A forwarding mechanism and transmission method of network data packets with concern for latency

Abstract

This invention discloses a forwarding mechanism and transmission method for network data packets with latency concerns, comprising a control plane and a data plane: In the control plane, a link-state database is generated using link latency as the link state, then a forwarding path database with latency concerns is generated from the link-state database, followed by a routing information table with latency concerns, and finally a forwarding information table with latency concerns. In the data plane, two regions are added to the network data packets to carry the latency target and latency error target of the data packet transmission, respectively. When a data packet passes through a network device, the target address of the data packet is searched and matched in the forwarding information table with latency concerns to obtain the corresponding forwarding information, and the data packet is sent according to the forwarding information. This invention can be implemented within network devices and deployed in large area networks covering hundreds or even thousands of kilometers, solving the latency guarantee problem within large areas.

Description

Technical Field

[0001] This invention belongs to the technical field of data communication networks, and relates to a forwarding mechanism and method for network data packets in routers and switches, specifically to a forwarding mechanism and transmission method for network data packets with latency concerns. Background Technology

[0002] In existing data communication networks, network packets are forwarded along the shortest path. If low-priority packets and important / critical packets are forwarded along the same shortest path, the low-priority packets may hinder the forwarding performance of the important / critical packets. Low-priority packets and important / critical packets typically have different latency requirements. Important / critical packets often have shorter latency requirements, needing to be delivered within a certain time, while low-priority packets can have longer delivery times; early or late arrival is not important. For example, a command to control a remote vehicle is a high-priority, critical packet that must be delivered on time; otherwise, the vehicle may be involved in an accident, resulting in loss of life and property. A packet querying weather temperature, on the other hand, is low-priority, and a slightly slower delivery time will not cause loss of life or property to the application and services.

[0003] To ensure the delivery of critical data packets, various technologies have been proposed in the data network communication field. These technologies revolve around Quality of Service (QoS). The earliest and most well-known technology is Integrated Service (IntServ), an end-to-end flow-based QoS technology. Before sending network data packets, network devices (routers and switches) need to request reserved network resources, including bandwidth and traffic, from other network devices via a network protocol called RSVP. After the network has reserved certain resources for the application, the network device sends the packet according to these resources. Integrated Service needs to reserve resources for a specific "flow" at every network device in the network's transmission path, resulting in significant network overhead. These reservations need to be refreshed periodically, adding further overhead to the network. During network data packet forwarding, received data packets are first classified into the corresponding flow and then shaped. Since various types of flows share a physical exit point, and their reserved bandwidth varies, they also need to be scheduled before being sent out to the corresponding physical exit point. There are usually many "flows" in the network, often reaching millions, tens of millions, or even hundreds of millions. Therefore, the scalability of this technology cannot meet the actual needs.

[0004] To reduce the network overhead of IntServ and address its scalability issues, the networking community proposed DiffServ (Differentiated Service). In DiffServ, the original ToS (Type of Service) field is used to carry the DiffServ service classification value (CodePoint). Since the original ToS field is only one byte (8 bits), the number of queues maintained and managed by the router for DiffServ service classification is smaller, and complex packet classification algorithms are not required. DiffServ significantly reduces the complexity of IntServ, has low overhead, and increases scalability, leading to its widespread deployment. However, it also has drawbacks: DiffServ is implemented on each router, and it can only guarantee differentiated forwarding performance on the local router, but it cannot guarantee the performance of the entire end-to-end forwarding path.

[0005] If the data communication network is an MPLS network, it also provides a technique called Traffic Engineering to implement QoS. This Traffic Engineering technique extends IntServ's RSVP to obtain a new signaling protocol, RSVP-TE. RSVP-TE establishes and reserves a forwarding path with pre-reserved bandwidth in the network. Unlike IntServ, this forwarding path does not correspond to a flow, but rather to a client or a class of applications. For example, when providing a Virtual Private Network (VPN), the traffic engineering tunnel reserved by RSVP-TE can be used to transmit data packets of the same client from one edge node to another between network edge nodes.

[0006] In recent years, the industry has also developed a routing technology called Segment Routing (SR). Segment Routing can guarantee that network traffic is forwarded along specific routes. These specific paths are not necessarily the shortest paths; they can be determined by the user through programming. For different purposes, the forwarding paths can be programmed to pass through certain specific nodes and links.

[0007] However, while all the technologies mentioned above can guarantee the forwarding path and bandwidth along that path, none can guarantee the specific latency of network data packet forwarding. Some emerging network applications require that the transmission latency of data packets within the network meet certain requirements. For example, using the network to issue control commands to remote vehicles: left turn, right turn, stop, accelerate, decelerate, etc.; in industrial control and automation, many machine tools and assembly lines can be operated and controlled remotely by workers via the network; and in hazardous or toxic environments, robots can be used to replace humans in labor and work, and these robots are controlled remotely. In the Industrial Internet and Industry 4.0, many application scenarios require data networks to provide end-to-end forwarding and transmission services with guaranteed latency.

[0008] To achieve end-to-end latency-guaranteed services, the industry has proposed "Time-Sensitive Networks" (TSNs). TSNs support Layer 2 switching and connectivity. The Internet Engineering Task Force (IETF) has also established the Deterministic Networking (DetNet) working group, aiming to provide TSN-like forwarding services in Layer 3 networks. Currently, most systems use Cyclic Queuing and Forwarding (CQF). CQF combines Per-Stream Filtering and Policing (PSFP) at the ingress end with Time-Aware Shaping (TAS) at the egress end, and utilizes time synchronization and Time-Division Multiplexing (TDM) to regulate, select, and schedule data traffic forwarding. The PSFP, a data flow-based filtering and control mechanism, and its associated technologies are complex and expensive to implement, and can only handle a limited amount of traffic. This means that they can only be used in local area networks (LANs) in smaller areas to date, and it cannot be demonstrated that they can be used in large area network deployments covering hundreds or even thousands of kilometers. Summary of the Invention

[0009] This invention addresses the limitations of existing technologies in ensuring latency over large areas by providing a latency-sensitive network packet forwarding mechanism and transmission method, comprising a control plane and a data plane. In the control plane, a link-state database is generated using link latency as the link state. Then, a latency-sensitive forwarding path database is generated from the link-state database, followed by a latency-sensitive routing information table, and finally, a latency-sensitive forwarding information table. In the data plane, two regions are added to the network packet to carry the packet's latency target and latency error target, respectively. Each time a packet passes through a network device, the destination address of the packet is searched and matched in the latency-sensitive forwarding information table to obtain the corresponding forwarding information, which is then used to send the packet. This transmission method can be implemented within network devices (routers and switches) without requiring circular queues, CQF (Continuous Quality Forwarding), or PSFP (Power Flow Based Filtering and Control) mechanisms. It can be deployed in large area networks covering hundreds or even thousands of kilometers, solving the latency guarantee problem over large areas.

[0010] To achieve the above objectives, the technical solution adopted by this invention is as follows: a network data packet forwarding mechanism concerned with latency, which includes a control plane and a data plane:

[0011] In the control plane, link state database is generated using link delay as the link state, then forwarding path database with relevant delay is generated from the link state database, then routing information table with relevant delay is generated, and finally forwarding information table with relevant delay is generated.

[0012] In the data plane, two regions are added to the network data packets to carry the data packet transmission delay target and delay error target, respectively. When a data packet passes through a network device, the target address of the data packet is searched and matched in the forwarding information table that is concerned with delay to obtain the corresponding forwarding information, and the data packet is sent according to the forwarding information.

[0013] As an improvement of the present invention, the latency of the link is treated as the state of the link, and then OSPF or IS-IS is used to generate a link state database LSDB, which contains latency information of all links.

[0014] Based on the link state database LSDB, a forwarding path database LC-PDB for which latency is a concern is constructed.

[0015] Based on the forwarding path database LC-PDB, a routing information table LC-RIB for latency is constructed and generated; in the routing information table LC-RIB for latency, there is a corresponding ordinary routing information table for different network latency parameters.

[0016] Based on the routing information table LC-RIB which is concerned with delay, a forwarding information table LC-FIB which is concerned with delay is constructed. The forwarding information table LC-FIB which is concerned with delay includes a delay target table and a normal forwarding information table. For a delay value in the delay target table, there is a corresponding normal forwarding information table.

[0017] As another improvement of the present invention, the link state database LSDB includes at least two types of information: (1) next hop and delay information; (2) DASN information of subnets directly connected to the router;

[0018] The next hop and latency information are formatted as follows: Router, Outgoing Interface, Next-Hop, and Latency. Here, Router is a network device, Next-Hop is the next network device directly connected to Router, Outgoing Interface is the connection from the exit point of Router to Next-Hop, and Latency is the latency value of the link from Router to Next-Hop.

[0019] The format of the DASN information of the subnet directly connected to the router is as follows: Router, Next-Hop, Subnet Address, Subnet Address Prefix Length, Latency, and Outgoing Interface. Next-Hop is set to "directed", Latency is set to 0, Subnet Address is the IP address of the subnet directly connected to the Router, Subnet Address Prefix Length is the effective prefix length of the Subnet Address, and Outgoing Interface is the exit point on the Router that is directly connected to the subnet represented by Subnet Address and Subnet Address Prefix Length.

[0020] As another improvement of the present invention, the latency-sensitive routing information table LC-RIB includes at least IP Address, IP Address Prefix Length, Outgoing Interface, Next-Hop, and Latency.

[0021] The general forwarding information table includes at least Address Prefix, Address Prefix Length, Outgoing Interface, and Next-Hop.

[0022] To achieve the above objectives, the present invention also adopts the following technical solution: a method for transmitting network data packets of concern to latency using the above-mentioned forwarding mechanism, comprising the following steps:

[0023] S1: Configure the maximum latency target (MLT) in the network;

[0024] S2: Configure the state parameters of each link in the network to the latency value of that link;

[0025] S3: Based on OSPF or IS-IS protocols, publish and exchange the status of each link, or generate a Link State Database (LSDB) through network management system or network controller configuration;

[0026] S4: Generate a routing information table (LC-RIB) for latency concerns from the Link State Database (LSDB). The routing table in the LC-RIB contains the latency for forwarding to the next hop. This includes the following steps:

[0027] S41: Generate a forwarding path database (LC-PDB) from the LSDB, focusing on latency; a forwarding path is defined as a sequence of network devices: <R1,R2,…,R n >, where R1 is the current router, R i+1 It is R i The next hop; a forwarding path of concern is defined as a tuple (p,t), where p is a forwarding path, t is the sum of the delays on each link in forwarding path p, and t<=MLT;

[0028] S42: Generate a routing information table LC-RIB (which controls latency) from the forwarding path database LC-PDB (which controls latency).

[0029] Steps S41 and S42 can be completed separately, or the two steps can be combined into one step.

[0030] S5: Generate a delay-sensitive forwarding information table (LC-FIB) from the delay-sensitive routing information table (LC-RIB); each delay-sensitive forwarding information table (LC-FIB) consists of multiple ordinary forwarding information tables (FIB). For each routing information (prefix, prefix-length, out-if, next-hop, latency) in the LC-RIB, add (prefix, prefix-length, out-if, next-hop) to the FIB. latency middle;

[0031] S6: Generate the link delay table NHLLT for directly connected links from the LSDB;

[0032] S7: Embed the latency requirement for the data packet to reach the target location in the network data packet, wherein the data packet contains transmission latency parameters: latency and jitter; or the data packet contains minimum latency parameters and maximum latency parameters;

[0033] S8: When a network device receives a data packet with embedded latency requirements, it forwards the data packet according to the LC-FIB (Local Latency Information Table) for which latency is a concern.

[0034] Compared with existing technologies, the technical advantages and effects of this invention are as follows: This invention discloses a forwarding mechanism and transmission method for network data packets with latency concerns. Addressing the problems that existing network technologies (IntServ, DiffServ, Traffic Engineering, Segment Routing, TSN, etc.) either cannot guarantee end-to-end transmission latency, or whose technical complexity and cost limit their applicability to small areas (e.g., local area networks) and thus cannot be used in large area networks (e.g., wide area networks), this invention provides a low-cost, highly compatible, highly scalable forwarding mechanism and transmission method applicable to various regional networks (including wide area networks spanning thousands of kilometers), thus guaranteeing network latency.

[0035] For situations with strict latency requirements for network data packet transmission, current data communication networks, while guaranteeing data packets will arrive earlier or later, cannot guarantee on-time delivery. If a data packet fails to reach its destination on time, it becomes unusable upon arrival, sometimes causing accidents and even loss of life and property. Examples include remotely controlling robots, vehicles, factory production lines, mining robots, port handling machines, and robots operating in hazardous environments. Using the forwarding mechanism and transmission method of this invention, the network can employ different data transmission paths according to the varying latency requirements of data packets, thereby ensuring timely delivery.

[0036] The method of this invention can be implemented within network devices (routers and switches) without the need for circular queues and forwarding (CQF) and flow-based filtering and control (PSFP) mechanisms, and can be deployed in large area networks covering hundreds or even thousands of kilometers. Attached Figure Description

[0037] Figure 1 This is a schematic diagram of data packet transmission for two references in Embodiment 2 of the present invention. Detailed Implementation

[0038] The present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are for illustrative purposes only and are not intended to limit the scope of the invention.

[0039] Example 1

[0040] A network packet forwarding mechanism and transmission method that addresses latency concerns is typically implemented in the control card of a network device, and specifically includes the following steps:

[0041] Step S1: Configure the maximum latency target (MLT) that the network system can support. Once the MLT is configured, the latency target value in each packet must be less than or equal to the MLT. A network has only one MLT.

[0042] Step S2: For each link in the network, configure its link state parameters to the latency value of that link. That is, if a data packet is transmitted from a router A to its next-hop router B, then the latency consumed by the data packet on the link from A to B is the link state value of that link.

[0043] Step S3: Generate a Link State Database (LSDB) using a routing protocol. For example, OSPF or IS-IS can be used to generate the LSDB. The LSDB can store various related information. Corresponding to this invention, the LSDB should at least store the following information:

[0044]

[0045] Here, Router is a network device (router or switch), Next-Hop is the next network device directly connected to Router (usually called the next hop), Outgoing Interface is the connection from the exit point of Router to Next-Hop, and Latency is the latency value of the link from Router to Next-Hop.

[0046] When generating the routing information table, we also need the DASN (Directly Attached Subnets) information of the subnets directly connected to the router, in the following format:

[0047]

[0048] Here, Next-Hop is set to "directed", Latency is set to 0. Router and Next-Hop refer to the router and the next-hop router, respectively. Subnet Address is the IP address of the subnet directly connected to the Router. SubnetAddress Prefix Length is the effective prefix length of the Subnet Address. Outgoing Interface is the exit point on the Router that is directly connected to the subnet represented by Subnet Address and Subnet Address Prefix Length.

[0049] Information about subnets directly connected to the router is typically considered part of the LSDB and stored within it. For ease of explanation, we will distinguish this information by storing it in a separate table, which we will call the Directly Connected Subnets to Router (DASN) table.

[0050] Step S4: Generate a latency-carrying forwarding path database (LC-PDB) from the LSDB. A forwarding path is defined as a sequence of network devices: <R1,R2,…,R n >, where R1 is the current router, R i+1 It is R i The next hop. A latency-carrying path is defined as a tuple (p, t), where p is a forwarding path and t is the sum of the delays on all links in forwarding path p. The method for generating the latency-carrying path database LC-PDB is as follows:

[0051] Step S4-1: Define a set P to store forwarding paths for which latency is a concern, and define another set PC to store candidates for forwarding paths for which latency is a concern. Assume the current router is s.

[0052] Step S4-2: Place ( <s>,0) is added to set P.

[0053] Step S4-3: For all connected next-hop routers v of the current router s, assuming the link-state value from s to v is t, if t ≤ MLT, then (<s,v> ,t) is added to PC.

[0054] Step S4-4: If PC is empty, then jump to step 4-6; otherwise, execute the following steps in sequence:

[0055] Step S4-4-1: Extract a candidate forwarding path (q, t) from the PC;

[0056] Step S4-4-2: Remove (q,t) from set PC;

[0057] Step S4-4-3: Add (q,t) to P;

[0058] Step S4-4-4: Assume the last router in the forwarding path q is U. Search the LSDB to find all routers V1, V2, ... connected to U, excluding directly connected subnets. For each router V directly connected to U... i Assuming a transition from U to V i If the link delay is m, and t+m≤MLT and V i Since it didn't appear in q, then we'll leave ( <q,V i >,t+m) are added to PC, where <q,V i > indicates that V i A new sequence generated by appending to the end of sequence q.

[0059] Step S4-5: Return to step 4-4 until PC is empty.

[0060] Step S4-6: At this point, the set P is the forwarding path database LC-PDB for which latency is of concern.

[0061] Step S5: Generate a latency-carried routing information table (LC-RIB) from the latency-carried forwarding path database (LC-PDB). The LC-RIB contains the following information: (IP Address, IP Address Prefix Length, Outgoing Interface, Next-Hop, Latency).

[0062] Assuming the current router is A, the LC-RIB generation method follows these steps:

[0063] Step S5-1: If the LC-PDB database for forwarding paths of concern regarding latency is empty, proceed to step 5-3. Otherwise, for each forwarding path (LCP) of concern regarding latency in the LC-PDB, perform the following steps:

[0064] Step S5-1-1: Assume LCP = ( ,0). A is the current network router.

[0065] Step S5-1-1-1: Add the routing information of router A's own address (also known as the loopback address) to the LC-RIB:

[0066] (A, 32, loopback, directed, 0)

[0067] Where A is its loopback address. Assuming A is an IPv4 address, 32 is the prefix length of its loopback address. In an IPv6 network, the prefix length of A's loopback address is 128. For other network types, the prefix length depends on the specific network address.

[0068] A router typically has one or more loopback addresses. All of these loopback addresses are added in the manner described above.

[0069] Step S5-1-1-2: Look up the DASN. For each subnet directly connected to A: Assuming the subnet is S / p, where S is an IP address, p is the address prefix length of S, and the connected interface is IF, then add the following information to the LC-RIB:

[0070] (S,p,IF,directed,0)

[0071] Step S5-1-2: Assume LCP = (<A,B> Let A be the current router, and B be the direct next-hop router of A. Assume the network exit point from A to B is IF.

[0072] Step S5-1-2-1: Add the routing information of all self-addresses (also known as loopback addresses) of router B to the LC-RIB:

[0073] (B, 32, IF, B, L)

[0074] If this is an IPv4 network, the prefix length of B's ​​own address is 32. If it is an IPv6 network, the prefix length of B's ​​own address is 128. If it is another type of network, the prefix length depends on that type of network address.

[0075] Step S5-1-2-2: For each subnet directly connected to B: Assuming the subnet is S / p, where S is an IP address and p is the address prefix length of S, then add the following information to LC-RIB:

[0076] (S,p,IF,B,L)

[0077] Step S5-1-3: Assume LCP = (<A,B,…,X> Let A be the current router, B be the direct next-hop router, and X be the last router on this forwarding path. Assume the network exit from A to B is IF.

[0078] Step S5-1-3-1: Add the routing information of all loopback addresses of router X to the LC-RIB:

[0079] (X,32,IF,B,L)

[0080] If this is an IPv4 network, the prefix length of X's own address is 32. If it is an IPv6 network, the prefix length of X's own address is 128. If it is another type of network, the prefix length depends on that type of network address.

[0081] Step S5-1-3-2: For each subnet directly connected to X: Assuming the subnet is S / p, where S is an IP address and p is the address prefix length of S, then add the following information to the LC-RIB:

[0082] (S,p,IF,B,L)

[0083] Step S5-2: Return to step S5-1 until all forwarding paths of concern regarding latency in LC-PDB have been processed.

[0084] Step S5-3: LC-RIB is the routing information table on the current router that is concerned with latency.

[0085] Step S6: Generate a latency-carried forwarding information base (LC-FIB) from the latency-carried routing information base (LC-RIB). A latency-carried forwarding information base consists of multiple latency-related general forwarding information bases (FIBs). A general forwarding information base (FIB) contains at least the following information:

[0086] Address Prefix Address Prefix Length Outgoing Interface Next-Hop

[0087] The method for generating the LC-FIB (Leadership Information Table) for latency concerns is as follows:

[0088] Step S6-1: For each possible delay, generate an empty general forwarding information table. When the network's maximum delay target is MLT, then there are 0 to MLT general forwarding information tables. We use FIB. t Let FIB0, FIB1, ..., FIB be the ordinary forwarding information table corresponding to the time delay t. MLT This corresponds to the forwarding information table from 0 to MLT with a latency of 1.

[0089] Step S6-2: For each routing information (prefix, prefix-length, out-if, next-hop, latency) in LC-RIB, add (prefix, prefix-length, out-if, next-hop) to FIB. latency Go to the middle.

[0090] Step S6-3: Generate FIB0, FIB1, ..., FIB... MLT Installed in the data plane of network equipment, it is usually its line card.

[0091] Step S7: Generate the Next-Hop LinkLatency Table (NHLLT) from the LSDB for directly connected links. Each entry in the NHLLT must contain at least the following information:

[0092] (Next-Hop, Link-Latency)

[0093] Generate NHLLT by following these steps:

[0094] Step S7-1: If all entries in the LSDB have been processed, step S7 is complete. Otherwise, extract one unprocessed entry from the LSDB, let's say (Router, Outgoing-Interface, Next-Hop, Latency).

[0095] Step S7-2: If the Router is not the current network device, mark the entry to indicate that it has been processed. Return to step S7-1.

[0096] Step S7-3: If the Router is the current network device itself, add (Next-Hop, Latency) to the NHLLT. Mark this entry to indicate that it has been processed.

[0097] Step S7-4: Return to step S7-1 until all entries in LSDB have been processed.

[0098] Step S7-5: Install the generated NHLLT in the data plane of the network device, usually its line card.

[0099] Step S8: Whenever the MLT configuration changes, steps S4 to S7 above must be executed again.

[0100] Step S9: Whenever the network topology or link status changes, steps S3 to S7 above must be executed again.

[0101] When the network topology and link states remain stable, the LC-FIB and NHLLT obtained through the above steps will be in a stable state. They can then be used to forward and transmit network data packets with latency requirements.

[0102] LC-FIB can be used to support various packet forwarding and transmissions where latency is a concern. Different applications have different latency requirements and different requirements for latency error. In the following steps, we use (t,j) to represent the application's latency requirement, where t is the latency (in milliseconds) and j is the latency error (in milliseconds). This means that the data packet should be delivered within t milliseconds, and the latency error should not exceed j milliseconds. In other words, the latency to the destination can be within [tj,t+j], including tj milliseconds and t+j milliseconds. The latency error is also called latency jitter. For example, "latency 20 milliseconds, speed not exceeding 2 milliseconds" can be expressed as (20,2).

[0103] Steps S10 and S11 below are used on the data plane of the network device. The data plane is usually implemented using a line card.

[0104] Step S10: Add two new fields to the data packet to carry the delay and delay error values. For example, if a data packet requires a delay of 20 milliseconds, with a speedup of no more than 2 milliseconds, then place 20 and 2 in these two fields. Although the specific format of various data packets differs, as long as the data packet contains two fields to hold the delay parameter, it is acceptable. If IPv4 and IPv6 are used for data packet encapsulation, extensions to IPv4 and IPv6 are required. This invention does not involve specific data packet formats. The general format of a data packet containing delay fields is as follows:

[0105]

[0106] The delay target and delay error target can be encapsulated as separate independent regions, or they can be encapsulated in the header of the data packet, or the regions in the original data packet header can be reused to carry delay and delay error.

[0107] The following step S11 describes the processing of the received data packets.

[0108] Step S11: If there is no packet in the Incoming Packet Buffer, wait. If there is a packet, then take out a new packet from it and forward and transmit it by following the steps below:

[0109] Step S11-1: Parse the received packet. According to the specific encapsulation format of the packet, parse the packet header to obtain the destination address DEST, latency target LAT, and jitter target JIT of the packet

[0110] Step S11-2: If the destination address DEST is the current router itself, then the packet is handed over to the current router (usually its control card) for its own processing. After the packet is processed, return to step 11 to process the next packet.

[0111] Step S11-3: In each of the following FIB M if there exists a 0 < T ≤ JIT such that (LAT - T) < 0, then set FIB LAT-T to be empty. Sequentially search for the best match of DEST (the longest address prefix match of DEST) in FIB LAT , FIB LAT-1 , FIB LAT+1 , ……, FIB LAT-JIT , FIB LAT+JIT ; or sequentially search for the best match of DEST (the longest address prefix match of DEST) in FIB LAT , FIB LAT+1 , FIB LAT-1 , ……, FIB LAT+JIT , FIB LAT-JIT .

[0112] Step S11-3-1: If no best match of DEST is found in all of the above FIB LAT , FIB LAT-1 , FIB LAT+1 , ……, FIB LAT-JIT , FIB LAT+JIT , discard the packet. After the packet is processed, return to step S11 to process the next packet.

[0113] Step S11-3-2: If there exists a FIB T in which the best match of DEST is found, thus obtaining the following forwarding information entry:

[0114] (Prefix, Prefix-Length, Outgoing-Interface, Next-Hop)

[0115] Step S11-3-2-1: If Next-Hop is "directed", it means there is a directly connected subnet. Send this data packet from the outgoing-interface to the directly connected subnet. After processing this data packet, return to step S11 to process the next data packet.

[0116] Step S11-3-2-2: Find the delay table entry corresponding to the next-hop of the directly connected link delay table NHLLT from the directly connected link delay table:

[0117] (Next-Hop,LL)

[0118] Step S11-3-2-3: If LL>LAT+JIT, return to step S11-3 and query the next normal forwarding information table in sequence.

[0119] Step S11-3-2-4: If LL <= LAT, modify the delay target value in the data packet, replacing the original delay target value LAT in the data packet with the LAT-LL value, and construct a new data packet. According to the encapsulation requirements of the Outgoing-Interface, send the newly constructed data packet out of the Outgoing-Interface. After this data packet is processed, return to step S11 to process the next data packet.

[0120] Step S11-3-2-5: If LL = LAT + JIT, replace the target delay value LAT in the original data packet with 0, and replace the delay error value JIT with 0, constructing a new data packet. According to the encapsulation requirements of the Outgoing-Interface, send the newly constructed data packet out of the Outgoing-Interface. After processing this data packet, return to step 11 to process the next data packet.

[0121] Step S11-3-2-6: Set NLH = LAT + JIT – LL. Arbitrarily set NLAT and NJIT such that the effective positive value interval of the mathematical interval [NLAT - NJIT, NLAT + NJIT] is equivalent to the mathematical interval [0, NLH], where the effective positive value interval is defined as follows:

[0122] If m≥0 and n>0, then the effective positive value interval of the mathematical interval [m,n] is equal to its mathematical interval [m,n].

[0123] If m < 0 and n > 0, then the valid positive value interval of the mathematical interval [m, n] is equal to the mathematical interval [0, n].

[0124] In other cases, the valid positive value interval of the mathematical interval [m,n] is equal to the mathematical interval [0,0].

[0125] A simple way to set this is to set NLAT=0 and NJIT=NLH.

[0126] Step S11-3-2-7: Replace the target latency value (LAT) in the original data packet with the NLAT value, and replace the latency error value (JIT) with the NJIT value to construct a new data packet. According to the encapsulation requirements of the Outgoing-Interface, send the newly constructed data packet out through the Outgoing-Interface. This data packet processing is complete; return to step S11 to process the next data packet.

[0127] Example 2

[0128] This implementation example Figure 1 , Figure 1 It contains a demonstrative network in which the parameters in each link represent the latency. Figure 1 The system also includes two applications: one for querying temperature and the other for controlling a car to turn left. The temperature query application has no strict latency requirements and can be delivered later, while the car direction control command application has strict latency requirements. If the command is delivered late, it may cause traffic accidents, resulting in damage to people and property. This embodiment uses the forwarding mechanism and transmission of this invention in this application scenario. Assuming A is our current router, we will now construct a latency-sensitive forwarding information table in router A, specifically including the following steps:

[0129] Step S1: Configure the maximum latency target (MLT) in the network. In this embodiment, MLT = 22 is configured.

[0130] Step S2: Configure the status parameters of each link in the network to the corresponding delay value of that link.

[0131] Step S3: Generate the network link-state database (LSDB), as follows:

[0132]

[0133]

[0134] Network links are typically bidirectional. In this embodiment, for illustrative purposes, only unidirectional link states are listed above.

[0135] The subnet information table directly connected to the router is as follows:

[0136]

[0137] Step S4: Generate a forwarding path database (LC-PDB) for which latency is a concern from the LSDB:

[0138] Taking A as the current router, the LC-PDB forwarding path database, which is of concern regarding latency, is as follows:

[0139] ( ,0);

[0140] (<A,D> ,2);

[0141] (<A,B> ,3);

[0142] (<A,D,C> ,7);

[0143] (<A,B,F> ,10);

[0144] (<A,D,E> ,12);

[0145] (<A,D,C,E> ,13);

[0146] (<A,B,C> ,15);

[0147] (<A,D,E,C> ,18);

[0148] (<A,D,C,B> ,19);

[0149] (<A,B,F,E> ,20);

[0150] (<A,B,C,D> ,20);

[0151] (<A,B,C,E> ,twenty one);

[0152] (<A,D,E,F> ,twenty two);

[0153] In this context, router names A, B, C, D, E, and F represent their respective IP addresses.

[0154] Step S5: Generate a routing information table LC-RIB for latency concerns from the forwarding path database for which latency is a concern.

[0155]

[0156]

[0157] Step S6: Generate a delay-sensitive forwarding information table (LC-FIB) from the delay-sensitive routing information table (LC-RIB). In this embodiment, MLT = 22, therefore, we have one ordinary forwarding information table corresponding to each possible delay. In this example, we have a total of 23 ordinary forwarding information tables as follows:

[0158] FIB0:

[0159]

[0160] FIB1: Empty

[0161] FIB2:

[0162] IP Address IP Address Prefix Length Outgoing Interface Next-Hop 192.168.100.40 32 eth3 192.168.100.40 10.110.0.0 16 eth3 192.168.100.40

[0163] FIB3:

[0164]

[0165]

[0166] FIB4, FIB5, and FIB6 are empty.

[0167] FIB7:

[0168] IP Address IP Address Prefix Length Outgoing Interface Next-Hop 192.168.100.30 32 eth3 192.168.100.40

[0169] FIB8 and FIB9 are empty.

[0170] FIB 10 :

[0171] IP Address IP Address Prefix Length Outgoing Interface Next-Hop 192.168.100.200 32 eth2 192.168.100.20 10.200.0.0 16 eth2 192.168.100.20

[0172] FIB 11 empty

[0173] FIB 12 :

[0174] IP Address IP Address Prefix Length Outgoing Interface Next-Hop 192.168.100.50 32 eth3 192.168.100.40 10.130.0.0 16 eth3 192.168.100.40

[0175] FIB 13 :

[0176] IP Address IP Address Prefix Length Outgoing Interface Next-Hop 192.168.100.50 32 eth3 192.168.100.40 10.130.0.0 16 eth3 192.168.100.40

[0177] FIB 14 empty

[0178] FIB 15 :

[0179] IP Address IP Address Prefix Length Outgoing Interface Next-Hop 192.168.100.30 32 eth2 192.168.100.20

[0180] FIB 16 FIB 17 empty

[0181] FIB 18 :

[0182] IP Address IP Address Prefix Length Outgoing Interface Next-Hop 192.168.100.30 32 eth3 192.168.100.40

[0183] FIB 19 :

[0184] IP Address IP Address Prefix Length Outgoing Interface Next-Hop 192.168.100.20 32 eth3 192.168.100.40 10.150.0.0 16 eth3 192.168.100.40

[0185] FIB 20 :

[0186] IP Address IP Address Prefix Length Outgoing Interface Next-Hop 10.130.0.0 16 eth2 192.168.100.20 10.110.0.0 16 eth2 192.168.100.20 192.168.100.50 32 eth2 192.168.100.20 192.168.100.40 32 eth2 192.168.100.20

[0187] FIB 21 :

[0188]

[0189]

[0190] FIB 22 :

[0191] IP Address IP Address Prefix Length Outgoing Interface Next-Hop 10.200.0.0 16 eth3 192.168.100.40 192.168.100.200 32 eth3 192.168.100.40

[0192] Step S7: Generate the Next-Hop Link Latency Table (NHLLT) from the LSDB as follows:

[0193] Next-Hop Latency 192.168.100.20 3 192.168.100.40 2

[0194] At this point, we have completed the control information required for forwarding latency-sensitive packets: LC-FIB and NHLLT. We will now use them to forward packets corresponding to... Figure 1 Data packets for two applications.

[0195] Step S8: When the network device receives a data packet with embedded latency requirements, it forwards the data packet according to the LC-FIB (Latency-Concerned Forwarding Information Table).

[0196] First, construct the data packets corresponding to the two applications:

[0197] The first data packet corresponds to the car control command, and we require a latency of 10 milliseconds and a latency error of 1 millisecond.

[0198] 10.100.210.20 10.200.165.50 10 1 Command "Turn Left”

[0199] The second data packet, corresponding to the temperature query, requires a latency of 20 milliseconds and a latency error of 2 milliseconds.

[0200] 10.120.21.80 10.200.200.30 20 2 "What’s the temperature”

[0201] Of course, 20 milliseconds is still a very small value in practical applications; it is used here as an example.

[0202] For transmitting the first data packet:

[0203] Suppose that at this point, the network device receives the first data packet.

[0204] Step S11-1: Parse the received data packet. Based on the specific encapsulation format of the data packet, parse the packet header to obtain the destination address DEST = 10.200.165.50, the delay target LAT = 10, and the delay error target JIT = 1.

[0205] Step S11-2: The destination address DEST=10.200.165.50 is not the current router itself, continue to step S11-3;

[0206] Step S11-3: In FIB 10 FIB9, FIB 11 The best match is found in the sequence 10.200.165.50.

[0207] Step S11-3-2: In FIB 10 The best match was found to be 10.200.165.50, resulting in the following forwarding information entry:

[0208] IP Address IP Address Prefix Length Outgoing Interface Next-Hop 10.200.0.0 16 eth2 192.168.100.20

[0209] Step S11-3-2-1: Next-Hop is not "directed", continue to the next step.

[0210] Step S11-3-2-2: Find the delay table entry for the directly connected next hop corresponding to Next-Hop192.168.100.20 from the directly connected link delay table NHLLT:

[0211] Next-Hop Latency 192.168.100.20 3

[0212] The delay LL = 3 in NHLLT.

[0213] Step S11-3-2-3: The condition is not met, continue to step S11-3-2-4.

[0214] Step S11-3-2-4: Since 3 <= 10, LL <= LAT. Modify the delay target value in the data packet, replacing the original delay target value of 10 with 7, and construct a new data packet:

[0215] 10.100.210.20 10.200.165.50 7 1 Command "Turn Left”

[0216] According to the encapsulation requirements of the eth2 egress, the newly constructed data packet is sent to router B (192.168.100.20) through the eth2 egress. After processing, return to step S11 to process the next data packet.

[0217] Suppose that at this point, the network device receives the second data packet.

[0218] Step S11-1: Parse the second data packet to obtain its destination address DEST = 10.200.200.30, delay target LAT = 20, and delay target error JIT = 2.

[0219] Step S11-2: The target address DEST=10.200.200.30 is not the current router itself, continue to step S11-3.

[0220] Step S11-3: In FIB 20 FIB 19 FIB 21 FIB 18 FIB 22 The best match is found in the sequence 10.200.200.30.

[0221] Step S11-3-2: In FIB 22 The best match was found to be 10.200.200.30. The following forwarding information table was obtained:

[0222] IP Address IP Address Prefix Length Outgoing Interface Next-Hop 10.200.0.0 16 eth3 192.168.100.40

[0223] Step S11-3-2-1: If Next-Hop is not "directed", proceed to the next step.

[0224] Step S11-3-2-2: Find the delay table entry of the directly connected next hop corresponding to Next-Hop from the directly connected link delay table NHLLT:

[0225] Next-Hop Latency 192.168.100.40 2

[0226] The link delay LL of the link connected to the next hop 192.168.100.40 is 2.

[0227] Step S11-3-2-3: The condition is not satisfied, continue to Step S11-3-2-4.

[0228] Step S11-3-2-4: Since 2 < 20, that is, LL < LAT, modify the delay target value in the data packet, replace the original delay target value LAT in the data packet with the value of LAT–LL, that is, replace 20 with 18, and construct a new data packet:

[0229] 10.120.21.80 10.200.200.30 18 2 "What’s the temperature”

[0230] According to the encapsulation requirements of the egress eth3, send the newly constructed data packet to the router D (192.168.100.40) through the egress eth3. This data packet processing is completed, and return to Step S8 to process the next data packet.

[0231] In summary, this invention extends the Link State Database (LSDB) beyond the traditional network layer packet processing pipeline. It designs processes and methods for generating a Latency-Cared Path Database (LC-PDB) from the LSDB, generating a Latency-Cared Routing Information Base (LC-RIB) from the LC-PDB, generating a Latency-Cared Forwarding Information Base (LC-FIB) from the LC-RIB, and finally, designing the processing steps for routers and switches to forward and transmit data packets using the LC-FIB. After executing the methods and processes of this invention, the time taken for a data packet to reach its destination is guaranteed to be within the specified latency. This invention is applicable to wide area networks (WANs) for data communication and can support and implement network applications spanning thousands of kilometers.

[0232] It should be noted that the above content merely illustrates the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. For those skilled in the art, various improvements and modifications can be made without departing from the principle of the present invention, and all such improvements and modifications fall within the scope of protection of the claims of the present invention.< / s>

Claims

1. A network data packet forwarding mechanism that considers latency, characterized by: This forwarding mechanism comprises a control plane and a data plane: In the control plane, link state database is generated using link delay as the link state, then forwarding path database with relevant delay is generated from the link state database, then routing information table with relevant delay is generated, and finally forwarding information table with relevant delay is generated. In the data plane, two regions are added to the network data packets to carry the data packet transmission delay target and delay error target, respectively. When a data packet passes through a network device, the target address of the data packet is searched and matched in the forwarding information table that is concerned with delay to obtain the corresponding forwarding information, and the data packet is sent according to the forwarding information.

2. The network data packet forwarding mechanism for network packets with latency concerns as described in claim 1, characterized in that: The forwarding information table is generated in the following specific way: Generate a link-state database (LSDB) containing information about the latency of all links; Based on the link state database LSDB, a forwarding path database LC-PDB for which latency is a concern is constructed. Based on the forwarding path database LC-PDB, a routing information table LC-RIB for latency is constructed and generated; in the routing information table LC-RIB for latency, there is a corresponding ordinary routing information table for different network latency parameters. Based on the routing information table LC-RIB which is concerned with latency, a forwarding information table LC-FIB which is concerned with latency is constructed. The forwarding information table LC-FIB which is concerned with latency includes a latency target table and a normal forwarding information table. For each latency value in the latency target table, there is a corresponding normal forwarding information table.

3. The network data packet forwarding mechanism for latency-sensitive networks as described in claim 2, characterized in that: The Link State Database (LSDB) includes at least two types of information: (1) next-hop and latency information; and (2) DASN information of subnets directly connected to the router. The next hop and latency information are formatted as follows: Router, Outgoing Interface, Next-Hop, and Latency. Here, Router is a network device, Next-Hop is the next network device directly connected to Router, Outgoing Interface is the connection from the exit point of Router to Next-Hop, and Latency is the latency value of the link from Router to Next-Hop. The format of the DASN information of the subnet directly connected to the router is as follows: Router, Next-Hop, Subnet Address, Subnet Address Prefix Length, Latency, and Outgoing Interface. Next-Hop is set to "directed", Latency is set to 0, Subnet Address is the IP address of the subnet directly connected to the Router, Subnet Address Prefix Length is the effective prefix length of the Subnet Address, and Outgoing Interface is the exit point on the Router that is directly connected to the subnet represented by Subnet Address and Subnet Address Prefix Length.

4. The network data packet forwarding mechanism for latency-sensitive networks as described in claim 2, characterized in that: The routing information table LC-RIB, which is concerned with latency, includes at least IP Address, IP Address Prefix Length, OutgoingInterface, Next-Hop, and Latency. The general forwarding information table includes at least Address Prefix, Address Prefix Length, OutgoingInterface, and Next-Hop.

5. A method for transmitting network data packets of concern to latency using the forwarding mechanism as described in claim 1, characterized in that, Includes the following steps: S1: Configure the maximum latency target (MLT) in the network; S2: Configure the state parameters of each link in the network to the latency value of that link; S3: Based on OSPF or IS-IS protocols, publish and exchange the status of each link, or generate a Link State Database (LSDB) through network management system or network controller configuration; S4: Generate a routing information table (LC-RIB) for latency concerns from the Link State Database (LSDB). The routing table in the LC-RIB contains the latency for forwarding to the next hop. This includes the following steps: S41: Generate a forwarding path database (LC-PDB) from the LSDB, focusing on latency; a forwarding path is defined as a sequence of network devices: <R1,R2,…,R n >, where R1 is the current router, R i+1 It is R i The next jump; A forwarding path of concern is defined as a tuple (p,t), where p is a forwarding path and t is the sum of the delays on each link in forwarding path p, and t<=MLT; S42: Generate a routing information table LC-RIB (which controls latency) from the forwarding path database LC-PDB (which controls latency). Steps S41 and S42 can be completed separately, or the two steps can be combined into one step. S5: Generate a delay-sensitive forwarding information table LC-FIB from the delay-sensitive routing information table LC-RIB; the delay-sensitive forwarding information table LC-FIB is composed of multiple ordinary forwarding information tables FIB: set (MLT+1) empty ordinary forwarding information tables FIB0, FIB1, ..., FIB corresponding to delays of 0, 1 to MLT. MLT， For each routing information (prefix, prefix-length, out-if, next-hop, latency) in LC-RIB, add (prefix, prefix-length, out-if, next-hop) to FIB. latency middle; S6: Generate the link delay table NHLLT for directly connected links from the LSDB; S7: Embed the latency requirement for the data packet to reach the target location in the network data packet, wherein the data packet contains transmission latency parameters: latency and jitter; or the data packet contains minimum latency parameters and maximum latency parameters; S8: When a network device receives a data packet with embedded latency requirements, it forwards the data packet according to the LC-FIB (Local Latency Information Table) for which latency is a concern.

6. The method for transmitting network data packets with latency concerns as described in claim 5, characterized in that: Step S41 specifically includes the following steps: S411: Set set P to store forwarding paths for which latency is a concern, and set another set PC to store candidates for forwarding paths for which latency is a concern; S412: Will ( <s> ,0) are added to set P, where s is the current router;< / s> <s> S413: For all connected next-hop routers v of the current router s, assuming the link-state value from s to v is t, if t ≤ MLT, then (<s,v> Add ,t) to PC; S414: If PC is empty, proceed to step S416; otherwise, execute the following steps in sequence: S414-1: Take a candidate forwarding path (q, t) from the PC; where q is the forwarding path; S414-2: Remove (q,t) from set PC; S414-3: Add (q,t) to P; S414-4: Assuming the last router in the forwarding path q is U, search the LSDB to find all routers V1, V2, ... connected to U, excluding directly connected subnets; for each router V directly connected to U... i Assuming a transition from U to V i If the link delay is m, and t+m≤MLT and V i If it has not appeared in q, then ( <q,V i >,t+m) are added to PC, where <q,V i > indicates that V i A new sequence generated by appending to the end of sequence q; S415: Return to step S414 until PC is empty; S416: Set P is the forwarding path database LC-PDB for which latency is a concern; S417: Whenever the network maximum latency target (MLT) changes, the above steps S411-S416 need to be executed again. S418: Whenever the network status changes and causes the LSDB to change, the above steps S411-S416 need to be executed again.

7. The method for transmitting network data packets with latency concerns according to claim 5, characterized in that: Step S42 specifically includes the following steps: S420: Set LC-RIB to empty; S421: If the LC-PDB database of forwarding paths of concern regarding latency is empty, proceed to step S423; otherwise, for each forwarding path LCP of concern regarding latency in the LC-PDB, perform the following steps: S421-1: Assume LCP = ( ,0), where A is the current network router; S421-1-1: Add the routing information of router A's own address to the LC-RIB. The routing information includes (A, 32, loopback, directed, 0), where A is its own address. Assuming A is an IPv4 address, 32 is the prefix length of its own address. If it is an IPv6 network, the prefix length of A's own address is 128. If it is another type of network, the prefix length depends on the network address of that other type. S421-1-2: Locate the subnet DASN directly connected to the router. For each subnet directly connected to A: Assume that the subnet is S / p, where S is an IP address, p is the address prefix length of S, and the connected interface is IF. Then add (S, p, IF, directed, 0) to LC-RIB. S421-2: Assume LCP = (<A,B> Given a network path (A, B), where A is the current router and B is A's direct next-hop router, and assuming the network exit point from A to B is IF, then perform the following steps: S421-2-1: Add all routing information (B, 32, IF, B, L) for router B's own address to the LC-RIB. If it is an IPv4 network, the prefix length of B's ​​own address is 32; if it is an IPv6 network, the prefix length of B's ​​own address is 128; if it is another type of network, the prefix length depends on the network address of that other type. S421-2-2: For each subnet directly connected to B: Assuming the subnet is S / p, where S is an IP address and p is the address prefix length of S, then add (S, p, IF, B, L) to LC-RIB. S421-3: Assume LCP = (<A,B,…,X> Let A be the current router, B be the direct next-hop router, and X be the last router on the forwarding path. Assuming the network exit from A to B is IF, then the following steps are performed: S421-3-1: Add all routing information (X, 32, IF, B, L) for router X's own address to the LC-RIB. When it is an IPv4 network, the prefix length of X's own address is 32; when it is an IPv6 network, the prefix length of X's own address is 128; when it is another type of network, the prefix length depends on the network address of that other type. S421-3-2: For each subnet directly connected to X: Assuming the subnet is S / p, where S is an IP address and p is the address prefix length of S, then add (S, p, IF, B, L) to LC-RIB. S422: Return to step S421 until all forwarding paths of concern regarding latency in LC-PDB have been processed; S423: LC-RIB is the routing information table on the current router that addresses latency concerns.

8. The method for transmitting network data packets with latency concerns according to claim 5, characterized in that: Step S6 specifically includes the following steps: S61: Each entry in the NHLLT (Link Latency Table) must include at least Next-Hop and Link-Latency; S62: For each entry in the LSDB, assuming it is (Router, Outgoing-Interface, Next-Hop, Latency), if Router is not the current network device, ignore it; otherwise, add (Next-Hop, Latency) to NHLLT.

9. The method for transmitting network data packets with latency concerns according to claim 5, characterized in that: In step S8, when the data packet contains the destination address DEST, the delay target LAT, and the delay error target JIT, the following steps are performed: S81: In each of the following FIBs M if there exists 0 < T ≤ JIT such that (LAT - T) < 0, then set FIB LAT-T to be empty; sequentially search for the best matching DEST in FIB LAT , FIB LAT-1 , FIB LAT+1 , ……, FIB LAT-JIT , FIB LAT+JIT , or sequentially search for the best matching DEST in FIB LAT , FIB LAT+1 , FIB LAT-1 , ……, FIB LAT+JIT , FIB LAT-JIT ; where FIB M represents the general forwarding information table corresponding to the delay M. S82: If in all of the above FIBs LAT FIB LAT-1 FIB LAT+1 ..., FIB LAT-JIT FIB LAT+JIT If no best match for DEST is found in the data packet, the packet is discarded and the data packet processing is complete. S83: If a FIB exists T The best match DEST is found in the table, which yields the following forwarding information entry: (Prefix, Prefix-Length, Outgoing-Interface, Next-Hop). The packet being processed is forwarded according to this forwarding information.

10. The method for transmitting network data packets with latency concerns according to claim 9, characterized in that: Step S83 specifically includes the following steps: S831: If Next-Hop is "directed", it means there is a directly connected subnet. Send this packet from the outgoing-interface to the directly connected subnet; the packet processing is complete. S832: Find the delay entry corresponding to the directly connected next hop of Next-Hop from the NHLLT generated in step S62, that is, the delay value from this device to Next-Hop, set as LL, and make a judgment: If LL > LAT + JIT, query the next normal forwarding information table in sequence; If LL <= LAT, modify the delay target value in the data packet, replace the original delay target value LAT in the data packet with the LAT-LL value, construct a new data packet, and send the newly constructed data packet out of the Outgoing-Interface according to the encapsulation requirements of the Outgoing-Interface. The data packet processing is complete. If LL = LAT + JIT, replace the delay target value LAT in the original data packet with 0, replace the delay error value JIT with 0, construct a new data packet, and send the newly constructed data packet out of the Outgoing-Interface according to the encapsulation requirements of the Outgoing-Interface. The data packet is then processed. S833: Set NLH = LAT + JIT – LL, and arbitrarily set NLAT and NJIT such that the effective positive value interval of the mathematical interval [NLAT - NJIT, NLAT + NJIT] is equivalent to the mathematical interval [0, NLH], where the effective positive value interval is defined as follows: If m≥0 and n>0, then the effective positive value interval of the mathematical interval [m,n] is equal to its mathematical interval [m,n]. If m < 0 and n > 0, then the valid positive value interval of the mathematical interval [m, n] is equal to the mathematical interval [0, n]. In other cases, the valid positive value interval of the mathematical interval [m,n] is equal to the mathematical interval [0,0]. S834: Replace the target delay value LAT in the original data packet with the NLAT value, replace the delay error value JIT with the NJIT value, construct a new data packet, and send the newly constructed data packet out of the Outgoing-Interface according to the encapsulation requirements of the Outgoing-Interface. The processing of this data packet is complete. < / s>

Images