A system and method for dynamic link quality perception optimization in a mesh topology

Abstract

This invention relates to the field of communication network technology, specifically disclosing a dynamic link quality sensing and optimization system and method in a mesh topology. The optimization system includes: a round-trip time (RTT) sampling module for acquiring the RTT values ​​of each communication route link between the originating end and the target receiving end in the mesh topology; and a routing selection module for calculating the cost of the corresponding communication route links based on each set of RTT values, dynamically selecting communication route links with low-latency links as the preferred selection criterion. This invention measures the RTT of different routing paths between the originating end and the target receiving end in a mesh topology and uses the Babel routing protocol to select low-latency links as the preferred selection criterion, thereby dynamically optimizing the quality of each communication route link in a mesh topology communication network based on the Babel routing protocol. This effectively improves the accuracy of routing path optimization and the quality of communication route links in the communication network.

Description

Technical Field

[0001] This invention relates to the field of communication network technology, specifically to a dynamic sensing and optimization system and method for link quality in a mesh topology. Background Technology

[0002] In the mesh topology of a communication network, different arrangements and combinations of communication nodes form the communication routing links from the originating end to the target receiving end. The Babel routing protocol, as a distance-vector routing protocol that avoids loops, has the ability to converge quickly and update routes efficiently. It plays a crucial role in the efficiency and stability of communication routing links in mesh topologies and is an important technical means for constructing communication routing links.

[0003] However, while the Babel routing protocol provides a flexible routing mechanism, in practical deployments, it uses packet loss rate to evaluate the quality of wireless communication routing links, while for other types of communication routing links, it uses a simple hop count as the metric. This hybrid metric mechanism performs well in conventional communication network environments, but in special mesh topologies based on tunneling services—such as Virtual Private Networks (VPNs)—its routing path selection capability has significant limitations, leading to suboptimal routing decisions. Specific examples include... Figure 1 As shown: exist Figure 1 In the communication mesh topology shown, there are three routers A, B and C located in the same country (region) and router D located in another country (region). They are connected in a diamond mesh topology based on tunneling service technology. Assuming router A is the originating end and router C is the destination receiving end, the local routing path through router B is obviously the optimal communication route, which usually provides better service quality and reduces costs, while the remote routing path through router D is not. However, the existing Babel routing protocol treats the two routing paths between the originating point A and the destination receiver C as having the same metric. This results in a near 50% probability that traffic will be routed to a remote router D, and then from the remote router D to the local destination receiver C, leading to suboptimal routing decisions.

[0004] Therefore, in mesh topologies based on tunneling services, such as VPN networks, the existing Babel routing protocol suffers from suboptimal routing decisions due to the use of packet loss rate or hop count as metrics. This results in significant limitations in routing capabilities, with low accuracy and poor performance, which need to be improved and addressed. Summary of the Invention

[0005] The technical objective of this invention is to address the specific characteristics of the mesh topology of the aforementioned communication network and the Babel routing protocol, as well as the technical deficiencies of the Babel routing protocol in metric decision-making among different routing paths in the mesh topology. The invention provides a dynamic link quality awareness optimization system for mesh topologies that dynamically selects communication routing links based on low-latency links, and an optimization method based on this system.

[0006] The technical objective of this invention is achieved through the following technical solution: a dynamic sensing and optimization system for link quality in a mesh topology, the optimization system comprising: The round-trip time sampling module is used to obtain the RTT value of each communication route link between the originating end and the target receiving end in the mesh topology; The routing module calculates the cost of the corresponding communication route link based on each group of RTT values ​​and dynamically selects the communication route link with low-latency links as the preferred criterion.

[0007] Furthermore, the round-trip time sampling module adds an original timestamp and a received timestamp to the node neighbor table data structure of the Babel routing protocol; The original timestamp is a 32-bit unsigned integer that records the local clock value of the first node of the neighboring node in the corresponding communication routing link when the current message is sent. The received timestamp is a 32-bit unsigned integer that records the local clock value of the second node of the neighboring node in the corresponding communication routing link when the message is received.

[0008] Furthermore, the timestamp is the corresponding timestamp in the Hello TLV data packet and the IHUTLV data packet.

[0009] Furthermore, the mesh topology is a mesh topology based on tunnel service technology.

[0010] A dynamic link quality sensing and optimization method for a mesh topology, the optimization method being based on the aforementioned dynamic link quality sensing and optimization system for a mesh topology, and executing the following sequential process: Based on the timestamps in the Babel routing protocol, obtain the RTT values ​​of each communication routing link between the originating end and the target receiving end; The cost of the corresponding communication routing link is calculated based on the RTT value of each group, and the selection of the communication routing link is dynamically realized by following the selection rule of low latency link preference.

[0011] Furthermore, the process of obtaining the RTT value of each communication route link between the originating end and the target receiving end specifically includes the following sequential steps: S1. In the sending direction, according to the arrangement of neighboring nodes in the communication route link between the originating end and the target receiving end, the first node of each neighboring node in the Babel routing protocol periodically sends a Hello TLV data packet to the second node. The Hello TLV data packet sent contains a timestamp t1 based on the local clock of the first node; S2. When the second neighboring node receives the Hello TLV data packet sent by the first node, it calculates the reception time t1' using its local clock and performs the following operations: - Record t1 into the original timestamp field of the first node in the neighbor table data structure; - Record t1´ in the receiving timestamp field of the corresponding first node in the neighbor table data structure; If there are multiple sets of neighboring nodes in the communication route link between the originating end and the target receiving end, then the same applies. S3. In the transmission return direction, according to the arrangement order of neighboring nodes in the communication route link between the originating end and the target receiving end, the second node of each neighboring node in the Babel routing protocol sends an IHU TLV data packet to the first node; When sending IHU TLV data packets, the second node needs to check whether the original timestamp and the received timestamp in the neighbor table data structure have been defined. If they have been defined, the following requirements must be met: The -IHU TLV data packet is sent in the same group of data packets as the Hello TLV data packet; - The IHU TLV data packet contains the original timestamp and the received timestamp stored in the neighbor table data structure; S4. When the first node among the neighboring nodes receives a data packet from the second node, it performs the following operations: - Calculate the receiving time t2 based on the local clock; - Verify that the data packet contains both the Hello TLV data packet with timestamp t2´ and the IHU TLV data packet with timestamps t1 and t1´. - If the verification passes, the RTT value of the neighboring nodes is calculated using the following formula: RTT=(t2-t1)-(t2´-t1´); S5. The first node in the neighboring nodes also performs the following operations: - Record t2´ into the original timestamp field of the corresponding second node in the neighbor table data structure; - Record t2 in the receiving timestamp field of the corresponding second node in the neighbor table data structure; If there are multiple groups of neighboring nodes in the communication route link between the originating end and the target receiving end, the same principle applies, and the RTT values ​​of each group of neighboring nodes are added together to form the total RTT value.

[0012] Furthermore, the process of selecting communication routing links based on RTT values ​​specifically includes the following sequential steps: S1. Smooth the obtained RTT values; S2. Convert the smoothed RTT value into a cost value using a bounded nonlinear mapping; S3. Use a hysteresis filter to limit the oscillation amplitude in the middle of the RTT range, and dynamically optimize the communication route link between the originating end and the target receiving end.

[0013] Furthermore, the smoothing of RTT values ​​satisfies the following relationship: RTT = βRTT0 + (1-β)RTT n ; In the formula, RTT is the RTT value of the corresponding communication routing link; It is a smoothing constant; RTT0 is the RTT value that was not initially defined in the corresponding communication routing link; RTT n This corresponds to the RTT value of subsequent new samples in the communication routing link.

[0014] Furthermore, the smoothed RTT value is converted into a cost value, specifically including the mapping process from the smoothed RTT value to the link cost, the link cost mapping method process, and the link cost calculation process; Among them, the process of smoothing the mapping from RTT value to link cost satisfies monotonicity, upper limit saturation and lower limit stability; The monotonicity refers to the fact that the mapping must maintain a monotonically increasing characteristic. The larger the RTT value, the higher the cost of the corresponding communication routing link. The upper limit saturation means that when the RTT value exceeds the set threshold, the mapping result should remain constant; The lower limit stability means that the mapping result should remain constant when the RTT value is close to 0. The link cost mapping method uses a segmented parameterized mapping approach, which has three parameters: the minimum segment delay threshold rtt-min, the maximum segment delay threshold rtt-max, and the link cost penalty constant max-rtt-penalty. The segment cost calculation rules are as follows: - When RTT < rtt-min, it is classified as a low-latency link, and the communication routing link cost under this state remains the single-hop nominal cost; - When rtt-min≤RTT≤rtt-max is classified as a medium-latency link, the communication routing link cost under this state increases linearly with RTT; - When RTT > rtt-max, it is classified as a high-latency link. In this state, the communication routing link is subject to a fixed penalty value max-rtt-penalty on top of the nominal cost. The link cost calculation process uses an additive approach to calculate the link cost, satisfying the following relationship: M(c,m) = c + m; In the formula, M(c,m) is a function of the final metric value; C represents the locally calculated link cost; m is the metric for neighbor notifications.

[0015] Furthermore, the dynamic optimization selection of communication routing links satisfies the following rules: - Under instantaneous conditions, m(R´)<m(R); - Under long-term conditions, ms(R´)<ms(R); In the formula, m(R´) is the metric value of the backup communication route link under instantaneous conditions; m(R) is the metric of the current communication routing link in the instantaneous condition; ms(R´) is the metric for the backup communication route link in the long-term condition; ms(R) is the metric for the current communication route link in the long-term condition.

[0016] The beneficial technical effects of the present invention are as follows: The above-mentioned technical measures are designed to address the specific characteristics of the mesh topology of the communication network and the Babel routing protocol. By measuring the round-trip time (RTT) of different routing paths between the originating end and the target receiving end in the mesh topology, and by using the Babel routing protocol with low-latency links as the preferred selection criterion, the quality of each communication routing link in the mesh topology communication network based on the Babel routing protocol is dynamically optimized, thereby effectively improving the optimization accuracy of the communication network routing path and the quality of the communication network routing links. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of a network topology for a wireless communication network.

[0018] Figure 2 This is an architecture diagram of the dynamic perception optimization system of the present invention.

[0019] Figure 3 This is an architecture diagram of the sub-TLV format in the Babel protocol.

[0020] Figure 4 This is an architecture diagram of the timestamp sub-TLV format in the Hello TLV data packet structure.

[0021] Figure 5 This is an architecture diagram of the timestamp sub-TLV format in the IHU TLV data packet structure.

[0022] Figure 6 This is a flowchart illustrating the Mills algorithm. Detailed Implementation

[0023] This invention relates to the field of communication network technology, specifically to a dynamic sensing and optimization system for link quality in a mesh communication topology, and an optimization method based on this system. The following description is in conjunction with the accompanying drawings. Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 and Figure 6 The technical solution of this invention will be clearly and thoroughly explained.

[0024] It should be noted that the accompanying drawings of this invention are schematic, and unnecessary details have been simplified to clarify the technical objectives of this invention, so as to avoid obscuring the technical solutions contributed by this invention to the prior art. Furthermore, the terms "approximately" or "basically" used below to refer to quantities or fit relationships mean that reasonable assembly and processing errors are allowed in the industry, and do not literally describe absolute quantities or fit relationships.

[0025] The mesh topology of the communication network targeted in this invention is a mesh topology based on tunneling service technology, such as a VPN network.

[0026] The mesh topology of the present invention is as follows Figure 1 As shown, between the originating router A and the target receiving router C, there are multiple communication routing links (if there were only one communication routing link, there would be no room for dynamic sensing and optimization). One communication routing link is formed through router B (i.e., composed of neighboring nodes AB and BC), and another communication routing link is formed through router D (i.e., composed of neighboring nodes AD and DC). Therefore, how to dynamically sense and determine the optimal communication routing link between the originating end A and the target receiving end C is the focus of this invention. Of course, besides… Figure 1 Besides the mesh topology shown, other multi-path communication routing link structures from the originating end to the target receiving end can also be used, including communication routing links with only a single set of neighboring nodes between the originating end and the target receiving end. Figure 1For example, in the mesh topology shown, assume that there is a communication routing link between the originating router A and the target receiving router B, with neighboring nodes AB forming one link, and another communication routing link formed by the combination of neighboring nodes AD, DC, and CB.

[0027] This invention relates to a dynamic sensing and optimization system for link quality in mesh topology communication networks, comprising a round-trip time sampling module and a routing selection module. (See also...) Figure 2 As shown.

[0028] The round-trip time sampling module, based on the network data traffic between the originating end and the target receiving end in the mesh topology, periodically samples the round-trip time data of each communication routing link using the timestamp in the Babel routing protocol to obtain the RTT value of each communication routing link between the originating end and the target receiving end. In other words, the round-trip time sampling module is used to obtain the RTT value of each communication routing link between the originating end and the target receiving end in the mesh topology.

[0029] The routing module uses the input RTT values ​​to map the links to costs according to preset rules, and calculates the corresponding link costs based on the mapping results, dynamically implementing route selection with low-latency links as the benchmark.

[0030] Specifically, in the Babel routing protocol, each node needs to maintain a neighbor table. This data structure uses (interface, address) tuples as index keys to store information on all recently active neighboring interfaces. Neighbor table entries contain multi-dimensional network state data, including: interface reachability information, Hello message reception status records, link quality metrics, sequence number management mechanisms, and various timers and other key parameters. The round-trip time sampling module of this invention adds a raw timestamp and a received timestamp to the neighbor table data structure of each node in the Babel routing protocol to achieve accurate measurement of the round-trip time (RTT) between nodes. The raw timestamp is a 32-bit unsigned integer (modulo...). The operation records the local clock value of the first node (i.e., node A in neighbor node AB) of the corresponding neighbor node in the communication routing link when the current message is sent; the receive timestamp is a 32-bit unsigned integer (modulo 1). (Calculation), record the local clock value of the second node (i.e. node B in neighbor node AB) of the corresponding communication routing link when receiving the current message.

[0031] The timestamps mentioned above refer to the corresponding timestamps in the Hello TLV data packet and the IHU (i.e., "I Heard You") TLV data packet.

[0032] See Figure 3 As shown, the packet format in the Babel routing protocol is sub-TLV format. In the sub-TLV format packet architecture, Packet is the packet format of the Babel routing protocol. The packet body consists of a set of TLV (Type-Length-Value) structures arranged in sequence, and the packet tail is another set of TLV structures arranged in sequence. The payload part of the TLV structure contains a body and a list of sub-TLVs that can be selected for a specific TLV type. The specific parsing rules of the sub-TLV body depend on both the type of the sub-TLV itself and the type of the parent TLV it is embedded in.

[0033] Because the payload portion of a TLV structure can contain a list of sub-TLVs optional for a specific TLV type, users can extend it according to their own business needs. Therefore, this invention, based on the flexible and scalable functionality of TLV and its sub-TLV structures, designs dedicated fields in Hello TLV and IHU TLV packets to achieve inter-node round-trip time (RTT) measurement and optimal route selection. Specifically: The Hello TLV data packet, while fulfilling the basic function of neighbor discovery, supports reception cost calculation through extended fields; The IHU TLV data packet adds transmission cost measurement capabilities while maintaining the core function of bidirectional reachability verification.

[0034] The timestamp sub-TLV design in the Hello TLV packet is such that, when included in a Hello TLV packet, it needs to carry a timestamp, and its timestamp sub-TLV format is as follows: Figure 4 As shown: "Type" indicates that it is set to a specific number to indicate the timestamp sub-TLV; "Length" refers to the number of bytes in octet, excluding the main bytes of the type and length fields; "Send timestamp" indicates the time when the data packet containing this sub-TLV was sent, according to the sender's clock.

[0035] The timestamp sub-TLV design in IHU TLV packets is such that, when included in an IHU TLV packet, it carries two timestamps, with the timestamp sub-TLV format as follows: Figure 5 As shown: "Type" indicates that it is set to a specific number to indicate the timestamp sub-TLV; "Length" indicates the length of the body, in eight bytes, excluding the type and length fields; "Original timestamp" refers to a copy of the transmission timestamp of the last timestamp sub-TLV in the Hello TLV received from a node that has applied the IHU TLV; "Received timestamp" indicates the time when the last timestamped Hello TLV was received from the node that applied the included IHU TLV, according to the sender's clock.

[0036] Based on the above-mentioned dynamic sensing and optimization system for link quality in a mesh topology, the dynamic sensing and optimization method for link quality in a mesh topology of the present invention executes the following sequential process: Based on the timestamps in the Babel routing protocol, obtain the RTT values ​​of each communication routing link between the originating end and the target receiving end; The cost of the corresponding communication routing link is calculated based on the RTT value of each group, and the selection of the communication routing link is dynamically realized by following the selection rule of low latency link preference.

[0037] Specifically, the process of obtaining the RTT value of each communication route link between the originating end and the target receiving end includes the following sequential steps (such as...). Figure 6 (as shown) S1. In the forward direction of transmission, follow the order of neighboring nodes in the communication route link between the originating end and the target receiving end (e.g., ...). Figure 1 As shown, between the originating end A and the target receiving end C, the neighbor nodes of one of the communication routing links are arranged as AB and BC (taking AB as an example below). In the Babel routing protocol, the first node (i.e., node A) of each neighbor node (i.e., neighbor node AB in the example) periodically sends a Hello TLV data packet (i.e., Hello message) to the second node (i.e., node B). The Hello TLV data packet sent contains a timestamp t1 based on the local clock of the first node (i.e., node A); S2. When the second node (node ​​B) in the neighboring nodes (i.e., neighboring nodes AB in the example) receives the Hello TLV data packet sent by the first node (i.e., node A), it calculates the reception time t1' using its local clock and performs the following operations: - Record t1 in the original timestamp field of the first node (i.e., node A) in the neighbor table data structure; - Record t1' in the receiving timestamp field of the first node (i.e., node A) in the neighbor table data structure; If there are multiple sets of neighboring nodes in the communication route link between the originating end and the target receiving end, the same applies (i.e., the same applies to BC neighboring nodes). S3. In the transmission return direction, according to the arrangement order of neighbor nodes in the communication route link between the originating end and the target receiving end, the second node (i.e., node B) of each neighbor node (i.e., neighbor node AB in the example) in the Babel routing protocol sends an IHU TLV data packet (i.e., IHU message) to the first node (i.e., node A). When sending IHU TLV data packets, the second node (i.e., node B) needs to check whether the original timestamp and the received timestamp in the neighbor table data structure have been defined. If they have been defined, the following requirements must be met: - The IHU TLV data packet and the Hello TLV data packet (which can be a regularly scheduled Hello message or an unscheduled Hello message) are sent together in the same set of data packets; - The IHU TLV data packet contains the original timestamp and the received timestamp stored in the neighbor table data structure; S4. When the first node (node ​​A) in the neighboring nodes (i.e., neighboring nodes AB in the example) receives a data packet from the second node (i.e., node B), it performs the following operations: - Calculate the receiving time t2 based on the local clock; - Verify that the data packet contains both the Hello TLV data packet with timestamp t2´ and the IHU TLV data packet with timestamps t1 and t1´. - If the verification passes, the RTT value of the neighboring nodes is calculated using the following formula: RTT=(t2-t1)-(t2´-t1´); In the formula, all calculations use modular arithmetic, and this value is the RTT value in the neighboring node; S5. The first node (node ​​A) in the neighboring nodes (i.e., neighboring nodes AB in the example) also performs the following operations: - Record t2´ in the original timestamp field of the corresponding second node (i.e., node B) in the neighbor table data structure; - Record t2 in the receiving timestamp field of the corresponding second node (i.e., node B) in the neighbor table data structure; If there are multiple groups of neighboring nodes in the communication route link between the originating end and the target receiving end, the same applies (i.e., the same applies to BC neighboring nodes), and the RTT values ​​of each group of neighboring nodes are added together to form the total RTT value.

[0038] Since sampling is periodic, this invention employs the following method to discard outdated samples. When the corresponding node receives a data packet containing Hello and IHU packets, it performs the following steps: - IHU packet timestamp verification: Compare the current local time t2 with the original timestamp contained in the IHU packet; if the origin timestamp is a future time, or the deviation from the current time exceeds the set threshold T (e.g., the set threshold T = 3 minutes), the timestamp will still be recorded in the neighbor table, but it is prohibited from being used for RTT calculation; Hello packet timestamp verification: Similarly, the node compares the Hello timestamp with the received timestamp recorded in the neighbor table; if the Hello timestamp is earlier than the recorded timestamp, or if the time interval between the Hello timestamp and the recorded timestamp exceeds a set threshold T, then these timestamps are also included in the RTT calculation.

[0039] In addition, to ensure consistency of RTT calculation results across different implementations, timestamps should be collected at the same level of the protocol stack: - The sending timestamp should be collected immediately before the data packet is delivered to the protocol stack (to avoid the impact of queuing delays); - The receive timestamp should be collected immediately after receiving the data packet from the protocol stack.

[0040] In actual operating conditions, the RRT value measured above is prone to noise and negative feedback loops, which can lead to frequent oscillations. Therefore, it is not suitable to use it directly as input for the Babel routing protocol's selection process. Therefore, selecting communication routing links based on RTT values ​​should also include the following sequential steps: S1. Smooth the obtained RTT values ​​using the following formula to eliminate instability caused by outliers: RTT = βRTT0 + (1-β)RTT n ; In the formula, RTT is the RTT value of the corresponding communication routing link; This is a smoothing constant used to control the intensity of the smoothing process. The smoothing constant is selected between 0.8 and 0.9, with a default value of 0.836. RTT0 is the RTT value that was not initially defined in the corresponding communication routing link; RTT n This corresponds to the RTT value of subsequent new samples in the corresponding communication routing link; S2. Using a bounded nonlinear mapping, the smoothed RTT value is converted into a cost value to avoid instability at the upper and lower boundaries of the RTT range; More specifically, the smoothed RTT value is converted into a cost value, which includes the mapping process from the smoothed RTT value to the link cost, the link cost mapping method process, and the link cost calculation process. Among them, the process of smoothing the mapping from RTT value to link cost satisfies monotonicity, upper limit saturation and lower limit stability; The monotonicity refers to the fact that the mapping must maintain a monotonically increasing characteristic. The larger the RTT value, the higher the cost of the corresponding communication routing link. The upper limit saturation means that when the RTT value exceeds the set threshold, the mapping result should remain constant (i.e., all inferior links are considered equally inferior) in order to avoid routing oscillations caused by congested links. The lower limit stability means that the mapping result should remain constant when the RTT value is close to 0, thereby eliminating the impact of small fluctuations in low RTT links on routing stability. The link cost mapping method uses a segmented parameterized mapping approach, which has three parameters: the minimum segment delay threshold rtt-min, the maximum segment delay threshold rtt-max, and the link cost penalty constant max-rtt-penalty. The segment cost calculation rules are as follows: - When RTT < rtt-min, it is classified as a low-latency link, and the communication routing link cost under this state remains the single-hop nominal cost; - When rtt-min≤RTT≤rtt-max is classified as a medium-latency link, the communication routing link cost under this state increases linearly with RTT; - When RTT > rtt-max, it is classified as a high-latency link. In this state, the communication routing link is subject to a fixed penalty value max-rtt-penalty on top of the nominal cost. The aforementioned RTT-min value should be slightly larger than the RTT of the local uncongested link; The above-mentioned RTT-max value setting should avoid using the link RTT threshold (which may be due to long distances or congestion). The max-rtt-penalty value mentioned above is used to control the penalty intensity for long-distance links; The link cost calculation process uses an additive approach to calculate the link cost, satisfying the following relationship: M(c,m) = c + m; In the formula, M(c,m) is a function of the final metric value; C represents the locally calculated link cost; m is the metric for neighbor notifications; S3. When the link RTT is in the RTT-min and RTT-max range, the cost value may still fluctuate. In order to solve this kind of fluctuation problem, the present invention adopts a robust hysteresis method for processing. By using a hysteresis filter to limit the oscillation amplitude in the middle of the RTT range, the communication route link between the originating end and the target receiving end is dynamically optimized and selected. More specifically, the dynamic optimization selection of communication routing links satisfies the following rules: - Under instantaneous conditions, m(R´)<m(R); - Under long-term conditions, ms(R´)<ms(R); In the formula, m(R´) is the metric value of the backup communication route link under instantaneous conditions; m(R) is the metric of the current communication routing link in the instantaneous condition; ms(R´) is the metric for the backup communication route link in the long-term condition; ms(R) is the metric for the current communication route link in the long-term condition.

[0041] The above specific technical solutions are only used to illustrate the present invention, and are not intended to limit it.

[0042] Although the present invention has been described in detail with reference to the specific technical solutions described above, those skilled in the art should understand that modifications can still be made to the specific technical solutions described above, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the present invention.

Claims

1. A dynamic link quality sensing and optimization system for a mesh topology, characterized in that, The optimization system includes: The round-trip time sampling module is used to obtain the RTT value of each communication route link between the originating end and the target receiving end in the mesh topology; The routing module calculates the cost of the corresponding communication route link based on each group of RTT values ​​and dynamically selects the communication route link with low-latency links as the preferred criterion.

2. The link quality dynamic sensing and optimization system in the mesh topology according to claim 1, characterized in that, The round-trip time sampling module adds an original timestamp and a received timestamp to the node neighbor table data structure of the Babel routing protocol. The original timestamp is a 32-bit unsigned integer that records the local clock value of the first node of the neighboring node in the corresponding communication routing link when the current message is sent. The received timestamp is a 32-bit unsigned integer that records the local clock value of the second node of the neighboring node in the corresponding communication routing link when the message is received.

3. The link quality dynamic sensing and optimization system in the mesh topology according to claim 2, characterized in that, The timestamps are the corresponding timestamps in the Hello TLV data packets and IHUTLV data packets.

4. The link quality dynamic sensing and optimization system in the mesh topology according to claim 1, characterized in that, The mesh topology is a mesh topology based on tunnel service technology.

5. A method for dynamic quality sensing and optimization of links in a mesh topology, characterized in that, The optimization method, based on the link quality dynamic awareness and optimization system in the mesh topology according to any one of claims 1 to 4, executes the following sequential process: Based on the timestamps in the Babel routing protocol, obtain the RTT values ​​of each communication routing link between the originating end and the target receiving end; The cost of the corresponding communication routing link is calculated based on the RTT value of each group, and the selection of the communication routing link is dynamically realized by following the selection rule of low latency link preference.

6. The method for dynamic sensing and optimization of link quality in a mesh topology according to claim 5, characterized in that, The process of obtaining the RTT value of each communication route link between the originating end and the target receiving end includes the following sequential steps: S1. In the sending direction, according to the arrangement of neighboring nodes in the communication route link between the originating end and the target receiving end, the first node of each neighboring node in the Babel routing protocol periodically sends a Hello TLV data packet to the second node. The Hello TLV data packet sent contains a timestamp t1 based on the local clock of the first node; S2. When the second neighboring node receives the Hello TLV data packet sent by the first node, it calculates the reception time t1' using its local clock and performs the following operations: - Record t1 into the original timestamp field of the first node in the neighbor table data structure; - Record t1´ in the receiving timestamp field of the corresponding first node in the neighbor table data structure; If there are multiple sets of neighboring nodes in the communication route link between the originating end and the target receiving end, then the same applies. S3. In the transmission return direction, according to the arrangement order of neighboring nodes in the communication route link between the originating end and the target receiving end, the second node of each neighboring node in the Babel routing protocol sends an IHU TLV data packet to the first node; When sending IHU TLV data packets, the second node needs to check whether the original timestamp and the received timestamp in the neighbor table data structure have been defined. If they have been defined, the following requirements must be met: The -IHU TLV data packet is sent in the same group of data packets as the Hello TLV data packet; - The IHU TLV data packet contains the original timestamp and the received timestamp stored in the neighbor table data structure; S4. When the first node among the neighboring nodes receives a data packet from the second node, it performs the following operations: - Calculate the receiving time t2 based on the local clock; - Verify that the data packet contains both the Hello TLV data packet with timestamp t2´ and the IHU TLV data packet with timestamps t1 and t1´. - If the verification passes, the RTT value of the neighboring nodes is calculated using the following formula: RTT=(t2-t1)-(t2´-t1´); S5. The first node in the neighboring nodes also performs the following operations: - Record t2´ into the original timestamp field of the corresponding second node in the neighbor table data structure; - Record t2 in the receiving timestamp field of the corresponding second node in the neighbor table data structure; If there are multiple groups of neighboring nodes in the communication route link between the originating end and the target receiving end, the same principle applies, and the RTT values ​​of each group of neighboring nodes are added together to form the total RTT value.

7. The method for dynamic sensing and optimization of link quality in a mesh topology according to claim 5, characterized in that, The process of selecting communication routing links based on RTT values ​​includes the following sequential steps: S1. Smooth the obtained RTT values; S2. Convert the smoothed RTT value into a cost value using a bounded nonlinear mapping; S3. Use a hysteresis filter to limit the oscillation amplitude in the middle of the RTT range, and dynamically optimize the communication route link between the originating end and the target receiving end.

8. The method for dynamic sensing and optimization of link quality in a mesh topology according to claim 7, characterized in that, The smoothing of RTT values ​​satisfies the following relationship: RTT=βRTT0+(1-β)RTT n ; In the formula, RTT is the RTT value of the corresponding communication routing link; It is a smoothing constant; RTT0 is the RTT value that was not initially defined in the corresponding communication routing link; RTT n This corresponds to the RTT value of subsequent new samples in the communication routing link.

9. The method for dynamic sensing and optimization of link quality in a mesh topology according to claim 7, characterized in that, The smoothed RTT value is converted into a cost value, which specifically includes the mapping process from the smoothed RTT value to the link cost, the link cost mapping method process, and the link cost calculation process; Among them, the process of smoothing the mapping from RTT value to link cost satisfies monotonicity, upper limit saturation and lower limit stability; The monotonicity refers to the fact that the mapping must maintain a monotonically increasing characteristic. The larger the RTT value, the higher the cost of the corresponding communication routing link. The upper limit saturation means that when the RTT value exceeds the set threshold, the mapping result should remain constant; The lower limit stability means that the mapping result should remain constant when the RTT value is close to 0. The link cost mapping method uses a segmented parameterized mapping approach, which has three parameters: the minimum segment delay threshold rtt-min, the maximum segment delay threshold rtt-max, and the link cost penalty constant max-rtt-penalty. The segment cost calculation rules are as follows: - When RTT < rtt-min, it is classified as a low-latency link, and the communication routing link cost under this state remains the single-hop nominal cost; - When rtt-min≤RTT≤rtt-max is classified as a medium-latency link, the communication routing link cost under this state increases linearly with RTT; - When RTT > rtt-max, it is classified as a high-latency link. In this state, the communication routing link is subject to a fixed penalty value max-rtt-penalty on top of the nominal cost. The link cost calculation process uses an additive approach to calculate the link cost, satisfying the following relationship: M(c,m) = c + m; In the formula, M(c,m) is a function of the final metric value; C represents the locally calculated link cost; m is the metric for neighbor notifications.

10. The method for dynamic sensing and optimization of link quality in a mesh topology according to claim 7, characterized in that, The dynamic optimization selection of communication routing links satisfies the following rules: - Under instantaneous conditions, m(R´)<m(R); - Under long-term conditions, ms(R´)<ms(R); In the formula, m(R´) is the metric value of the backup communication route link under instantaneous conditions; m(R) is the metric of the current communication routing link in the instantaneous condition; ms(R´) is the metric for the backup communication route link in the long-term condition; ms(R) is the metric for the current communication route link in the long-term condition.

Images