabilities.
[0003] In summary, the existing ECMP and VRRP technologies have the following problems: In the traditional VRRP protocol, all traffic is processed by the Master router, while the Backup router is only in a listening state. This results in the Backup device's CPU, memory, bandwidth and other resources being completely idle during non-fault periods, leading to a waste of resources. The VRRP protocol lacks load balancing capabilities and cannot dynamically switch the Master role or allocate traffic based on real-time load, which makes high-priority but low-performance devices potentially become performance bottlenecks. ECMP has limitations; its static hashing mechanism cannot detect changes in link state, such as congestion or failure, which leads to problems such as load imbalance and long-tail latency. ECMP technology lacks sufficient support for business priority and cannot allocate the optimal path for critical business, thus making it difficult to guarantee the Service Level Agreement (SLA). In ECMP, the network and service requirements are disconnected, and the path weights cannot be dynamically adjusted according to real-time network conditions, resulting in low resource allocation efficiency. Summary of the Invention
[0004] To address the shortcomings of existing technologies, this invention provides a load adjustment method that coordinates multiple VRRP groups with ECMP.
[0005] Firstly, the method includes, Multiple ECMP equivalent paths are generated based on the ECMP protocol principle, and a corresponding VRRP group is created for each ECMP equivalent path. The mapping relationship between ECMP equivalent paths and VRRP groups is established, and a ring primary-backup topology is constructed among multiple virtual routers. Based on the extended routing protocol, cross-VRRP group state synchronization is achieved, and a global VRRP state information table is generated; Establish a bidirectional forwarding detection session on each ECMP equivalent path and periodically collect quality parameters from the global VRRP state information table; Construct a dynamic weight model, and based on the collected quality parameters, calculate and normalize the weight values of each ECMP equivalent path. Detect business traffic, identify DSCP fields, and classify business types; Based on the weight values and service types of each ECMP equivalent path, a differentiated traffic scheduling strategy is established to dynamically adjust the ECMP equivalent paths.
[0006] Furthermore, the implementation of cross-VRRP group state synchronization according to the extended routing protocol specifically includes defining a VRRP-specific TLV in the Opaque LSA of the OSPFv2 / OSPFv3 protocol.
[0007] Furthermore, the generation of the global VRRP status information table also includes an aging mechanism in which the global VRRP status information table adopts an aging mechanism. If no update on the status of a VRID group is received within a certain notification period, the entry is marked as suspicious. If no update is received in the next period, the VRID group entry is deleted.
[0008] Furthermore, the quality parameters include: link latency value, available bandwidth value, path packet loss rate, and BFD session status.
[0009] Furthermore, the process of detecting service traffic, identifying the DSCP field, and classifying service types specifically includes performing deep packet inspection on service traffic to identify the DSCP value of the ToS field in the IP header; where DSCP=46, it is marked as deterministic service; DSCP=32, it is marked as latency-sensitive service; and DSCP=0, it is marked as best-effort service; for unmarked service traffic, it is mapped to the above three types of service roaming according to the source / destination IP and port policy.
[0010] Furthermore, the establishment of a differentiated traffic scheduling strategy based on the weight values and service types of each ECMP equivalent path, and the dynamic adjustment of the ECMP equivalent paths, specifically includes: For deterministic business operations, the ECMP equivalent path with the highest ECMP equivalent path weight is strongly bound to ensure low packet loss and low jitter. For latency-sensitive services, the ECMP equivalent path with the lowest link latency is dynamically selected to ensure latency sensitivity. For best-effort operations, the normalized ECMP equivalent path weights are used as the basis for traffic allocation ratios, and weighted round-robin distribution is performed.
[0011] Furthermore, the method also includes fault handling and recovery by defining a three-level fault circuit interruption mechanism.
[0012] Secondly, the system includes: a topology generation module, a status information generation module, a data acquisition module, a weight calculation module, a service partitioning module, and a dynamic adjustment module; The topology generation module is used to generate multiple ECMP equivalent paths according to the ECMP protocol principle, create a corresponding VRRP group for each ECMP equivalent path, establish a mapping relationship between ECMP equivalent paths and VRRP groups, and build a ring primary-backup topology structure among multiple virtual routers. The state information generation module is used to realize cross-VRRP group state synchronization according to the extended routing protocol and generate a global VRRP state information table. The acquisition module is used to establish a bidirectional forwarding detection session on each ECMP equivalent path and periodically collect quality parameters from the global VRRP status information table. The weight calculation module is used to construct a dynamic weight model and, based on the collected quality parameters, calculate and construct the dynamic weight model to obtain and normalize the weight values of each ECMP equivalent path. The service segmentation module is used to detect service traffic, identify the DSCP field, and segment service types. The dynamic adjustment module is used to establish a differentiated traffic scheduling strategy based on the weight value and service type of each ECMP equivalent path, and to dynamically adjust the ECMP equivalent path.
[0013] Furthermore, the status information generation module is specifically used to define a VRRP-specific TLV in the Opaque LSA of the OSPFv2 / OSPFv3 protocol.
[0014] Furthermore, the status information generation module is also used to ensure that the global VRRP status information table adopts an aging mechanism. If no update on the status of a VRID group is received within a certain notification period, the entry is marked as suspicious, and if no update is received in the next period, the VRID group entry is deleted.
[0015] Furthermore, the quality parameters include: link latency value, available bandwidth value, path packet loss rate, and BFD session status.
[0016] Furthermore, the service segmentation module is specifically used to perform deep packet inspection on service traffic and identify the DSCP value of the ToS field in the IP header; wherein, when DSCP=46, it is marked as deterministic service; when DSCP=32, it is marked as latency-sensitive service; when DSCP=0, it is marked as best-effort service; for unmarked service traffic, it is mapped to the above three types of service roaming according to the source / destination IP and port policy.
[0017] Furthermore, the dynamic adjustment module is specifically used for, For deterministic business operations, the ECMP equivalent path with the highest ECMP equivalent path weight is strongly bound to ensure low packet loss and low jitter. For latency-sensitive services, the ECMP equivalent path with the lowest link latency is dynamically selected to ensure latency sensitivity. For best-effort operations, the normalized ECMP equivalent path weights are used as the basis for traffic allocation ratios, and weighted round-robin distribution is performed.
[0018] Furthermore, the system also includes a fault handling module, which is used to handle and recover from faults by defining a three-level fault circuit interruption mechanism.
[0019] Compared with the prior art, the present invention has the following advantages: 1. This invention utilizes a multi-VRRP group ring backup architecture, in which all routers participate in service forwarding, increasing resource utilization from less than 50% in traditional VRRP to 100%, doubling network capacity with the same hardware investment; based on DSCP's refined service classification and differentiated scheduling strategy, deterministic services can obtain exclusive guarantee of the path with the highest weight, while latency-sensitive services always select the optimal latency path, and the service-level packet loss rate can be controlled below 0.001%, with latency jitter less than 5 milliseconds.
[0020] 2. Achieve a scheduling shift from "equal distribution" to "value-first" scheduling. Based on a comprehensive weighted calculation model of service priority, latency, and bandwidth, paths are evaluated and selected. This scheduling strategy, which comprehensively considers multiple key factors, can better guarantee the SLA of critical services and ensure that important services receive higher-quality network services. 3. A three-level fault circuit interruption mechanism is proposed to actively monitor the link status and adjust the ECMP weight or VRRP group policy in a timely manner, thereby actively avoiding faulty links and preventing service interruption.
[0021] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description
[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0023] Figure 1 A schematic diagram of the load adjustment method for multi-VRRP groups and ECMP coordination according to the present invention is shown.
[0024] Figure 2 A schematic diagram of a load adjustment system module for multi-VRRP groups and ECMP collaboration according to the present invention is shown.
[0025] Figure 3 The diagram illustrates a ring master-slave topology constructed in a specific embodiment of the present invention.
[0026] Figure 4 This illustrates a specific ring master-slave topology in an embodiment of the present invention. Detailed Implementation
[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0028] like Figure 1 As shown, this invention discloses a load balancing method that coordinates multiple VRRP groups and ECMP. By deploying a multi-VRRP group policy and linking it with routing protocols, and combining ECMP with a ring-shaped mutual backup strategy among multiple VRRP groups, a multi-path dynamic load balancing system is constructed. Based on VRRP group status information and service priorities, the weights of ECMP equal-cost paths are dynamically adjusted. Combined with a VRRP protocol-level fault isolation mechanism, the impact of ECMP equal-cost path failures on the network is reduced. Furthermore, a differentiated traffic scheduling strategy is used to achieve fine-grained classification of service types, ensuring the SLA of critical services. The steps include... S1. Based on the ECMP protocol principle, multiple ECMP equivalent paths are generated, and a corresponding VRRP group is created for each ECMP equivalent path. The mapping relationship between ECMP equivalent paths and VRRP groups is established, and a ring primary-backup topology is constructed among N virtual routers.
[0029] In this embodiment, as Figure 3 As shown, N VRRP backup groups are configured on N physical routers to form a logical ring master-slave topology. The i-th router is configured as the master router in the i-th VRRP group and as the backup router in the (i-1)-th VRRP group. The value of i ranges from 1 to N, and the 0th VRRP group is the Nth VRRP group, ensuring that the forwarding plane of all routers is activated.
[0030] Optionally, the constructed ring master-slave topology also includes the following configuration constraints: a single router cannot simultaneously serve as the Master role in multiple VRRP groups; if a configuration conflict occurs, a verification error will be triggered and the router will be refused to send a message; each VRRP group must contain at least one Master and one Backup role, and the Backup priority must be at least 30 points lower than the Master priority to ensure switching stability; the VRRP advertisement interval is uniformly configured to 1 second, and the preempt mode is disabled by default to prevent frequent switching.
[0031] In this embodiment, in the ring primary-backup topology, each VRRP group is abstracted as a logical node of ECMP, and the next-hop address of the ECMP equivalent path is configured as the virtual IP address of the VRRP group; and the VRRP virtual IP address is broadcast to the entire network through the free ARP mechanism to build a dynamic mapping table from virtual IP to physical path, wherein the dynamic mapping table includes VRID, virtual IP, physical IP of the primary router and interface index.
[0032] S2. Based on the extended routing protocol, achieve cross-VRRP group state synchronization and generate a global VRRP state information table.
[0033] In this embodiment, the present invention defines a VRRP-specific TLV (Type=0x0B01) in the Opaque LSA (Type-11) of the OSPFv2 / OSPFv3 protocol; wherein, the VRRP-specific TLV includes a VRID field (2 bytes), a virtual IP address field (4 bytes), a VRRP priority field (1 byte), a VRRP status field (1 byte), and a checksum field (2 bytes), as follows. Figure 4 As shown, each router periodically broadcasts its own VRRP status information through this extended LSA message, and the receiving router parses the TLV and updates its locally maintained global VRRP status information table.
[0034] In this embodiment, to prevent inconsistent VRRP status information, when the same VRRP group receives Master status announcements from multiple routers, the one with the highest priority is used. If the priorities are the same, the one with the larger Router ID is used. The global VRRP status information table adopts an aging mechanism. If no update to the status of a VRID group is received within 3 announcement periods, the entry is marked as suspicious. If no update is received in the 4th period, the VRID group entry is deleted.
[0035] S3. Establish a bidirectional forwarding detection (BFD) session on each ECMP equivalent path and periodically collect quality parameters from the global VRRP status information table. The quality parameters include: link delay RTT value (measured by BFD echo function), available bandwidth B value (calculated by interface counter, total bandwidth minus used bandwidth), path packet loss rate (statistical on BFD packet loss), and BFD session status (UP / DOWN / ADMIN_DOWN).
[0036] S4. Construct a dynamic weight model, and based on the collected quality parameters, calculate and normalize the weight values of each ECMP equivalent path, which are expressed as follows: ECMP weight = α × (Priority - 1 / 254) + β × (1 - RTT / RTT_max) + γ × (B / Bt) Where α, β, and γ are adjustment coefficients (α+β+γ=1). α is sensitive to service type and applies to different priority services; β is sensitive to latency and applies to latency requirements; γ is sensitive to bandwidth and applies to bandwidth requirements. Recommended configurations are α = 0.5±0.1, β = 0.3±0.1, and γ = 0.2±0.1. Priority is the VRRP Master route priority in the ECMP equivalent path. RTT is the current latency, obtained through BFD detection. RTT_max is the maximum path latency, with an initial recommended value of 50±5ms. B is the available bandwidth, obtained through BFD detection, calculated as total bandwidth minus the already occupied bandwidth. Bt is the total bandwidth; interface bandwidth can be directly viewed.
[0037] Optionally, for deterministic services (DE), the adjustment coefficient α is increased to 0.7, β is decreased to 0.2, and γ is decreased to 0.1 to strengthen priority weight; for latency-sensitive services (LS), the adjustment coefficient β is increased to 0.5, α is decreased to 0.3, and γ remains at 0.2 to strengthen latency weight; for best-effort services (BE), the default coefficients α=0.5, β=0.3, and γ=0.2 are used to balance the three.
[0038] S5. Detect business traffic, identify the DSCP field, and classify business types.
[0039] In this embodiment, deep packet inspection is performed on traffic at the core router ingress to identify the DSCP value of the ToS field in the IP header. When DSCP=46 (DE class), it is marked as a deterministic service (DE); when DSCP=32 (LS class), it is marked as a latency-sensitive service (LS); when DSCP=0 (BE class), it is marked as a best-effort service (BE). For unmarked traffic, it is mapped to the above three types of services according to the source / destination IP and port policy.
[0040] S6. Based on the weight values and service types of each ECMP equivalent path, establish a differentiated traffic scheduling strategy and dynamically adjust the ECMP equivalent paths.
[0041] In this embodiment, the present invention establishes a differentiated traffic scheduling strategy based on the weight values of each ECMP equivalent path obtained in step S4, specifically including: 1. For DE-type services (DSCP=46, such as voice / video), strongly bind the highest weight ECMP equivalent path to ensure low packet loss and low jitter; 2. For LS-type services (DSCP=32, such as real-time data), dynamically select the ECMP equivalent path with the lowest RTT to ensure latency sensitivity. 3. For BE-type services (DSCP=0, normal traffic), the normalized ECMP equivalent path weight is used as the basis for traffic allocation ratio, and weighted round-robin (WRR) distribution is performed.
[0042] In another embodiment of the invention, the invention further defines a three-level fault circuit interruption mechanism for fault handling and recovery, thereby minimizing the impact of the fault; wherein, 1. Level 1 Fault: When the single-path packet loss rate increases beyond the threshold, ECMP weight downgrade is triggered, reducing the VRRP group priority of the faulty link, decreasing the ECMP weight, and reducing the forwarding of single-faulty path traffic. 2. Level 2 Fault: When the BFD session goes down, VRRP primary / backup switch is triggered. Traffic on the original path is temporarily allocated to the highest priority Backup device to ensure that traffic continues to be forwarded. During the next weight allocation, traffic on the original path will be allocated to other normal paths according to ECMP weights. No traffic will be allocated during the fault. 3. Level 3 Fault: When all devices in a single VRRP group are down, ECMP equal path recalculation is immediately triggered, multi-VRRP group policies are reissued, traffic is redistributed, alarms and logs are generated, and the network administrator is notified to handle the situation.
[0043] In another embodiment of the invention, Figure 4 Taking the ring master-slave topology shown as an example, 1. When the packet loss rate of Path1 increases to over 10%, a Level 1 fault alarm is triggered, and fault handling and recovery are performed. Fault handling includes: a. Adjust the route priority of VRRP group 1, reduce the priority of the Master route of VRRP group 1 to 100, and the priority of the Backup route to 50; b. The next cycle triggers ECMP weight calculation, the network topology remains unchanged, and the ECMP path remains unchanged; c. Distribute traffic to Path1, Path2, Path3, and Path4 according to ECMP weights; Fault recovery includes, a. Once the V-router1 link stabilizes, adjust the priority of the VRRP group 1 Master route to 150 and the Backup route to 100; b. In the next cycle of ECMP weight calculation, the ECMP weights of the four paths will be normalized; c. Distribute traffic to Path1, Path2, Path3, and Path4 according to the normalization ratio; 2. When Path1 link fails, a secondary fault alarm is triggered, and fault handling and recovery are performed. Fault handling includes: a. When VRRP group 1 performs a master / standby switch, V-router2 is both the master route for VRRP group 1 and the master route for VRRP group 2.
[0044] bV-router2 temporarily takes over all traffic forwarding for Path1; c. In the next cycle of ECMP weight calculation, VRRP group 1 does not process traffic, and performs ECMP weight normalization calculation on the other three paths. The network topology and ECMP paths remain unchanged. d. Distribute the traffic from Path1 to the three paths Path2, Path3, and Path4 according to a differentiated scheduling strategy; Fault recovery includes, After aV-router1 is restored, V-router2 switches to Backup; b. The next cycle triggers ECMP weight calculation, normalizing the ECMP weights of the four paths; c. Distribute traffic to Path1, Path2, Path3, and Path4 according to the differentiated scheduling strategy; 3. When a single VRRP group goes down (i.e., v-router1 and v-router2 both go down), a level 3 fault alarm is triggered, and fault handling and recovery are performed. Fault handling includes... a. IP address lost due to a single VRRP group going down; b. ECMP node failure, resulting in network topology changes; c. Immediately trigger ECMP path recalculation; d. Generate a new ECMP path, and then proceed to steps S1-S6; Fault recovery includes, a. Faulty equipment recovery, network topology changes, ECMP path calculation; b. Generate a new ECMP path, and then proceed to steps S1-S6.
[0045] like Figure 2 As shown, the present invention discloses a load adjustment system for multi-VRRP groups and ECMP collaboration, which includes a topology generation module, a status information generation module, an acquisition module, a weight calculation module, a service partitioning module, and a dynamic adjustment module.
[0046] 1. The topology generation module generates multiple ECMP equivalent paths based on the ECMP protocol principle, creates a corresponding VRRP group for each ECMP equivalent path, establishes the mapping relationship between ECMP equivalent paths and VRRP groups, and constructs a ring primary-backup topology among N virtual routers.
[0047] In this embodiment, as Figure 3 As shown, N VRRP backup groups are configured on N physical routers to form a logical ring master-slave topology. The i-th router is configured as the master router in the i-th VRRP group and as the backup router in the (i-1)-th VRRP group. The value of i ranges from 1 to N, and the 0th VRRP group is the Nth VRRP group, ensuring that the forwarding plane of all routers is activated.
[0048] Optionally, the constructed ring master-slave topology also includes the following configuration constraints: a single router cannot simultaneously serve as the Master role in multiple VRRP groups; if a configuration conflict occurs, a verification error will be triggered and the router will be refused to send a message; each VRRP group must contain at least one Master and one Backup role, and the Backup priority must be at least 30 points lower than the Master priority to ensure switching stability; the VRRP advertisement interval is uniformly configured to 1 second, and the preempt mode is disabled by default to prevent frequent switching.
[0049] In this embodiment, in the ring primary-backup topology, each VRRP group is abstracted as a logical node of ECMP, and the next-hop address of the ECMP equivalent path is configured as the virtual IP address of the VRRP group; and the VRRP virtual IP address is broadcast to the entire network through the free ARP mechanism to build a dynamic mapping table from virtual IP to physical path, wherein the dynamic mapping table includes VRID, virtual IP, physical IP of the primary router and interface index.
[0050] 2. The status information generation module is used to realize cross-VRRP group status synchronization according to the extended routing protocol and generate a global VRRP status information table.
[0051] In this embodiment, the present invention defines a VRRP-specific TLV (Type=0x0B01) in the Opaque LSA (Type-11) of the OSPFv2 / OSPFv3 protocol; wherein, the VRRP-specific TLV includes a VRID field (2 bytes), a virtual IP address field (4 bytes), a VRRP priority field (1 byte), a VRRP status field (1 byte), and a checksum field (2 bytes), as follows. Figure 4 As shown, each router periodically broadcasts its own VRRP status information through this extended LSA message, and the receiving router parses the TLV and updates its locally maintained global VRRP status information table.
[0052] In this embodiment, to prevent inconsistent VRRP status information, when the same VRRP group receives Master status announcements from multiple routers, the one with the highest priority is used. If the priorities are the same, the one with the larger Router ID is used. The global VRRP status information table adopts an aging mechanism. If no update for a certain VRID is received within 3 announcement periods, the entry is marked as suspicious. If no update is received in the 4th period, the entry is deleted.
[0053] 3. The acquisition module is used to establish a bidirectional forwarding detection (BFD) session on each ECMP equivalent path and periodically collect quality parameters from the global VRRP status information table. The quality parameters include: link delay RTT value (measured by BFD echo function), available bandwidth B value (calculated by interface counter, total bandwidth minus used bandwidth), path packet loss rate (statistical on BFD packet loss), and BFD session status (UP / DOWN / ADMIN_DOWN).
[0054] 4. The weight calculation module is used to construct a dynamic weight model. Based on the collected quality parameters, it calculates and constructs the dynamic weight model, obtains and normalizes the weight values of each ECMP equivalent path, which are expressed as follows: ECMP weight = α × (Priority - 1 / 254) + β × (1 - RTT / RTT_max) + γ × (B / Bt) Where α, β, and γ are adjustment coefficients (α+β+γ=1). α is sensitive to service type and applies to different priority services; β is sensitive to latency and applies to latency requirements; γ is sensitive to bandwidth and applies to bandwidth requirements. Recommended configurations are α = 0.5±0.1, β = 0.3±0.1, and γ = 0.2±0.1. Priority is the VRRP Master route priority in the ECMP equivalent path. RTT is the current latency, obtained through BFD detection. RTT_max is the maximum path latency, with an initial recommended value of 50±5ms. B is the available bandwidth, obtained through BFD detection, calculated as total bandwidth minus the already occupied bandwidth. Bt is the total bandwidth; interface bandwidth can be directly viewed.
[0055] Optionally, for deterministic services (DE), the adjustment coefficient α is increased to 0.7, β is decreased to 0.2, and γ is decreased to 0.1 to strengthen priority weight; for latency-sensitive services (LS), the adjustment coefficient β is increased to 0.5, α is decreased to 0.3, and γ remains at 0.2 to strengthen latency weight; for best-effort services (BE), the default coefficients α=0.5, β=0.3, and γ=0.2 are used to balance the three.
[0056] 5. The business segmentation module is used to detect business traffic, identify the DSCP field, and classify business types.
[0057] In this embodiment, deep packet inspection is performed on traffic at the core router ingress to identify the DSCP value of the ToS field in the IP header. When DSCP=46 (DE class), it is marked as a deterministic service (DE); when DSCP=32 (LS class), it is marked as a latency-sensitive service (LS); when DSCP=0 (BE class), it is marked as a best-effort service (BE). For unmarked traffic, it is mapped to the above three types of services according to the source / destination IP and port policy.
[0058] 6. The dynamic adjustment module is used to establish differentiated traffic scheduling strategies based on the weight values and service types of each ECMP equivalent path, and to dynamically adjust the ECMP equivalent paths.
[0059] In this embodiment, the present invention establishes a differentiated traffic scheduling strategy based on the weight values of each ECMP equivalent path obtained in the weight calculation module, specifically including: For DE-type services (DSCP=46, such as voice / video), the highest weight ECMP equivalent path is strongly bound to ensure low packet loss and low jitter. For LS-type services (DSCP=32, such as real-time data), dynamically select the ECMP equivalent path with the lowest RTT to ensure latency sensitivity. For BE-type services (DSCP=0, normal traffic), the normalized ECMP equivalent path weight is used as the basis for traffic allocation ratio, and weighted round-robin (WRR) distribution is performed.
[0060] In another embodiment of the invention, the system further includes a fault handling module, which minimizes the impact of faults by defining a three-level fault circuit interruption mechanism; wherein, 1. Level 1 Fault: When the single-path packet loss rate increases beyond the threshold, ECMP weight downgrade is triggered, reducing the VRRP group priority of the faulty link, decreasing the ECMP weight, and reducing the forwarding of single-faulty path traffic. 2. Level 2 Fault: When the BFD session goes down, VRRP primary / backup switch is triggered. Traffic on the original path is temporarily allocated to the highest priority Backup device to ensure that traffic continues to be forwarded. During the next weight allocation, traffic on the original path will be allocated to other normal paths according to ECMP weights. No traffic will be allocated during the fault. 3. Level 3 Fault: When all devices in a single VRRP group are down, ECMP equal path recalculation is immediately triggered, multi-VRRP group policies are reissued, traffic is redistributed, alarms and logs are generated, and the network administrator is notified to handle the situation.
[0061] In another embodiment of the invention, Figure 4 Taking the ring master-slave topology shown as an example, 1. When the packet loss rate of Path1 increases to over 10%, a Level 1 fault alarm is triggered, and fault handling and recovery are performed. Fault handling includes: a. Adjust the route priority of VRRP group 1, reduce the priority of the Master route of VRRP group 1 to 100, and the priority of the Backup route to 50; b. The next cycle triggers ECMP weight calculation, the network topology remains unchanged, and the ECMP path remains unchanged; c. Distribute traffic to Path1, Path2, Path3, and Path4 according to ECMP weights; Fault recovery includes, a. Once the V-router1 link stabilizes, adjust the priority of the VRRP group 1 Master route to 150 and the Backup route to 100; b. In the next cycle of ECMP weight calculation, the ECMP weights of the four paths will be normalized; c. Distribute traffic to Path1, Path2, Path3, and Path4 according to the normalization ratio; 2. When Path1 link fails, a secondary fault alarm is triggered, and fault handling and recovery are performed. Fault handling includes: a. When VRRP group 1 performs a master / standby switch, V-router2 is both the master route for VRRP group 1 and the master route for VRRP group 2.
[0062] bV-router2 temporarily takes over all traffic forwarding for Path1; c. In the next cycle of ECMP weight calculation, VRRP group 1 does not process traffic, and performs ECMP weight normalization calculation on the other three paths. The network topology and ECMP paths remain unchanged. d. Distribute the traffic from Path1 to the three paths Path2, Path3, and Path4 according to a differentiated scheduling strategy; Fault recovery includes, After aV-router1 is restored, V-router2 switches to Backup; b. The next cycle triggers ECMP weight calculation, normalizing the ECMP weights of the four paths; c. Distribute traffic to Path1, Path2, Path3, and Path4 according to the differentiated scheduling strategy; 3. When a single VRRP group goes down (i.e., v-router1 and v-router2 both go down), a level 3 fault alarm is triggered, and fault handling and recovery are performed. Fault handling includes... a. IP address lost due to a single VRRP group going down; b. ECMP node failure, resulting in network topology changes; c. Immediately trigger ECMP path recalculation; d. Generate a new ECMP path and then dynamically adjust it; Fault recovery includes, a. Faulty equipment recovery, network topology changes, ECMP path calculation; b. Generate a new ECMP path and then dynamically adjust it.
[0063] The foregoing description and accompanying drawings fully illustrate embodiments of the invention to enable those skilled in the art to practice them. Other embodiments may include structural and other changes. The embodiments represent only possible variations. Individual components and functions are optional unless explicitly required, and the order of operation may vary. Some portions and features of some embodiments may be included or substituted for portions and features of other embodiments. Embodiments of the invention are not limited to the structures described above and shown in the accompanying drawings, and various modifications and changes may be made without departing from their scope. The scope of the invention is limited only by the appended claims.
Claims
1. A load adjustment method for multi-VRRP groups and ECMP coordination, characterized in that, The method includes, Multiple ECMP equivalent paths are generated based on the ECMP protocol principle, and a corresponding VRRP group is created for each ECMP equivalent path. The mapping relationship between ECMP equivalent paths and VRRP groups is established, and a ring primary-backup topology is constructed among multiple virtual routers. Based on the extended routing protocol, cross-VRRP group state synchronization is achieved, and a global VRRP state information table is generated; Establish a bidirectional forwarding detection session on each ECMP equivalent path and periodically collect quality parameters from the global VRRP state information table; Construct a dynamic weight model, and based on the collected quality parameters, calculate and normalize the weight values of each ECMP equivalent path. Detect business traffic, identify DSCP fields, and classify business types; Based on the weight values and service types of each ECMP equivalent path, a differentiated traffic scheduling strategy is established to dynamically adjust the ECMP equivalent paths.
2. The load adjustment method for multi-VRRP groups and ECMP coordination according to claim 1, characterized in that, The method of achieving cross-VRRP group state synchronization according to the extended routing protocol specifically includes defining a VRRP-specific TLV in the Opaque LSA of the OSPFv2 / OSPFv3 protocol.
3. The load adjustment method for multi-VRRP groups and ECMP coordination according to claim 1, characterized in that, The generation of the global VRRP status information table also includes an aging mechanism in which the global VRRP status information table adopts an aging mechanism. If no update on the status of a VRID group is received within a certain notification period, the VRID group is marked as suspicious. If no update is received in the next period, the entry for the VRID group is deleted.
4. The load adjustment method for multi-VRRP groups and ECMP coordination according to claim 1, characterized in that, The quality parameters include: link latency, available bandwidth, path packet loss rate, and BFD session status.
5. The load adjustment method for multi-VRRP groups and ECMP coordination according to claim 1, characterized in that, The process of detecting service traffic, identifying the DSCP field, and classifying service types specifically includes performing deep packet inspection on service traffic to identify the DSCP value of the ToS field in the IP header; where DSCP=46, it is marked as deterministic service; DSCP=32, it is marked as latency-sensitive service; and DSCP=0, it is marked as best-effort service; for unmarked service traffic, it is mapped to the above three service types according to the source / destination IP and port policy.
6. The load adjustment method for multi-VRRP groups and ECMP coordination according to claim 1, characterized in that, The aforementioned method establishes a differentiated traffic scheduling strategy based on the weight values and service types of each ECMP equivalent path, dynamically adjusting the ECMP equivalent paths. Specifically, this includes... For deterministic business operations, the ECMP equivalent path with the highest ECMP equivalent path weight is strongly bound to ensure low packet loss and low jitter. For latency-sensitive services, the ECMP equivalent path with the lowest link latency is dynamically selected to ensure latency sensitivity. For best-effort operations, the normalized ECMP equivalent path weights are used as the basis for traffic allocation ratios, and weighted round-robin distribution is performed.
7. The load adjustment method for multi-VRRP groups and ECMP coordination according to claim 1, characterized in that, The method also includes fault handling and recovery by defining a three-level fault circuit interruption mechanism.
8. A load conditioning system that coordinates multiple VRRP groups with ECMP, characterized in that, The system includes: a topology generation module, a status information generation module, a data acquisition module, a weight calculation module, a service partitioning module, and a dynamic adjustment module; The topology generation module is used to generate multiple ECMP equivalent paths according to the ECMP protocol principle, create a corresponding VRRP group for each ECMP equivalent path, establish a mapping relationship between ECMP equivalent paths and VRRP groups, and build a ring primary-backup topology structure among multiple virtual routers. The state information generation module is used to realize cross-VRRP group state synchronization according to the extended routing protocol and generate a global VRRP state information table. The acquisition module is used to establish a bidirectional forwarding detection session on each ECMP equivalent path and periodically collect quality parameters from the global VRRP status information table. The weight calculation module is used to construct a dynamic weight model and, based on the collected quality parameters, calculate and construct the dynamic weight model to obtain and normalize the weight values of each ECMP equivalent path. The service segmentation module is used to detect service traffic, identify the DSCP field, and segment service types. The dynamic adjustment module is used to establish a differentiated traffic scheduling strategy based on the weight value and service type of each ECMP equivalent path, and to dynamically adjust the ECMP equivalent path.
9. The load adjustment system with multiple VRRP groups and ECMP coordination according to claim 8, characterized in that, The status information generation module is specifically used to define a VRRP-specific TLV in the Opaque LSA of the OSPFv2 / OSPFv3 protocol.
10. The load adjustment system with multiple VRRP groups and ECMP coordination according to claim 8, characterized in that, The status information generation module is further configured to use an aging mechanism in the global VRRP status information table. If no update on the status of a VRID group is received within a certain notification period, the VRID group is marked as suspicious. If no update is received in the next period, the entry for the VRID group is deleted.
11. The load adjustment system with multiple VRRP groups and ECMP coordination according to claim 8, characterized in that, The quality parameters include: link latency, available bandwidth, path packet loss rate, and BFD session status.
12. The load adjustment system with multiple VRRP groups and ECMP coordination according to claim 8, characterized in that, The service segmentation module is specifically used to perform deep packet inspection on service traffic and identify the DSCP value of the ToS field in the IP header. Specifically, when DSCP=46, it is marked as deterministic service; when DSCP=32, it is marked as latency-sensitive service; when DSCP=0, it is marked as best-effort service. For unmarked service traffic, it is mapped to the above three types of services according to the source / destination IP and port policy.
13. The load adjustment system with multiple VRRP groups and ECMP coordination according to claim 8, characterized in that, The dynamic adjustment module is specifically used for, For deterministic business operations, the ECMP equivalent path with the highest ECMP equivalent path weight is strongly bound to ensure low packet loss and low jitter. For latency-sensitive services, the ECMP equivalent path with the lowest link latency is dynamically selected to ensure latency sensitivity. For best-effort operations, the normalized ECMP equivalent path weights are used as the basis for traffic allocation ratios, and weighted round-robin distribution is performed.
14. The load adjustment system with multiple VRRP groups and ECMP coordination according to claim 9, characterized in that, The system also includes a fault handling module, which is used to handle and recover from faults by defining a three-level fault circuit interruption mechanism.