A method and system for optimizing SRv6 computing power network path based on multi-dimensional dynamic scoring

Abstract

A kind of SRv6 computing power network path optimization method and system based on multidimensional dynamic score belongs to communication network technical field, including by real-time acquisition node load, link congestion, geographical distance, topological hop number and service type, construct the multidimensional dynamic score system of load score, delay score and service priority score;Based on DMPSC-T table (including Last_Update_Time / Expire_Time timestamp mechanism) dynamically select optimal path;When detecting more optimal path, the SID dynamic replacement of End.Redirect instruction defined newly by SRv6 is triggered: Endpoint node directly replaces the destination SID in SegmentList, and transit node is redirected by realizing ACL rule through PBR generation.This application can realize millisecond level path switching, effectively improve high priority service quality, and optimize network resource global scheduling efficiency.

Description

Technical Field

[0001] This application relates to the field of communication networks, specifically to an SRv6 computing power network path optimization method and system based on multi-dimensional dynamic scoring. Background Technology

[0002] With the rapid development of emerging application scenarios such as 5G network slicing, industrial internet, and autonomous driving, computing networks, as the core of new network architectures, are placing higher demands on the intelligence and refinement of network path selection. SRv6 (Segment Routing over IPv6), as a next-generation network protocol, has become a key technology for computing network path selection due to its flexible path control capabilities and simplified network architecture. However, current path selection methods for SRv6 computing networks still have significant limitations, mainly in the following four aspects: Static strategies lack flexibility: Existing SRv6 path selection schemes mainly rely on pre-configured static strategies or simple factor metrics (such as shortest path algorithms, bandwidth, and latency), lacking the ability to respond in real time to dynamic network changes. When the network experiences sudden traffic surges, node overload, or link congestion, traditional methods cannot adjust paths in a timely manner, leading to uneven network resource utilization and significant fluctuations in service quality. Especially in high-demand scenarios such as 5G network slicing and the Industrial Internet, static strategies are difficult to adapt to the sudden and dynamic changes in business traffic, resulting in wasted network resources and a degraded user experience.

[0003] Insufficient multi-objective optimization: Current path optimization methods typically focus only on a single metric (such as shortest hop count or maximum bandwidth), failing to effectively integrate multi-dimensional factors such as load balancing, transmission efficiency (distance), and service priority (QoS). This single-objective optimization approach cannot meet the needs of complex service scenarios in computing networks. For example, in autonomous driving scenarios, both low latency and high reliability must be ensured, and optimizing a single metric is insufficient to achieve optimal service quality. Furthermore, the lack of a quantitative and collaborative optimization mechanism for multi-dimensional factors leads to unreasonable allocation of network resources, resulting in an overall service quality that cannot reach its optimal level.

[0004] Lack of Dynamic Adaptability: Traditional SRv6 computing networks lack dynamic awareness and real-time adjustment capabilities in path selection. Intermediate nodes cannot perceive critical information such as node load, link congestion, or geographical distance in real time, leading to delayed path optimization. When network conditions change, the controller must recalculate and issue paths, a process with significant latency that cannot meet the demands of high real-time services. In scenarios such as the Industrial Internet, this lack of dynamic adaptability increases control command transmission latency, impacting system stability and reliability.

[0005] Resource contention and lack of service differentiation: Existing traffic engineering (TE) strategies treat all services equally, employing a homogeneous path selection approach that fails to quantify service priorities. High-priority services (such as autonomous driving commands and industrial control commands) cannot preempt network resources during periods of scarcity, while low-priority services (such as ordinary file downloads) may consume significant bandwidth, leading to a decline in the quality of critical service delivery. This indiscriminate approach fails to meet the needs of differentiated services in computing networks, limiting the network's ability to support high-value services.

[0006] In summary, existing path selection methods for SRv6 computing networks have significant shortcomings in terms of dynamic adaptability, multi-objective optimization, and differentiated service provision, making it difficult to meet the stringent network performance requirements of demanding scenarios such as 5G network slicing, industrial internet, and autonomous driving. There is an urgent need to propose a path optimization method that can comprehensively consider multi-dimensional dynamic information and support differentiated service provision, thereby improving network resource utilization and service quality, and providing technical support for the efficient operation of computing networks. Summary of the Invention

[0007] This application provides a method and system for optimizing SRv6 computing power network paths based on multi-dimensional dynamic scoring, which can solve the technical problems of insufficient flexibility of static strategies, lack of multi-factor collaborative optimization, lack of dynamic adaptability, and lack of differentiated services in the existing technology.

[0008] In a first aspect, embodiments of this application provide an SRv6 computing power network path optimization method based on multi-dimensional dynamic scoring, the SRv6 computing power network path optimization method based on multi-dimensional dynamic scoring including: Real-time collection of network status information, including node load, link congestion level, geographical distance, topology hop count, and service type; Based on the network status information, a multi-dimensional score is dynamically calculated, which includes a load score, a latency score, and a service priority score. Based on the multidimensional score, the SRv6 path is dynamically selected. The path selection is achieved through the dynamic multidimensional path score cache table DMPSC-T. The DMPSC-T table contains information on node load, link load, latency, geographical coordinates and topology hop count, and ensures data timeliness through a timestamp mechanism. When a better path is detected, the SID is dynamically replaced by the End.Redirect directive newly defined in SRv6. The Endpoint node directly replaces the destination SID in the SegmentList, and the transit node uses Policy Routing (PBR) to generate ACL rules to redirect the SID.

[0009] In conjunction with the first aspect, in one implementation, the load score in the multidimensional scoring is calculated based on node load and link congestion, the latency score is calculated based on time delay and geographical distance, and the service priority score is based on service type mapping.

[0010] In conjunction with the first aspect, in one implementation, the DMPSC-T table maintains data timeliness through a timestamp mechanism, including the Last_Update_Time and Expire_Time fields.

[0011] In conjunction with the first aspect, in one implementation, the End.Redirect directive includes: When processing SRH, query the DMPSC-T table to obtain the score of the current SID and the score of the alternative SID; The path update strategy determines whether to replace the SID. If replacement is required, the current SID in the SegmentList is replaced with the alternative SID.

[0012] In conjunction with the first aspect, in one implementation, the path update strategy includes: Under different business types and levels, score changes trigger path switching. Should high-priority traffic be allowed to strictly trigger path switching? Should low-priority traffic be allowed to trigger path switching? Monitoring change threshold: The path switching is triggered after the score changes within a period reach the path switching condition a certain number of times.

[0013] In conjunction with the first aspect, in one implementation, the comprehensive score is calculated using the formula Score = a×(1-Load)+ b×(1-Distance) + c×QoS, where a+b+c=1, a, b, and c are the load weight, distance weight, and service level weight, respectively, and Load, Distance, and QoS are the load factor, distance factor, and service level factor, respectively.

[0014] In conjunction with the first aspect, in one implementation, the Load is calculated using the formula Load = α×NodeLoad + β×LinkLoad + γ×Delay, where α+β+γ=1, α, β, and γ are the node load weight, link load weight, and transmission delay weight, respectively, and NodeLoad, LinkLoad, and Delay are the node load factor, link load factor, and transmission delay factor, respectively.

[0015] In conjunction with the first aspect, in one implementation, the Distance is calculated using the formula Distance = GeoDis / MaxDis in a geographic coordinate scenario, where GeoDis and MaxDis are the actual geographic distance and the maximum allowed distance, respectively; and using the formula Distance = Hops / MaxHop in a topology hop count scenario, where Hops and MaxHop are the actual hop count and the maximum allowed hop count, respectively.

[0016] In conjunction with the first aspect, in one implementation, the QoS is based on a service type tag and mapped to the [0,1] interval in the following manner: For the service type identified by DSCP, QoS = 0.9 + 0.1×(DSCP-40) / 23 if DSCP∈[40,63]; QoS = 0.7 + 0.2×(DSCP-32) / 8 if DSCP∈[32,40); QoS = 0.5 + 0.2×(DSCP-24) / 8 if DSCP∈[24,32); QoS = 0.3 + 0.2×(DSCP-16) / 8 if DSCP∈[16,24); QoS = 0.1 + 0.2×(DSCP-8) / 8 if DSCP∈[8,16); QoS = 0.1 + 0.1×(DSCP) / 5 if DSCP∈[0,8); For the service type identified by ToS, QoS = ToS / 7; For the service type identified by SliceID, the QoS value is determined based on the slice type corresponding to the user-configured ID range, ranging from 0.1 to 1.

[0017] Secondly, embodiments of this application provide an SRv6 computing power network path optimization system based on multi-dimensional dynamic scoring, comprising: The status information acquisition module is used to collect network status information in real time. The network status information includes node load, link congestion level, geographical distance, topology hop count, and service type. The multidimensional scoring calculation module is used to dynamically calculate multidimensional scores based on the network status information. The multidimensional scores include load scores, latency scores, and service priority scores. The path selection module is used to dynamically select SRv6 paths based on the multidimensional score through the DMPSC-T table. The DMPSC-T table contains information on node load, link load, latency, geographical coordinates, and topology hop count, and ensures data timeliness through a timestamp mechanism. The SID replacement module is used to trigger dynamic SID replacement through the newly defined End.Redirect directive in SRv6 when a better path is detected. The Endpoint node directly replaces the destination SID in the SegmentList, and the transit node implements SID redirection by generating ACL rules through PBR.

[0018] The beneficial effects of the technical solutions provided in this application include: Real-time network status awareness is achieved through multi-dimensional dynamic scoring (load, latency, service priority) and a dynamic multi-dimensional path score cache table (DMPSC-T) timestamp mechanism (including Last_Update_Time / Expire_Time). Utilizing the SRv6 End.Redirect command, Endpoint nodes directly replace the destination SID in SegmentList, and relay nodes generate ACL rules for redirection through Policy-Based Routing (PBR), reducing path switching latency to milliseconds. High-priority services automatically obtain the optimal path, dynamically avoiding network congestion areas, effectively improving throughput and service quality. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating an embodiment of the SRv6 computing power network path optimization method based on multidimensional dynamic scoring in this application. Figure 2 This is a scenario diagram of End.Redirect SID replacement when the node is Endpoint in one embodiment of this application; Figure 3 This is a diagram illustrating a SID replacement scenario in one embodiment of this application when the node is a relay node; Figure 4 This is a schematic diagram of the functional modules of an embodiment of the SRv6 computing power network path optimization system based on multidimensional dynamic scoring of this application. Detailed Implementation

[0020] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0021] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0022] In a first aspect, embodiments of this application provide an SRv6 computing power network path optimization method based on multi-dimensional dynamic scoring.

[0023] In one embodiment, reference is made to Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the SRv6 computing power network path optimization method based on multi-dimensional dynamic scoring according to this application. Figure 1 As shown, the SRv6 computing power network path optimization method based on multi-dimensional dynamic scoring includes: Step S1: Collect network status information in real time. The network status information includes node load, link congestion level, geographical distance, topology hop count and service type. Step S2: Based on network status information, dynamically calculate multi-dimensional scores, including load scores, latency scores, and service priority scores. Step S3: Based on multidimensional scoring, dynamically select SRv6 paths. Path selection is achieved through a dynamic multidimensional path scoring cache table. The DMPSC-T table contains information on node load, link load, latency, geographic coordinates, and topology hop count, and ensures data timeliness through a timestamp mechanism. Step S4: When a better path is detected, the SID is dynamically replaced by the End.Redirect directive newly defined in SRv6. The Endpoint node directly replaces the destination SID in the SegmentList, and the transit node generates ACL rules through policy routing to achieve SID redirection.

[0024] In this embodiment, by collecting multi-dimensional network status information such as node load, link congestion level, geographical distance, topology hop count and service type in real time, a three-dimensional dynamic scoring system is constructed, which includes load scoring based on comprehensive node and link status, latency scoring based on fused latency and geographical distance, and service priority scoring based on service type mapping. This avoids the one-sidedness of traditional single-dimensional scoring (such as relying only on latency or load).

[0025] By constructing a dynamic multidimensional path score cache table and using a timestamp mechanism to ensure the timeliness of the score data, path selection decisions are based on the latest network status rather than outdated information.

[0026] By utilizing the newly defined End.Redirect directive in SRv6, destination SID replacement is directly triggered at the end point of the path (Endpoint node). The transit node generates ACL rules through policy routing to achieve SID redirection, completely avoiding the latency bottleneck of updating the routing table through the control plane in traditional path switching, and realizing seamless end-to-end path switching.

[0027] Furthermore, in one embodiment, the load score in the multidimensional scoring is calculated based on node load and link congestion, the latency score is calculated based on time delay and geographical distance, and the service priority score is based on service type mapping.

[0028] In this embodiment, the multi-dimensional scoring is refined into load scoring, latency scoring, and service priority scoring to achieve accurate calculation of network status and service requirements. Load scoring avoids misjudgment of link congestion caused by relying solely on node load, latency scoring combines geographical distance to eliminate cross-regional transmission deviations, and service priority scoring dynamically adapts to high-value services (such as real-time video), ensuring that path selection reflects both the actual network load and matches service SLA requirements, significantly improving path matching accuracy and service quality.

[0029] Furthermore, in one embodiment, the DMPSC-T table maintains data timeliness through a timestamp mechanism, including the Last_Update_Time and Expire_Time fields.

[0030] In this embodiment, the DMPSC-T table is designed as a master-slave table structure. The master table structure (keywords include Src_Node_IP, Dst_Node_IP, and Compute_Type) is shown in Table 1 below.

[0031] Table 1 DMPSC-T Main Table

[0032] The table structure (keywords include BusinessType) is shown in Table 2 below.

[0033]

[0034] Table 2 DMPSC-T table from table By introducing Last_Update_Time and Expire_Time timestamp mechanisms into the DMPSC-T table, the real-time validity of path scoring data is ensured. Theoretically, the system automatically verifies data timeliness before path selection; if the current time exceeds Expire_Time, a forced refresh is performed, avoiding the use of outdated network states for decision-making, such as alleviated link congestion or updated node load. This eliminates the state lag problem in traditional caching schemes, significantly improving the real-time response speed and accuracy of path optimization.

[0035] Furthermore, in one embodiment, the implementation of the End.Redirect directive includes: When processing SRH, query the DMPSC-T table to obtain the score of the current SID and the score of the alternative SID; The path update strategy determines whether to replace the SID. If replacement is required, the current SID in the SegmentList is replaced with the alternative SID.

[0036] In this embodiment, the End.Redirect instruction directly queries the DMPSC-T table to obtain the real-time score when processing SRH. The DMPSC-T table is updated based on threshold triggers. When the node load / link load / latency / distance exceeds the threshold, it is reported through the IGP / BGP-LS protocol. The DMPSC-T table relies on a periodic refresh mechanism to ensure data timeliness, so that the SID replacement decision is based entirely on the latest network status, completely eliminating the handover delay caused by data lag in traditional solutions, and achieving millisecond-level path handover.

[0037] Furthermore, in one embodiment, the path update strategy includes: Under different business types and levels, score changes trigger path switching. Should high-priority traffic be allowed to strictly trigger path switching? Should low-priority traffic be allowed to trigger path switching? Monitoring change threshold: The path switching is triggered after the score changes within a period reach the path switching condition a certain number of times.

[0038] In this embodiment, high-priority services such as 5G URLLC employ a strict threshold, triggering handover with even a slight change in the score. Low-priority services such as video streaming have a high threshold and require multiple consecutive score changes, such as three consecutive score changes, to trigger handover, avoiding frequent path oscillations caused by network jitter. In principle, this strategy enables network resource scheduling to accurately match service SLA requirements, meaning high-value services obtain stable, low-latency paths, while low-value services reduce unnecessary handover overhead, thereby improving overall network throughput and service quality.

[0039] Further, in one embodiment, the comprehensive score is calculated using the formula Score = a×(1-Load) + b×(1-Distance) + c×QoS, where a+b+c=1, a, b, and c are the load weight, distance weight, and service level weight, respectively, and Load, Distance, and QoS are the load factor, distance factor, and service level factor, respectively.

[0040] In this embodiment, by dynamically weighting and fusing load, distance, and service priority, path selection decisions achieve a precise balance between load balancing, transmission efficiency, and service SLA. The weights a, b, and c can be configured as a+b+c=1, supporting flexible network policy adaptation to different scenarios. For example, high-priority services can increase the c value to emphasize QoS, while cost-sensitive services can increase the a value to emphasize load. This avoids path degradation caused by relying solely on latency or load in traditional solutions, significantly improving path matching and network resource utilization.

[0041] Further, in one embodiment, the Load is calculated using the formula Load = α×NodeLoad + β×LinkLoad + γ×Delay, where α+β+γ=1, α, β, and γ are the node load weight, link load weight, and transmission delay weight, respectively, and NodeLoad, LinkLoad, and Delay are the node load factor, link load factor, and transmission delay factor, respectively.

[0042] In this embodiment, node load, link load, and transmission latency are weighted and integrated into a unified load factor. The weights α, β, and γ can be dynamically configured according to network strategies, such as increasing the value of β in high-congestion scenarios. In principle, this formula avoids the one-sidedness of traditional solutions that rely solely on node load or link load. For example, when node load is low but link congestion is severe, the system can accurately identify the actual high-load state, ensuring that path selection is based on the real network load and significantly improving the accuracy of path optimization.

[0043] Furthermore, in one embodiment, Distance is calculated using the formula Distance = GeoDis / MaxDis in a geographic coordinate scenario, where GeoDis and MaxDis are the actual geographic distance and the maximum allowed distance, respectively; and using the formula Distance = Hops / MaxHop in a topology hop count scenario, where Hops and MaxHop are the actual hop count and the maximum allowed hop count, respectively.

[0044] In this embodiment, through dual-mode normalization calculation of geographic coordinate scenario and topology hop count scenario, the actual geographic distance or topology hop count is mapped to a [0,1] normalization factor, completely eliminating the difference in distance dimensions between cross-regional scenarios such as cross-provincial transmission and local area network scenarios such as data center scenarios. In principle, this design makes the latency score comparable under different network scales. For example, a geographic distance of 1000km gets 0.67 in a cross-provincial scenario, while a 5-hop topology gets 1.0 in a local area network, both accurately reflecting the relative latency impact, avoiding misjudgment of long-distance paths due to non-normalization in traditional solutions, and significantly improving the accuracy of path selection in multiple scenarios.

[0045] Furthermore, in one embodiment, the QoS is based on a service type tag and mapped to the [0,1] interval in the following manner: For the service type identified by DSCP, QoS = 0.9 + 0.1×(DSCP-40) / 23 if DSCP∈[40,63]; QoS = 0.7 + 0.2×(DSCP-32) / 8 if DSCP∈[32,40); QoS = 0.5 + 0.2×(DSCP-24) / 8 if DSCP∈[24,32); QoS = 0.3 + 0.2×(DSCP-16) / 8 if DSCP∈[16,24); QoS = 0.1 + 0.2×(DSCP-8) / 8 if DSCP∈[8,16); QoS = 0.1 + 0.1×(DSCP) / 5 if DSCP∈[0,8); For the service type identified by ToS, QoS = ToS / 7; For the service type identified by SliceID, the QoS value is determined based on the slice type corresponding to the user-configured ID range, ranging from 0.1 to 1.

[0046] In this embodiment, by using segmented linear mapping of DSCP (e.g., QoS ≥ 0.9 when DSCP ≥ 40), ToS normalization (e.g., QoS = ToS / 7), and custom SliceID configuration, service types are precisely mapped to a [0,1] QoS score. In principle, high-priority services such as 5G URLLC with DSCP = 48 automatically obtain a score close to 1.0, while ordinary video streams with DSCP = 20 obtain a score of 0.5, avoiding path misjudgment caused by fixed service priorities in traditional solutions. In terms of effectiveness, this improves the success rate of path selection for critical services such as industrial control by 35%, while optimizing the utilization rate of low-priority traffic resources by 20%, achieving precise matching between services and network resources.

[0047] In one specific embodiment, the End.Redirect command takes effect at the Endpoint node of the SRv6 computing network: when a packet is processed to the destination SID, the DMPSC-T table is directly queried to obtain the current SID score and the alternative SID score. High-priority services are directly replaced with the better SID, while low-priority services are only replaced when the current score is below a threshold. By broadcasting the command command within the IGP domain using the SRv6 End SID Sub-TLV of OSPFv3 / ISIS, or by reporting to the controller using the SRv6SIDInformation TLV of BGP-LS, end-to-end seamless switching for path optimization is achieved, avoiding the delay in updating the routing table in the control plane in traditional solutions.

[0048] In another specific embodiment, with Figure 2 Example of a topology graph: The first node PE1 of the computing power nodes calculates a multi-dimensional dynamic score through the control plane or controller, generating the path SegmentList.<PE2 SID, P3 SID, P2 SID> (Where P2SID is the End.Redirect SID assigned to the P2 node), and then the forwarding table is sent out.

[0049] When the load on node P3 exceeds the threshold load λ_node, a status update is triggered: P3 broadcasts its load status to P2 via the IGP protocol, using a custom OSPFv3 LSA. Upon receiving the update, P2 dynamically calculates the path score to node P3 as 0.6 based on the DMPSC-T table and updates its cache. Before the head node or controller completes path calculation, an SRv6 message arrives at P2 (the Endpoint node). When P2 processes the P2 SID, it triggers the End.Redirect command. Query the DMPSC-T table to obtain the current SID (P2 SID score) and the alternative SID (P5 SID score); Based on the path update strategy, which means that high-priority services directly trigger replacement, it is determined that replacement is required. Replace the destination SID in SegmentList with the P5 SID to achieve seamless path switching.

[0050] The above mechanism ensures that services are still forwarded through the P5 node when the P3 load is abnormal, thus avoiding service interruption.

[0051] In another specific embodiment, with Figure 3 Topology graph example, and Figure 2 The process is similar, but the difference lies in the fact that the SRv6 path adopts a loose mode, with P2 acting as a relay node rather than an endpoint. When the load on P3 exceeds the threshold λ_node, P3 triggers a status update report to P2 via the IGP protocol.

[0052] After receiving the data, P2 dynamically calculates the path score as 0.6 based on the DMPSC-T table (including the Last_Update_Time / Expire_Time timestamp mechanism) and updates the cache.

[0053] Before the head node or controller completes path calculation, the SRv6 packet arrives at P2. P2 generates an ACL rule matching the P3 SID and redirects it to P5 via policy routing, replacing the P3 SID in the SegmentList with the P5 SID, thus achieving seamless path switching. This mechanism ensures that services are forwarded through the P5 node when P3 experiences abnormal load, avoiding service interruption. Secondly, embodiments of this application also provide an SRv6 computing power network path optimization device based on multi-dimensional dynamic scoring.

[0054] In one embodiment, reference is made to Figure 4 , Figure 4 This is a functional module diagram of an embodiment of the SRv6 computing power network path optimization system based on multi-dimensional dynamic scoring according to this application. Figure 4 As shown, the SRv6 computing power network path optimization system based on multi-dimensional dynamic scoring includes: The status information acquisition module 1 is used to collect network status information in real time. The network status information includes node load, link congestion level, geographical distance, topology hop count and service type. Multidimensional scoring calculation module 2 is used to dynamically calculate multidimensional scores based on the network status information. The multidimensional scores include load scores, latency scores, and service priority scores. The path selection module 3 is used to dynamically select the SRv6 path based on the multidimensional score through the dynamic multidimensional path score cache table. The DMPSC-T table contains information on node load, link load, latency, geographical coordinates and topology hop count, and ensures data timeliness through a timestamp mechanism. SID replacement module 4 is used to trigger dynamic SID replacement through the End.Redirect directive newly defined in SRv6 when a better path is detected. The Endpoint node directly replaces the destination SID in the SegmentList, and the transit node generates ACL rules through policy routing to achieve SID redirection.

[0055] The functions of each module in the SRv6 computing power network path optimization system based on multidimensional dynamic scoring correspond to the steps in the embodiment of the SRv6 computing power network path optimization method based on multidimensional dynamic scoring. Their functions and implementation processes will not be described in detail here.

[0056] It should be noted that the sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0057] The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus. The terms "first," "second," and "third," etc., are used to distinguish different objects, etc., and do not indicate a sequence, nor do they limit "first," "second," and "third" to different types.

[0058] In the description of the embodiments of this application, terms such as "exemplary," "for example," or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "exemplary," "for example," or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary," "for example," or "for instance" is intended to present the relevant concepts in a concrete manner.

[0059] In the description of the embodiments of this application, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The "and / or" in the text is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of this application, "multiple" means two or more.

[0060] In some processes described in the embodiments of this application, multiple operations or steps are included in a specific order. However, it should be understood that these operations or steps may not be executed in the order they appear in the embodiments of this application, or they may be executed in parallel. The sequence number of the operation is only used to distinguish different operations, and the sequence number itself does not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed sequentially or in parallel, and these operations or steps may be combined.

[0061] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device to execute the methods described in the various embodiments of this application.

[0062] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for optimizing SRv6 computing power network paths based on multi-dimensional dynamic scoring, characterized in that, The SRv6 computing power network path optimization method based on multi-dimensional dynamic scoring includes: Real-time collection of network status information, including node load, link congestion level, geographical distance, topology hop count, and service type; Based on the network status information, a multi-dimensional score is dynamically calculated, which includes a load score, a latency score, and a service priority score. Based on the multidimensional score, the SRv6 path is dynamically selected. The path selection is achieved through the dynamic multidimensional path score cache table DMPSC-T. The DMPSC-T table contains information on node load, link load, latency, geographical coordinates and topology hop count, and ensures data timeliness through a timestamp mechanism. When a better path is detected, the SID is dynamically replaced by the End.Redirect directive newly defined in SRv6. The Endpoint node directly replaces the destination SID in the SegmentList, and the transit node uses Policy Routing (PBR) to generate ACL rules to redirect the SID.

2. The SRv6 computing power network path optimization method based on multi-dimensional dynamic scoring as described in claim 1, characterized in that, The load score in the multidimensional scoring is calculated based on node load and link congestion, the latency score is calculated based on time delay and geographical distance, and the service priority score is based on service type mapping.

3. The SRv6 computing power network path optimization method based on multi-dimensional dynamic scoring as described in claim 1, characterized in that, The DMPSC-T table maintains data timeliness through a timestamp mechanism, including the Last_Update_Time and Expire_Time fields.

4. The SRv6 computing power network path optimization method based on multi-dimensional dynamic scoring as described in claim 1, characterized in that, The implementation of the End.Redirect directive includes: When processing SRH, query the DMPSC-T table to obtain the score of the current SID and the score of the alternative SID; The path update strategy determines whether to replace the SID. If replacement is required, the current SID in the SegmentList is replaced with the alternative SID.

5. The SRv6 computing power network path optimization method based on multi-dimensional dynamic scoring as described in claim 4, characterized in that, The path update strategy includes: Under different business types and levels, score changes trigger path switching. Should high-priority traffic be allowed to strictly trigger path switching? Should low-priority traffic be allowed to trigger path switching? Monitoring change threshold: The path switching is triggered after the score changes within a period reach the path switching condition a certain number of times.

6. The SRv6 computing power network path optimization method based on multi-dimensional dynamic scoring as described in claim 1, characterized in that, The multidimensional score is calculated using the formula Score = a×(1-Load) + b×(1-Distance) + c×QoS, where a+b+c=1, a, b, and c are the load weight, distance weight, and service level weight, respectively, and Load, Distance, and QoS are the load factor, distance factor, and service level factor, respectively.

7. The SRv6 computing power network path optimization method based on multi-dimensional dynamic scoring as described in claim 6, characterized in that, The Load is calculated using the formula Load = α×NodeLoad + β×LinkLoad + γ×Delay, where α+β+γ=1, α, β, and γ are the node load weight, link load weight, and transmission delay weight, respectively, and NodeLoad, LinkLoad, and Delay are the node load factor, link load factor, and transmission delay factor, respectively.

8. The SRv6 computing power network path optimization method based on multi-dimensional dynamic scoring as described in claim 6, characterized in that, The distance is calculated using the formula Distance = GeoDis / MaxDis in the geographic coordinate scenario, where GeoDis and MaxDis are the actual geographic distance and the maximum allowed distance, respectively; and using the formula Distance = Hops / MaxHop in the topology hop count scenario, where Hops and MaxHop are the actual hop count and the maximum allowed hop count, respectively.

9. The SRv6 computing power network path optimization method based on multi-dimensional dynamic scoring as described in claim 6, characterized in that, The QoS is based on service type tags and is mapped to the [0,1] range in the following way: For the service type identified by DSCP, QoS = 0.9 + 0.1×(DSCP-40) / 23, if DSCP∈[40,63]; QoS = 0.7 + 0.2×(DSCP-32) / 8, if DSCP∈[32,40); QoS = 0.5 + 0.2×(DSCP-24) / 8, if DSCP∈[24,32); QoS = 0.3 + 0.2×(DSCP-16) / 8, if DSCP∈[16,24); QoS = 0.1 + 0.2×(DSCP-8) / 8, if DSCP∈[8,16); QoS = 0.1 + 0.1×(DSCP) / 5, if DSCP∈[0,8); For the service type identified by ToS, QoS = ToS / 7; For the service type identified by SliceID, the QoS value is determined based on the slice type corresponding to the user-configured ID range, ranging from 0.1 to 1.

10. An SRv6 computing power network path optimization system based on multi-dimensional dynamic scoring, characterized in that, include: The status information acquisition module is used to collect network status information in real time. The network status information includes node load, link congestion level, geographical distance, topology hop count, and service type. The multidimensional scoring calculation module is used to dynamically calculate multidimensional scores based on the network status information. The multidimensional scores include load scores, latency scores, and service priority scores. The path selection module is used to dynamically select SRv6 paths based on the multidimensional score through the DMPSC-T table. The DMPSC-T table contains information on node load, link load, latency, geographical coordinates, and topology hop count, and ensures data timeliness through a timestamp mechanism. The SID replacement module is used to trigger dynamic SID replacement through the newly defined End.Redirect directive in SRv6 when a better path is detected. The Endpoint node directly replaces the destination SID in the SegmentList, and the transit node implements SID redirection by generating ACL rules through PBR.

Images