An edge computing network architecture providing deterministic quality of service
By dividing the edge computing network architecture into wireless, wired, and computing domains, and employing deterministic networking technology and traffic shaping mechanisms, the problem of traditional deterministic networking technology being unable to adapt to heterogeneous media is solved, thus achieving deterministic quality of service assurance for edge computing networks.
Patent Information
- Application Number
- CN202210225755.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-03-09
- Publication Date
- 2026-02-10
- Estimated Expiration
- 2042-03-09
AI Technical Summary
Traditional edge computing network systems cannot provide deterministic quality of service guarantees in heterogeneous media scenarios. Existing deterministic network technologies are only applicable to wired networks and cannot be adapted to wireless and computing node components.
The edge computing network architecture is divided into a three-domain deployment approach, consisting of a wireless domain, a wired transmission domain, and a computing domain. Each domain uses LTE/5G standards, deterministic network technology, and equal-length time-period resource allocation. Deterministic transmission and processing are achieved through an inter-domain traffic shaping mechanism.
It provides deterministic quality of service guarantees in dynamic environments, with controllable and bounded system response time and jitter, and is suitable for heterogeneous media scenarios in edge computing networks.
Smart Images

Figure CN114650566B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of network architecture, and more specifically to an edge computing network architecture that provides deterministic quality of service. Background Technology
[0002] In edge computing network systems, users can upload locally generated tasks to computing nodes in the network, saving local resource costs (such as CPU, memory, and battery). However, traditional edge computing network systems employ a "best-effort" forwarding mechanism during transmission, and the quality of service (QoS) is affected by dynamic factors such as network load, failing to guarantee deterministic QoS. This means that system response latency and jitter are uncontrollable, unstable, and have no theoretical upper limit. Here, "response latency" is defined as the complete time interval (t2-t1) between the moment a task is sent from the user's local terminal (t1) and the moment the computing node completes processing (t2). With the future development of networks, more and more services require deterministic QoS guarantees, such as low-latency, high-reliability applications like remote surgery. Therefore, there is an urgent need to design edge computing network architectures that can provide deterministic QoS.
[0003] In this context, how to provide deterministic quality of service (QoS) guarantees in dynamic and unpredictable network environments has become a critical issue that future edge computing network systems urgently need to address. Deterministic networking technology precisely solves these problems. In a deterministic network, all nodes divide their time into equal-length time slots. Each pair of neighboring nodes establishes a time slot mapping relationship by pre-measuring the longest transmission delay. The definition and physical meaning of time slot mapping are as follows: If there is a link between nodes A and B (i.e., nodes A and B are neighboring nodes), then the mapping x→y means that a data packet sent by node A in time slot x will definitely be received by node B before the end of time slot y-1 (the longest transmission delay has been pre-measured). Regarding data packet forwarding, mapping x→y requires node B to transmit the aforementioned data packet to the next-hop node in time slot y. Similarly, mapping x→y requires that a data packet sent by node A in time slot x+1 can be re-forwarded by node B to the next-hop node in time slot y+1, and so on. Deterministic networks ensure deterministic end-to-end transmission by establishing hop-by-hop time slot mapping and forwarding relationships. This means that the delay and jitter between the source node sending a data packet and the destination node receiving it are controllable and have an upper bound.
[0004] Existing deterministic networking technologies only consider the transmission problems of wired networks and a single medium. Furthermore, during forwarding, nodes are required to divide time into equal-length time slots. However, edge computing network systems include wireless air interface components, wired transmission components, and computing node components. Task uploading and processing involves different media (wireless resource blocks, wired link bandwidth, and CPU resources). According to the requirements of current deterministic networks (due to the use of different transmission / processing media), nodes are neither suitable nor able to divide time into equal-length slots. Therefore, how to port deterministic networking technologies to edge computing network systems and ensure deterministic quality of service in edge computing network systems is the key problem addressed by this invention. Summary of the Invention
[0005] This invention provides an edge computing network architecture that provides deterministic quality of service, addressing the technical problem of porting deterministic network technology to edge computing network systems to provide deterministic quality of service.
[0006] The technical solution proposed in this invention is as follows:
[0007] This invention discloses an edge computing network architecture that provides deterministic quality of service, including a service provider, a user terminal, a base station, a router, and edge computing nodes. The service provider receives a task from the user terminal and feeds back the relevant configuration parameters for processing the task to the user terminal and relevant devices on the task transmission / processing path. The base station receives the task from the user terminal and forwards it to the router. The router receives the task sent by the base station and forwards it to the edge computing nodes.
[0008] The edge computing network architecture adopts a three-domain deployment approach for the data plane, including three independent domains: wireless domain, wired transmission domain, and computing domain. Among them, wireless domain data packets are transmitted using LTE / 5G standards, wired transmission domain data packets are transmitted using deterministic network technology, and the computing domain allocates computing resources with equal time periods.
[0009] The edge computing network architecture adopts an inter-domain traffic shaping mechanism, which includes a base station traffic shaping mechanism and an edge computing node traffic shaping mechanism. The base station traffic shaping mechanism forwards data packets based on specified time slots to achieve deterministic transmission from the wireless domain to the wired transmission domain. The computing node traffic shaping mechanism processes data packets based on specified time slots to achieve deterministic processing from the wired transmission domain to the computing domain.
[0010] Furthermore, the deterministic service ensures that after a user accesses the edge computing network, the system's response time and its jitter are controllable and have an upper bound.
[0011] Furthermore, the user terminal and base station belong to the wireless domain, the router belongs to the wired transmission domain, and the edge computing node belongs to the computing domain.
[0012] Furthermore, the service provider has resource allocation capabilities, configuring relevant decision variables for each task;
[0013] The user terminal has a time division function and a clock synchronization function. The time division function divides the time into time slots with a preset period length in the wireless domain, and the clock synchronization function keeps the start time of the user terminal and the associated base station time slot the same.
[0014] The base station has time division function, intra-domain clock synchronization function, inter-domain clock maintenance function, and base station traffic shaping function. The time division function divides time into time slots with a preset period length in the wireless domain. The intra-domain clock synchronization function keeps the start time of the time slot of the base station and the user terminals within its coverage area the same. The inter-domain clock maintenance function allows the base station to maintain a clock with a period of wired transmission domain time slots at the same time, keeping the clock frequency synchronized with that of adjacent routers.
[0015] The router has time partitioning function, intra-domain time synchronization function, inter-domain time synchronization function with time slots exceeding the time slot level, and deterministic forwarding function. The time partitioning function divides time into time slots of a preset period length in the wired transmission domain. The intra-domain time synchronization function keeps the router frequency synchronized with the adjacent router. The inter-domain time synchronization function with time slots exceeding the time slot level keeps the edge computing node and the router clock frequency synchronized with the time slot level.
[0016] The edge computing node has a time partitioning function, an inter-domain time synchronization function at the time slot level, and a time slot-based traffic shaping and task processing function. The time partitioning function divides time into time slots of a preset period length for the computing domain, and the inter-domain time synchronization function at the time slot level keeps the edge computing node and the router clock synchronized at the time slot level.
[0017] Furthermore, the time slot is defined as a time slot whose length is an integer multiple of the time slot, wherein the multiple is greater than or equal to 3.
[0018] Furthermore, the decision variables include: system parameters, transmission path, user terminal transmission time, base station traffic shaping parameters, and edge computing node traffic shaping parameters;
[0019] The system parameters include the time slot length and super-time slot length in the wireless domain, wired transmission domain, and computing domain; the transmission path is a set representing the task transmission / processing path, which includes the base stations, routers, and edge computing nodes traversed by the task; the user terminal transmission time is the specified time slot for the user terminal to send the task; the base station traffic shaping parameter is a positive integer indicating the base station traffic shaping operation; and the edge computing node traffic shaping parameter is a positive integer indicating the edge computing node traffic shaping operation.
[0020] Furthermore, the base station traffic shaping mechanism is as follows:
[0021] A cross-domain time index correspondence is established from the wireless domain to the wired transmission domain through a time slot mapping relationship, which is as follows:
[0022]
[0023] Where, Δ ran For the time slot length in the wireless domain, Δ dip N is the time slot length in the wired transmission domain. dip It is a positive integer greater than or equal to 3. Let x be the difference in the start time of the timeout slot between the two clocks on base station b, and let x be the period Δ. ran The time slot index, starting from zero, Φ b (x) represents the corresponding period Δ dip The time slot index;
[0024] According to the mapping relationship y = Φ b (x) indicates that the base station receiving a data packet in the wireless domain time slot x is equivalent to receiving it in the wired transmission domain time slot y, where y is a period of Δ dip The time slot index;
[0025] Specify data packet in Send to the router within the time slot. These are base station traffic shaping parameters.
[0026] Furthermore, the specific mechanism for shaping the computing node traffic is as follows:
[0027] A cross-domain time index correspondence is established from the wired transmission domain to the computing domain through a time slot mapping relationship, which is as follows:
[0028]
[0029] in, For router v i and edge computing node v j Link latency between For router v i and edge computing node v j The difference in start time of the upper time slot, Δ mec For the time slot length of the computation domain, Δ dip N is the time slot length in the wired transmission domain. mec x is a positive integer greater than or equal to 3, and x is a period of Δ. dip The time slot index starts from zero. The corresponding period is Δ mec The time slot index;
[0030] According to the mapping relationship Indicates router v i Data packets sent within time slot x of the wired transmission domain will be handled by edge computing node v. j Received within the computational domain y time slot, where y is a period of Δ. mec The time slot index;
[0031] Specify data packets in time slots The data is sent to the central processing unit for processing. This is a parameter for shaping the traffic of edge computing nodes.
[0032] The edge computing network architecture proposed in this invention provides deterministic quality of service, which transplants deterministic network technology to edge computing networks. It solves the problem that traditional deterministic network technology cannot be adapted to heterogeneous media scenarios. It can provide deterministic quality of service guarantees in dynamic and unpredictable environments, that is, the system "response time" and its jitter are controllable and have an upper bound. Attached Figure Description
[0033] The accompanying drawings, which form part of this application, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:
[0034] Figure 1 This is a model diagram of an edge computing network architecture that provides deterministic quality of service in an embodiment of the present invention;
[0035] Figure 2 This is a schematic diagram of a data plane transmission mechanism for an edge computing network architecture that provides deterministic quality of service in an embodiment of the present invention;
[0036] Figure 3 This is a schematic diagram illustrating the time partitioning, time synchronization, and frequency synchronization of an edge computing network architecture that provides deterministic quality of service in an embodiment of the present invention.
[0037] Figure 4 This is a schematic diagram illustrating the working steps of an edge computing network architecture that provides deterministic quality of service in an embodiment of the present invention;
[0038] Figure 5 This is a schematic diagram illustrating the process of end-to-end processing implemented by an edge computing network architecture that provides deterministic quality of service in an embodiment of the present invention. Detailed Implementation
[0039] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings in the embodiments of the present invention, and the embodiments will better describe the solution of the present invention.
[0040] Example 1
[0041] Reference Figure 4 This embodiment provides an edge computing network architecture that provides deterministic quality of service. The specific working steps of this network architecture are as follows:
[0042] S1: The user terminal submits a task processing request to the service provider;
[0043] Specifically, refer to Figure 5 A certain task will be performed in time slot a (time slot length is Δ). ran The service provider generates the relevant decision variables based on the current network resource utilization.
[0044] The decision variables include: system parameters, transmission path, user terminal transmission time, base station traffic shaping parameters, and edge computing node traffic shaping parameters. System parameters include the unit time slot length and time slot exceedance length in the wireless domain, wired transmission domain, and computing domain. The transmission path is a set representing the task transmission / processing path, including the base stations, routers, and edge computing nodes traversed by the task. The user terminal transmission time is the specified time slot for the user terminal to send the task. The base station traffic shaping parameter is a positive integer indicating the base station traffic shaping operation. The edge computing node traffic shaping parameter is a positive integer indicating the edge computing node traffic shaping operation.
[0045] S2: The service provider sends the relevant configuration parameters to the user terminal and all devices on the task transmission / processing path;
[0046] Specifically, refer to Figure 5 The service provider needs to send the relevant configuration parameters to the user terminal, corresponding base station, router, and edge computing node. The system parameters specify that the time slot lengths for the wireless domain, wired transmission domain, and computing domain correspond to Δ. ran Δ dip and Δ mec The time slot length is Δ hc ,Δ hc The specific value of Δ satisfies the condition: hc =N ran Δ ran =N dip Δ dip =N mec Δ mec , where N ran N dip N mec A positive integer greater than or equal to 3; devices in the transmission / processing path include base station v1, router v2, router v3, and edge computing node v4; for user terminals, service providers need to provide access permissions and transmission time slot indexes. For base stations and edge computing nodes, service providers need to provide base station / edge computing node traffic shaping parameters. and
[0047] S3: The user terminal sends the data packet corresponding to the task to the designated base station;
[0048] Specifically, refer to Figure 5 User terminals in time slots Within this timeframe, the data packet corresponding to the task is sent to the designated base station. During wireless transmission, the task occupies one radio resource block, and the usage time is one time slot length Δ. ran (To be specified in advance by the service provider).
[0049] S4: The base station completes the task reception and forwards it to the router within the specified time slot;
[0050] Specifically, refer to Figure 5 The base station will be in time slot a+2 (length Δ) ran Before the task ends, complete the wireless reception work.
[0051] The specified task transmission time slot is calculated using the base station traffic shaping mechanism:
[0052] Input the service provider's configured parameters into the following time slot mapping relationship:
[0053]
[0054] Where, Δ ran Δ is the unit time slot length in the wireless domain. dip N represents the unit time slot length in the wired transmission domain. dip It is a positive integer greater than or equal to 3. Let x be the difference in the start time of the timeout slot between the two clocks on base station b, and let x be the period Δ. ran The time slot index, starting from zero, Φ b (x) represents the corresponding period Δ dip The time slot index;
[0055] Calculated The base station believes that the wireless reception for the corresponding task can be performed in time slot b+3 (length Δ). dip Before the end, complete the reception process. The base station will then process the received data. Before time slot b+4 ends, forward the task to router v2.
[0056] S5: The router receives the task and forwards it to the edge computing node;
[0057] Specifically, refer to Figure 5Routers v2 and v3 use existing deterministic networking technologies to forward tasks to edge computing node v4 before time slot b+7 ends.
[0058] S6: The edge computing node receives the task and completes the processing within the specified time slot;
[0059] Specifically, refer to Figure 5 The specified task processing time slot is calculated through the edge computing node traffic shaping mechanism, and the service provider's configured parameters are input into the following time slot mapping relationship:
[0060]
[0061] in, For router v i and compute node v j Link latency between For router v i and compute node v j The difference in start time of the upper time slot, Δ mec For the time slot length of the computation domain, Δ dip N is the time slot length in the wired transmission domain. mec x is a positive integer greater than or equal to 3, and x is a period of Δ. dip The time slot index starts from zero. The corresponding period is Δ mec The time slot index;
[0062] Calculated This mapping relationship indicates that compute node v4 can be deployed in time slot c+5 (length Δ). mec Before the end, complete the task reception process. Compute node v4 receives the data based on the parameters. Assign the task in time slot c+7 (length Δ) mec Complete the processing work before the end.
[0063] Thus, the network architecture used in this embodiment provides end-to-end deterministic quality of service, meaning that the system's "response time" and its jitter are controllable and have an upper bound, as shown in the reference. Figure 5 You can see the system's maximum / minimum response time.
[0064] This embodiment adopts an edge computing network architecture that provides deterministic quality of service, transplanting deterministic network technology into the edge computing network. This solves the problem that traditional deterministic network technology cannot be adapted to heterogeneous media scenarios. It can provide deterministic quality of service guarantees in dynamic and unpredictable environments, that is, the system's "response time" and its jitter are controllable and have an upper bound.
[0065] In addition, embodiments of the present invention also provide a computer-readable storage medium, wherein the computer-readable storage medium may store a program that, when executed, includes any edge computing network architecture providing deterministic quality of service as described in the above method embodiments.
[0066] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.
[0067] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.
[0068] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage device, which may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.
[0069] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above description is only a specific embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made on the basis of the technical solution of this application should be included within the scope of protection of this application.
Claims
1. An edge computing network architecture providing deterministic quality of service, characterized in that, It includes service providers, user terminals, base stations, routers, and edge computing nodes. The service provider receives the task from the user terminal and feeds back the relevant configuration parameters for processing the task to the user terminal and the relevant devices on the task transmission / processing path. The base station receives the task from the user terminal and forwards it to the router. The router receives the task sent by the base station and forwards it to the edge computing node. The edge computing network architecture adopts a three-domain deployment approach for the data plane, including three independent domains: wireless domain, wired transmission domain, and computing domain. Among them, wireless domain data packets are transmitted using LTE / 5G standards, wired transmission domain data packets are transmitted using deterministic network technology, and the computing domain allocates computing resources with equal time periods. The edge computing network architecture adopts an inter-domain traffic shaping mechanism, which includes a base station traffic shaping mechanism and an edge computing node traffic shaping mechanism. The base station traffic shaping mechanism forwards data packets based on specified time slots to achieve deterministic transmission from the wireless domain to the wired transmission domain. The base station traffic shaping mechanism is as follows: A cross-domain time index correspondence is established from the wireless domain to the wired transmission domain through a time slot mapping relationship, which is as follows: Among them, △ ran For the time slot length in the wireless domain, Δ dip N is the time slot length in the wired transmission domain. dip It is a positive integer greater than or equal to 3. Let x be the difference in the start time of the timeout slot between the two clocks on base station b, and let x be the period of Δ. ran The time slot index, starting from zero, Φ b (x) represents the corresponding period Δ dip The time slot index; According to the mapping relationship y = Φ b (x) indicates that the base station receiving a data packet in the wireless domain x time slot is equivalent to receiving it in the wired transmission domain y time slot, where y is the time slot index with a period of △dip; Specify data packet in Send to the router within the time slot. For base station traffic shaping parameters; The compute node traffic shaping mechanism processes data packets based on specified time slots, enabling deterministic processing from the wired transmission domain to the compute domain; The specific computing node traffic shaping mechanism is as follows: A cross-domain time index correspondence is established from the wired transmission domain to the computing domain through a time slot mapping relationship, which is as follows: in, For router v i and edge computing node v j Link latency between For router v i and edge computing node v j The difference in start time of the upper time slot, △ mec For the time slot length of the computation domain, Δ dip N is the time slot length in the wired transmission domain. mec x is a positive integer greater than or equal to 3, and x is a period of Δ. dip The time slot index starts from zero. This is the slot index corresponding to a period of △mec; According to the mapping relationship Indicates router v i Data packets sent within time slot x of the wired transmission domain will be handled by edge computing node v. j Received within the computational domain y time slot, where y is a period of Δ. mec The time slot index; Specify data packets in time slots The data is sent to the central processing unit for processing. This is a parameter for shaping the traffic of edge computing nodes.
2. The edge computing network architecture providing deterministic quality of service according to claim 1, characterized in that, The deterministic service ensures that after a user accesses the edge computing network, the system's response time and jitter are controllable and have an upper bound.
3. The edge computing network architecture providing deterministic quality of service according to claim 1, characterized in that, The user terminal and base station belong to the wireless domain, the router belongs to the wired transmission domain, and the edge computing node belongs to the computing domain.
4. The edge computing network architecture providing deterministic quality of service according to claim 1, characterized in that, The service provider has resource allocation capabilities and can configure relevant decision variables for each task; The user terminal has time division and clock synchronization functions; The base station has time division function, intra-domain clock synchronization function, inter-domain clock maintenance function, and base station traffic shaping function. The router has time division function, intra-domain time synchronization function, inter-domain time synchronization function with time slots, and deterministic forwarding function; The edge computing node has time partitioning function, inter-domain time synchronization function at the time slot level, and time slot-based traffic shaping and task processing function.
5. An edge computing network architecture providing deterministic quality of service according to claim 4, characterized in that, The time slot is defined as a time slot whose length is an integer multiple of the time slot, where the multiple is greater than or equal to 3.
6. An edge computing network architecture providing deterministic quality of service according to claim 4, characterized in that, The decision variables include: system parameters, transmission path, user terminal transmission time, base station traffic shaping parameters, and edge computing node traffic shaping parameters. The system parameters include the time slot length and super-time slot length in the wireless domain, wired transmission domain, and computing domain; the transmission path is a set representing the task transmission / processing path, which includes the base stations, routers, and edge computing nodes traversed by the task; the user terminal transmission time is the specified time slot for the user terminal to send the task; the base station traffic shaping parameter is a positive integer indicating the base station traffic shaping operation; and the edge computing node traffic shaping parameter is a positive integer indicating the edge computing node traffic shaping operation.
7. A computer-readable storage medium, characterized in that, The storage medium contains an edge computing network architecture that provides deterministic quality of service as described in any one of claims 1 to 6.
Citation Information
Patent Citations
Enhanced NEF function, MEC and 5G integration
CN111684824A
FIWI network media access control system and method based on edge computing
CN112261667A