WO2022142631A1 - 任务处理方法、装置及系统、计算机存储介质 - Google Patents

任务处理方法、装置及系统、计算机存储介质 Download PDF

Info

Publication number
WO2022142631A1
WO2022142631A1 PCT/CN2021/125852 CN2021125852W WO2022142631A1 WO 2022142631 A1 WO2022142631 A1 WO 2022142631A1 CN 2021125852 W CN2021125852 W CN 2021125852W WO 2022142631 A1 WO2022142631 A1 WO 2022142631A1
Authority
WO
WIPO (PCT)
Prior art keywords
mec
mec device
neighbor
target
task request
Prior art date
Application number
PCT/CN2021/125852
Other languages
English (en)
French (fr)
Inventor
曹俊
王楠楠
Original Assignee
华为云计算技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为云计算技术有限公司 filed Critical 华为云计算技术有限公司
Publication of WO2022142631A1 publication Critical patent/WO2022142631A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/10—Protocols in which an application is distributed across nodes in the network
    • H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
    • H04L67/1004—Server selection for load balancing
    • H04L67/1012—Server selection for load balancing based on compliance of requirements or conditions with available server resources
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/50—Queue scheduling
    • H04L47/62—Queue scheduling characterised by scheduling criteria
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/10—Protocols in which an application is distributed across nodes in the network
    • H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
    • H04L67/1004—Server selection for load balancing
    • H04L67/1008—Server selection for load balancing based on parameters of servers, e.g. available memory or workload

Definitions

  • the present application relates to the field of communication technologies, and in particular, to a task processing method, device and system, and a computer storage medium.
  • Each data center is managed by distributed cloud technology, forming a hierarchical network system of edge cloud and central cloud.
  • the application instances (hereinafter referred to as instances) where the data center provides application services (application servers) for user equipment (UE) are deployed on each edge cloud, so that the instances of application services are deployed closer to the UE, reducing service costs. The delay improves the service experience.
  • the process for the UE to access application services includes: the UE sends a service routing request to an application controller that manages a multi-access edge computing (MEC) device, where the service routing request includes the application service requested by the UE to access The application service identification (ID).
  • the application controller obtains the MEC device on which the instance of the application service is deployed according to the application service ID, and then obtains the MEC device closest to the UE from the MEC devices on which the instance of the application service is deployed according to the geographic location information of the UE, and stores the The communication address information of the MEC device is sent to the UE.
  • the UE accesses the MEC device based on the communication address information of the MEC device, and then sends a task request to the MEC device to access the application service.
  • the present application provides a task processing method, device and system, and a computer storage medium, which can solve the problem of poor reliability of application services provided by current MEC devices.
  • a task processing method includes: the first MEC device receives a task request from the UE, where the task request includes an application service identifier of a target application service requested by the UE to access and SLA information corresponding to the task request, where the SLA information is used to determine the response to the task request. Satisfied SLA conditions.
  • the first MEC device determines a second MEC device among the neighbor MEC devices of the first MEC device in which the target is deployed A second instance of the service is applied and the second instance satisfies the SLA conditions. The first MEC device forwards the task request to the second MEC device.
  • the MEC device after receiving the task request from the UE, determines that it does not meet the SLA conditions that should be met in response to the task request, and schedules the task request to the neighboring MEC devices that meet the SLA conditions.
  • the neighbor MEC device processes the task request. Through reasonable scheduling of the task request, the response to the task request can meet the SLA conditions, and the reliability and availability of the task processing are improved.
  • the SLA information corresponding to the task request indicates the maximum allowable response delay of the task request
  • the SLA conditions satisfied by the response task request include: the total delay of responding to the task request is not greater than the maximum allowable response delay
  • selecting the MEC device processing the task request based on the maximum allowable response delay of the task request can ensure the access delay of the UE and improve the reliability of the application service provided by the MEC device.
  • the first MEC device may also acquire application information of the neighbor MEC device, where the application information includes the historical task processing time of the instance of the deployed application service.
  • the implementation process of the first MEC device determining the second MEC device in the neighbor MEC device of the first MEC device includes: the first MEC device determines the target neighbor MEC device in the neighbor MEC device according to the application information of the neighbor MEC device. Among them, the target neighbor MEC device satisfies: an instance of the target application service is deployed, and the historical task processing time of the deployed instance of the target application service is less than the remaining allowable response delay, and the remaining allowable response delay is obtained based on the maximum allowable response delay. The first MEC device determines the second MEC device among the target neighbor MEC devices.
  • the routing object of the task request is determined, and the routing object has a high probability of meeting the SLA conditions for responding to the task request, thereby improving the reliability of task processing.
  • the application information further includes the load of the deployed application service instance
  • the first MEC device determines the implementation process of the second MEC device in the target neighbor MEC device, including: in response to the number of the target neighbor MEC device being greater than 1, the first MEC device.
  • An MEC device is based on the historical task processing time of the instances of the target application service deployed in the multiple target neighbor MEC devices, the load of the instances of the target application service deployed in the multiple target neighbor MEC devices, and/or the relationship between the multiple target neighbor MEC devices and the The distance between the first MEC devices, and the second MEC device is determined among the multiple target neighbor MEC devices.
  • factors such as the status, load, and task processing time of MEC devices are comprehensively considered to perform routing selection for multiple traffic distribution, and the UE's access traffic is distributed to other appropriate MEC devices, so as to improve the reliability and efficiency of device-side collaborative task processing. availability.
  • the first MEC device determines the implementation process of the target neighbor MEC device in the neighbor MEC device according to the application information of the neighbor MEC device, including: the first MEC device according to the application information of the neighbor MEC device and the neighbor MEC device and the first MEC device.
  • the network response delay between MEC devices determines the target neighbor MEC device among the neighbor MEC devices.
  • the target neighbor MEC device satisfies: an instance of the target application service is deployed, and the sum of the network response delay with the first MEC device and the historical task processing time of the deployed instance of the target application service is less than the remaining allowable response delay .
  • the implementation process for the first MEC device to determine the second MEC device among the target neighbor MEC devices may include: in response to the number of the target neighbor MEC devices being greater than 1, the first MEC device compares the multiple target neighbor MEC devices with the The target neighbor MEC device with the smallest sum of the network response delay between the first MEC devices and the historical task processing time of the deployed instance of the target application service is determined as the second MEC device.
  • the routing object of the task request is determined, and the probability that the routing object can meet the SLA conditions for responding to the task request.
  • the way of considering the processing time of the task request by the MEC device is higher, thereby improving the reliability of the task processing.
  • scheduling task requests by taking into account the processing time of MEC devices for task requests and the network response delay between MEC devices can provide deterministic task response delays and improve user performance. experience.
  • the task request further includes request file information
  • the request file information includes a file identifier of the file to be requested
  • the application information further includes file information
  • the file information includes the file identifier of the local file and the file size of the local file.
  • the first MEC device determines the implementation process of the target neighbor MEC device in the neighbor MEC device according to the application information of the neighbor MEC device, including: the first MEC device according to the application information of the neighbor MEC device, the request file information and the neighbor MEC device and the first The network bandwidth between MEC devices, and the target neighbor MEC device is determined in the neighbor MEC device.
  • the target neighbor MEC device satisfies: an instance of the target application service is deployed, the local file includes the file to be requested, and the file transfer time for transmitting the to-be-requested file to the first MEC device and the historical task processing time of the deployed instance of the target application service The sum is less than the remaining allowable response delay, and the file transmission time is determined based on the file size of the file to be requested and the network bandwidth between the neighbor MEC device and the first MEC device.
  • the implementation process for the first MEC device to determine the second MEC device in the target neighbor MEC devices may include: in response to the number of the target neighbor MEC devices being greater than 1, the first MEC device transfers the multiple target neighbor MEC devices to the target neighbor MEC devices.
  • the target neighbor MEC device with the smallest sum of the file transmission time of the first MEC device transmitting the to-be-requested file and the historical task processing time of the deployed instance of the target application service is determined as the second MEC device.
  • the routing object of the task request is determined based on the processing time of the MEC device for the task request and the file transmission time for transferring the file to be requested between the MEC devices, and the probability that the routing object can satisfy the SLA condition of the task request is compared with The method of only considering the processing time of the task request by the MEC device is higher, thereby improving the reliability of the task processing.
  • the routing object of the task request may also be determined by considering the time when the task request is transmitted between the MEC devices.
  • the first MEC device may also broadcast and send application information of the first MEC device.
  • the first MEC device determines the implementation process of the second MEC device in the neighboring MEC devices of the first MEC device, including: In response to the first instance not meeting the SLA condition and the forwarded times of the task request received by the first MEC device being less than the forwarding times threshold, the first MEC device determines a second MEC device among neighbor MEC devices.
  • the first MEC device may also receive an MEC device list sent by the application controller, where the MEC device list includes communication address information and geographic location information of multiple MEC devices.
  • the first MEC device determines, among the multiple MEC devices, an MEC device whose distance from the first MEC device is smaller than the distance threshold as a neighbor MEC device of the first MEC device.
  • the first MEC device may further receive a response to the task request sent by the second MEC device.
  • the first MEC device forwards the response to the task request to the target device.
  • the target device is a device that sends a task request to the first MEC device.
  • the target device may be a UE, or may also be other MEC devices.
  • the routing path of the response to the task request sent by the UE is opposite to the routing path of the task request.
  • the UE sends the task request to the accessed MEC device, and finally the MEC device accessed by the UE also sends the task request to the UE.
  • the entire process of the MEC device processing the task request is unaware of the UE, so there is no need to change the way the UE accesses the application service, and the compatibility with the UE is good.
  • a possible implementation manner of the first MEC device receiving the task request from the UE includes: in response to the UE not being in the service area corresponding to the first MEC device, the first MEC device receives the task request forwarded by the third MEC device. Alternatively, in response to the UE being in the service area corresponding to the first MEC device, the first MEC device receives the task request sent by the UE or the task request forwarded by the fourth MEC device.
  • the service area corresponding to the MEC device that finally processes the task request from the UE may or may not cover the location of the UE, that is, the MEC device itself does not meet the SLA conditions that should be met to respond to the task request.
  • the MEC device that processes the task request can be flexibly selected among the neighboring MEC devices, so that the flexibility of the task processing is high.
  • the first MEC device processes the task request through the first instance.
  • a task processing device in a second aspect, includes a plurality of functional modules, and the plurality of functional modules interact to implement the methods in the first aspect and the various embodiments thereof.
  • the multiple functional modules may be implemented based on software, hardware, or a combination of software and hardware, and the multiple functional modules may be arbitrarily combined or divided based on specific implementations.
  • a MEC device including: a processor and a memory;
  • the memory for storing a computer program, the computer program including program instructions
  • the processor is configured to invoke the computer program to implement the methods in the first aspect and various embodiments thereof.
  • a task processing system includes: a plurality of multi-access edge computing MEC devices, the MEC devices include the apparatus according to the second aspect or the MEC according to the third aspect equipment.
  • a computer storage medium where instructions are stored on the computer storage medium, and when the instructions are executed by a processor, the methods in the foregoing first aspect and its various embodiments are implemented.
  • a chip in a sixth aspect, includes a programmable logic circuit and/or program instructions, and when the chip runs, the methods in the first aspect and the various embodiments thereof are implemented.
  • the route selection of the traffic distribution is carried out multiple times, and the access traffic of the UE is distributed to the appropriate Other MEC devices to improve the reliability and availability of end-to-end collaborative task processing.
  • the traffic is distributed to other available MEC devices through multiple traffic distribution to ensure the availability and continuity of services.
  • route selection takes into account factors such as the maximum allowable response latency, network response latency between MEC devices, task processing time, and file transfer time to provide deterministic task response latency and improve user experience. .
  • factors such as the maximum allowable response latency, network response latency between MEC devices, task processing time, and file transfer time to provide deterministic task response latency and improve user experience.
  • multiple traffic distributions are used to meet application latency requirements and ensure application user experience. The method is realized by mutual perception between MEC devices, but the UE has no perception, so there is no need to change the way the UE accesses the application service, and the compatibility with the UE is good.
  • the MEC device when the MEC device itself does not meet the SLA conditions that should be satisfied in responding to the task request, it is not limited by the geographical location of the UE, and can flexibly select the MEC device that processes the task request among the neighboring MEC devices, making the task processing more flexible. high.
  • FIG. 1 is a schematic diagram of an infrastructure of a communication network provided by an embodiment of the present application.
  • FIG. 2 is a schematic flowchart of a task processing method provided by an embodiment of the present application.
  • FIG. 3 is a schematic structural diagram of a task processing apparatus provided by an embodiment of the present application.
  • FIG. 4 is a schematic structural diagram of another task processing apparatus provided by an embodiment of the present application.
  • FIG. 5 is a schematic structural diagram of another task processing apparatus provided by an embodiment of the present application.
  • FIG. 6 is a schematic structural diagram of yet another task processing apparatus provided by an embodiment of the present application.
  • FIG. 7 is a block diagram of an MEC device provided by an embodiment of the present application.
  • FIG. 1 is a schematic diagram of an infrastructure of a communication network provided by an embodiment of the present application.
  • the communication network includes: a core network (core network, CN) device 101, base stations 102A-102B (collectively referred to as base stations 102) and MEC devices 103A-103B (collectively referred to as MEC devices 103).
  • each MEC device 103 is connected to the corresponding base station 102, for example, the MEC device 103A is connected to the base station 102A, and the MEC device 103B is connected to the base station 102B.
  • Each base station 102 is connected to the core network device 101 .
  • the number of each device in the figure is only used for exemplary description, and not as a limitation on the communication network involved in the embodiments of the present application.
  • the MEC device 103 is an infrastructure platform that implements edge cloud deployment.
  • the MEC device 103 may be composed of one or more servers, or composed of one or more servers and other network devices (such as gateways, switches and other devices), or a cloud computing platform, the network resources formed by the network architecture Used to provide a service platform for instances.
  • Each MEC device 103 has a corresponding service area, and the service area may be a cell corresponding to the base station 102 connected to the MEC device 103 .
  • the service area corresponding to the MEC device 103A includes the cell corresponding to the base station 102A
  • the service area corresponding to the MEC device 103B includes the cell corresponding to the base station 102B.
  • At least one instance is deployed in each MEC device 103 , and the instance is used to provide application services for UEs in a service area corresponding to the MEC device 103 .
  • the application service refers to: a service provided for the UE through various types of applications, such as a vehicle network application service operated by a certain developer.
  • An instance refers to a specific application of an application service deployed on different MEC devices. Therefore, an application service can correspond to multiple instances. For example, in the embodiment of the present application, a first instance is deployed in the first MEC device, and a second instance is deployed in the second MEC device, wherein the first instance and the second instance are instances of the same application service (target application service) .
  • the communication network as shown in Figure 1 can be extended in many variants.
  • the core network can be divided into a control plane (CP) and a user plane (UP) to form a control and user plane separation (CUPS) communication network.
  • the core network device 101 only includes the control plane function.
  • the user plane of the core network is implemented by each gateway (not shown in the figure), that is, each gateway belongs to the core network side device and performs the user plane function of the core network.
  • the data connection between the gateway and the base station is uniformly managed and controlled by the core network control plane.
  • the base station 102 is a new radio access base station (new radio gNodeB, NR gNB) or other radio access device, which is not limited in this embodiment of the present application.
  • new radio access base station new radio gNodeB, NR gNB
  • other radio access device which is not limited in this embodiment of the present application.
  • the instances of the application service are deployed on multiple MEC devices near the base station in a distributed manner, and are used to provide the application service to UEs in the cells of the connected base station.
  • instances of ultra-reliable and low-latency communication (URLLC) applications are deployed on multiple MEC devices close to base stations.
  • the URLLC instance deployed on each MEC device provides URLLC services for the UE respectively.
  • the application service ID of each application service may be described in the MEC device.
  • the application service ID is used to uniquely identify the corresponding application service to distinguish different application services.
  • the application service ID may be in the form of a domain name (domain name), or the MEC device may uniformly allocate an application service ID for each application service, for example, the application service ID may be an Internet Protocol (Internet Protocol, IP) address.
  • IP Internet Protocol
  • the application service ID may also be in other forms, which are not limited in this embodiment of the present application.
  • Instances of each application service also have an instance ID.
  • the instance ID is used to uniquely identify the corresponding instance to distinguish different instances of the same application service. That is, different instances of the same application service have the same application service ID but different instance IDs.
  • the instance ID may be in the form of the IP address of the instance, or may be in other forms, which are not limited in this embodiment of the present application. It should be noted that, when the application service ID and the instance ID are both in the form of IP addresses, the sender and/or the receiver can distinguish the two through different fields in the message.
  • the sending end can parse out the IP address of the application service and the IP address of the instance from the corresponding fields.
  • the communication network may also include an application controller 104 .
  • the application controller 104 is used to manage each MEC device in the communication network, and then manage and control all applications in the communication network.
  • the application controller 104 is a network device on the core network side, and the application controller 104 is connected to the core network device 101 for acquiring the geographic location information of the UE through the core network device 101 .
  • the application controller 104 is connected to each MEC device in the communication network, and is used for sending control information to each MEC device 103 .
  • the application controller 104 is also connected to network devices (such as base stations) on each base station side, and is used to make control decisions and the like of the application instance according to the requirements of the user side, for example, to decide which MEC device 102 the UE accesses.
  • network devices such as base stations
  • the application controller 104 stores the information of each MEC device 103 in the communication network, including the connection relationship between the MEC device 103 and the base station 102, the communication address information, geographic location information of the MEC device 103, and instances of deployed application services.
  • the communication address information may be an IP address.
  • the geographic location information can be represented by a city name and/or latitude and longitude.
  • the application controller 104 is also used to provide the MEC device 103 with the communication address information and geographic location information of other MEC devices.
  • the application controller 104 may also provide the MEC device 103 with information such as application service IDs of application services deployed in other MEC devices.
  • the UE when the UE wants to access the target application service, the UE sends a service routing request to the application controller 104, where the service routing request includes the application service ID of the target application service.
  • the application controller 104 acquires the geographic location information of the UE from the core network device 101 .
  • the core network device 101 may determine the geographic location of the UE according to the base station to which the UE is connected.
  • the application controller 104 can obtain the MEC device whose service area covers the location of the UE from the MEC device 103 in which the instance of the target application service is deployed according to the geographic location information of the UE, and the MEC device can be the location where the service area covers the UE and is closest to the UE.
  • the MEC device deployed with the instance of the target application service may also be any MEC device deployed with the instance of the target application service whose service area covers the location of the UE.
  • the application controller 104 sends the acquired communication address information of the MEC device to the UE.
  • the UE accesses the MEC device based on the communication address information of the MEC device, and then sends a task request to the MEC device.
  • both the MEC device 103A and the MEC device 103B are deployed with instances of the target application service.
  • the application controller 104 It will control and decide that the UE accesses the MEC device 103A.
  • the MEC device when it receives a task request from the UE, it first determines whether it satisfies the service level agreement (service level agreement, SLA) condition that should be satisfied in response to the task request. After it is determined that it satisfies the SLA condition, the task request is processed; after it is determined that it does not meet the SLA condition, the task request is scheduled to the neighbor MEC device that satisfies the SLA condition, and the neighbor MEC device that satisfies the SLA condition processes the task. ask.
  • SLA service level agreement
  • FIG. 2 is a schematic flowchart of a task processing method provided by an embodiment of the present application.
  • the method can be applied to the communication network as shown in FIG. 1 .
  • the method includes:
  • Step 201 The first MEC device receives the MEC device list sent by the application controller.
  • the MEC device list includes communication address information and geographic location information of multiple MEC devices.
  • the communication address information may be an IP address.
  • the geographic location information may include a city where the MEC device is located and/or the latitude and longitude of the location.
  • the MEC device list may further include application service IDs of application services deployed in the plurality of MEC devices.
  • the multiple MEC devices in the MEC device list may include all MEC devices managed by the application controller, or may only include MEC devices that are deployed with at least one instance of the same application service as the first MEC device.
  • the first instance of the target application service is deployed in the first MEC device.
  • the target application service can be any application service.
  • Step 202 The first MEC device determines, among the multiple MEC devices, a MEC device whose distance from the first MEC device is less than a distance threshold as a neighbor MEC device of the first MEC device.
  • the neighbor MEC device of the first MEC device is the MEC device whose distance from the first MEC device is smaller than the distance threshold.
  • the first MEC device can calculate the distance from each MEC device to the first MEC device according to the geographic location information of the first MEC device and the geographic location information of each MEC device in the MEC device list, and After the neighbor MEC device is determined, the communication address information and geographic location information of the neighbor MEC device are stored.
  • the first MEC device determines, among the multiple MEC devices, the MEC device whose distance from the first MEC device is less than the distance threshold and has at least one instance of the same application service deployed as the neighbor MEC device of the first MEC device. . Since the first MEC device acquires the neighbor MEC device for the purpose of scheduling the task request from the UE, and the task request can only be scheduled to the MEC device deployed with the instance of the application service that the task request requests to access, only the MEC device with the An MEC device deployed with at least one instance of the same application service is only likely to receive a task request scheduled by the first MEC device. Therefore, the neighbor MEC device of the first MEC device is set to meet the requirements of at least one MEC device deployed with the first MEC device.
  • the conditions of the instance of the same application service can reduce the content resources occupied by the first MEC device to store the information of the neighbor MEC device while ensuring the scheduling validity and comprehensiveness of the task request by the first MEC device, and can also reduce the neighbor MEC device.
  • the number of devices is used to improve the scheduling efficiency of subsequent first MEC devices for task requests.
  • Step 203 The first MEC device acquires the application information of the neighbor MEC devices of the first MEC device.
  • the application information of the neighbor MEC device of the first MEC device includes the historical task processing time of the instance of the application service deployed in the neighbor MEC device.
  • the historical task processing time may include the maximum processing time of the task request and/or the average processing time of the task request in the last N minutes, where N is a positive number.
  • the historical task processing time may also include the minimum processing time of the task request within the last N minutes.
  • the application information of the neighbor MEC device of the first MEC device also includes the load of the instance of the application service deployed in the neighbor MEC device, that is, the central processing unit corresponding to each instance of the application service deployed in the neighbor MEC device. (central processing unit, CPU) usage.
  • the CPU resources for running an instance of an application service in the MEC device are usually pre-divided.
  • the first MEC device may also collect statistics on application information of the first MEC device.
  • the first MEC device may establish a regional routing table after determining the neighbor MEC device of the first MEC device and acquiring the application information of the neighbor MEC device.
  • the routing table established by the first MEC may include the following contents:
  • IP address xx.xx.xx.xx
  • IP address xx.xx.xx.xx
  • the APP list includes information of instances of various types of application services deployed in the MEC device.
  • the first MEC device may send a network probe request to the neighbor MEC device to count the network response delay between the neighbor MEC device and the first MEC device.
  • the network response delay between the neighbor MEC device and the first MEC device may refer to: the total duration from the first MEC device sending the network probe request to receiving the network probe response replied by the neighbor MEC device based on the network probe request.
  • the network response delay may include the maximum network response delay, the average network response delay and/or the minimum network response delay in the last N minutes.
  • the first MEC device may periodically count the network response delay between the neighbor MEC device and the first MEC device, so as to update the network response delay in real time.
  • the routing table established by the first MEC may include the following contents:
  • IP address xx.xx.xx.xx
  • IP address xx.xx.xx.xx
  • the application information of the neighbor MEC device of the first MEC device further includes file information of the neighbor MEC device, where the file information includes the file identifier of the local file and the file size of the local file.
  • the file identifier can be a file name.
  • the file information can also include the directory of the local file.
  • the first MEC device may send a network probe request to the neighbor MEC device to count the network bandwidth between the neighbor MEC device and the first MEC device.
  • the network bandwidth may include a maximum network bandwidth, an average network bandwidth, and/or a minimum network bandwidth in the last N minutes.
  • the first MEC device may periodically count the network bandwidth between the neighbor MEC device and the first MEC device, so as to update the network bandwidth in real time.
  • the routing table established by the first MEC may include the following contents:
  • IP address xx.xx.xx.xx
  • IP address xx.xx.xx.xx
  • the implementation process of step 203 includes: the first MEC device sends an application information acquisition request to a neighbor MEC device; the first MEC device receives the neighbor MEC device's reply based on the application information acquisition request. Application information for MEC devices.
  • the first MEC device periodically sends an application information acquisition request to the neighbor MEC device, so as to update the application information of the neighbor MEC device in real time.
  • the first MEC device stores the latest application information of neighbor MEC devices.
  • the implementation process of step 203 includes: the first MEC device receives the application information actively broadcast and sent by the neighbor MEC device.
  • the first MEC device broadcasts the application information of the first MEC device.
  • the first MEC device broadcasts and sends the application information of the first MEC device, which means sending a packet containing the application information of the first MEC device and the packet is a group addressed (group addressed) packet
  • the destination address of the packet can be a multicast address or a broadcast address.
  • the destination address of the packet includes the IP addresses of multiple neighboring MEC devices.
  • each MEC device in the communication network periodically broadcasts and sends its own application information, or broadcasts and sends the updated application information every time its own application information is updated.
  • Each MEC device stores the latest application information of neighboring MEC devices.
  • Step 204 The first MEC device receives the task request from the UE.
  • the task request includes the application service identifier of the target application service that the UE requests to access and the SLA information corresponding to the task request, where the SLA information is used to determine the SLA condition satisfied in response to the task request.
  • the SLA information corresponding to the task request indicates the maximum allowable response delay of the task request
  • the SLA conditions satisfied by the response to the task request include: the total delay of responding to the task request is not greater than the maximum allowable response delay.
  • the maximum allowable response delay of the task request refers to the maximum time allowed for the entire process from the time when the UE sends the task request to the completion of the processing of the task request.
  • the task request being processed may be that the UE receives a response to the task request.
  • the SLA information corresponding to the task request may include the maximum allowable response delay of the task request and the sending time at which the UE sends the task request.
  • the task request further includes related information of the to-be-processed service, for example, may include request file information, where the request file information includes a file identifier of the to-be-requested file.
  • the file to be requested refers to the file that the UE requests to download.
  • the task request may also include a sequence number for identifying the task request, etc., which is not limited in this embodiment of the present application.
  • the first MEC device receives the task request from the UE, which may be the first MEC device receiving the task request directly sent by the UE; or the first MEC device receiving the task request from the UE forwarded by other MEC devices.
  • the first MEC device may receive the task request forwarded by the third MEC device, where the first MEC device is a neighbor MEC device of the third MEC device.
  • the first MEC device may receive a task request sent by the UE or a task request forwarded by the fourth MEC device, where the first MEC device is a neighbor MEC device of the fourth MEC device .
  • the service area corresponding to the MEC device that finally processes the task request from the UE may or may not cover the location of the UE.
  • SLA conditions it is not limited by the geographic location of the UE, and the MEC device that processes the task request can be flexibly selected among the neighboring MEC devices, so that the flexibility of task processing is high.
  • the first MEC device may also receive the number of forwardings sent together with the task request, where the number of forwardings is the number of times the task request has been forwarded.
  • the forwarding times received by the first MEC device and sent together with the task request is 0, indicating that the task request is directly sent by the UE to the first MEC device.
  • the forwarding times received by the first MEC device and sent together with the task request is 1, indicating that the task request is forwarded to the first MEC device via one MEC device.
  • Step 205 The first MEC device determines whether the first instance of the target application service deployed in the first MEC device satisfies the SLA condition; in response to the first instance of the target application service deployed in the first MEC device not meeting the SLA condition, execute Step 206; or, in response to the first instance of the target application service deployed in the first MEC device meeting the SLA condition, perform step 208.
  • the task request includes the maximum allowable response delay of the task request and the sending time when the UE sends the task request.
  • the implementation process of step 205 includes: the first MEC device calculates the remaining allowable response delay of the task request.
  • the remaining allowable response delay maximum allowable response delay-(current time-the sending time when the UE sends the task request).
  • the current time is the time at which the remaining allowable response delay is calculated, and may be, for example, the time at which it is judged whether the first instance satisfies the SLA condition.
  • the first MEC device determines that the first instance satisfies the SLA condition; or, in response to the historical task processing time of the first instance being greater than or equal to the first remaining allowable delay In response to the delay, the first MEC device determines that the first instance does not meet the SLA condition.
  • the current time is the time at which the remaining allowable response delay is calculated, for example, it may be the reception time when the first MEC device receives the task request.
  • the task request is stored in the pending task queue, wherein the task requests in the pending task queue are stored sequentially from the head to the tail according to the remaining allowable response delay from small to large.
  • the first MEC device sequentially fetches task requests from the head of the to-be-processed task queue for processing.
  • the MEC device may preferentially process the remaining task requests with a smaller allowable response delay, so as to meet the delay requirements of each task request as much as possible, and improve the reliability of the MEC device for processing tasks.
  • step 206 is performed.
  • the first MEC device determines that the task cannot be satisfied currently Requested SLA conditions and return processing failure information for this task request.
  • the forwarding times threshold may be 3, 4 or 5.
  • the forwarding times threshold is set to limit the forwarding times of the task request, so as to avoid excessive forwarding of the task request and additionally cause a larger transmission delay of the task request and/or the response to the task request, and also limit the In order to parse the number of MEC devices requested by the task, it avoids occupying too many processing resources of the MEC devices.
  • Step 206 The first MEC device determines the second MEC device among the neighboring MEC devices of the first MEC device.
  • a second instance of the target application service is deployed in the second MEC device and the second instance satisfies the SLA condition.
  • the implementation process of step 206 includes: the first MEC device determines a target neighbor MEC device among the neighbor MEC devices according to the application information of the neighbor MEC devices.
  • the first MEC device determines the second MEC device among the target neighbor MEC devices.
  • the target neighbor MEC device satisfies: an instance of the target application service is deployed, and the historical task processing time of the deployed instance of the target application service is less than the remaining allowable response delay.
  • the current time is the time at which the remaining allowable response delay is calculated.
  • the implementation process of the first MEC device determining the second MEC device in the target neighbor MEC device may include: in response to the number of the target neighbor MEC devices being greater than 1, the first MEC device according to the multiple target neighbor MEC devices the historical task processing time of the instance of the target application service deployed in the device, the load of the instance of the target application service deployed in the multiple target neighbor MEC devices and/or the distance between the multiple target neighbor MEC devices and the first MEC device, A second MEC device is determined among the plurality of target neighbor MEC devices.
  • the first MEC device may determine the target neighbor MEC device with the shortest historical task processing time of the instances of the target application service deployed among the multiple target neighbor MEC devices as the second MEC device.
  • the first MEC device may also determine any target neighbor MEC device as the second MEC device.
  • the first MEC device determines that the SLA condition of the task request cannot currently be satisfied, and returns a processing failure for the task request. information.
  • the routing object of the task request is determined, and the routing object has a high probability of satisfying the SLA condition for responding to the task request, thereby improving the reliability of task processing.
  • the implementation process of step 206 includes: the first MEC device, according to the application information of the neighbor MEC device and the network response delay between the neighbor MEC device and the first MEC device,
  • the target neighbor MEC device is determined among the neighbor MEC devices.
  • the first MEC device determines the second MEC device among the target neighbor MEC devices.
  • the target neighbor MEC device satisfies: an instance of the target application service is deployed, and the sum of the network response delay with the first MEC device and the historical task processing time of the deployed instance of the target application service is less than the remaining allowable response delay .
  • the implementation process of the first MEC device determining the second MEC device in the target neighbor MEC device may include: in response to the number of the target neighbor MEC devices being greater than 1, the first MEC device converts multiple target neighbor MEC devices to Among the devices, the target neighbor MEC device with the smallest sum of the network response delay with the first MEC device and the historical task processing time of the deployed instance of the target application service is determined as the second MEC device.
  • the first MEC device may also determine any target neighbor MEC device as the second MEC device.
  • the first MEC device determines that the SLA condition of the task request cannot currently be satisfied, and returns a processing failure for the task request. information.
  • the routing object of the task request is determined, and the probability that the routing object can satisfy the SLA condition for responding to the task request is compared with that of the MEC device.
  • the method of only considering the processing time of the task request by the MEC device is higher, thereby improving the reliability of task processing.
  • the processing time of task requests by MEC devices and the network between MEC devices are considered comprehensively.
  • Response delay is used to schedule task requests, which can provide deterministic task response delay and improve user experience.
  • the task request includes the requested file information
  • the requested file information includes the file identifier of the file to be requested
  • the file identifier of the file to be requested can be a directory, a list of file names or a file description .
  • the implementation process of step 206 includes: the first MEC device determines the target neighbor MEC device among the neighbor MEC devices according to the application information and request file information of the neighbor MEC device and the network bandwidth between the neighbor MEC device and the first MEC device. The first MEC device determines the second MEC device among the target neighbor MEC devices.
  • the target neighbor MEC device satisfies: an instance of the target application service is deployed, the local file includes the file to be requested, and the file transfer time for transmitting the to-be-requested file to the first MEC device and the historical task processing time of the deployed instance of the target application service The sum is less than the remaining allowable response delay, and the file transmission time is determined based on the file size of the file to be requested and the network bandwidth between the neighbor MEC device and the first MEC device.
  • the file transmission time for the neighbor MEC device to transmit the to-be-requested file to the first MEC device is equal to the file size of all the to-be-requested files divided by the network bandwidth between the neighbor MEC device and the first MEC device.
  • the process for the first MEC device to determine the target neighbor MEC device among the neighbor MEC devices includes: first, the first MEC device obtains, based on the routing table in the third possible situation in step 203, the neighbor MEC devices that contain all the files to be requested and each of them. The size of the file to be requested. Then, the first MEC device calculates the file transfer time of each requested file, and accumulates the file transfer time of all requested files. Finally, the first MEC device determines the neighbor MEC device whose sum of the file transmission time of all requested files and the historical task processing time of the deployed instance of the target application service is less than the remaining allowable response delay as the target neighbor MEC device.
  • the implementation process of the first MEC device determining the second MEC device in the target neighbor MEC device may include: in response to the number of the target neighbor MEC devices being greater than 1, the first MEC device converts multiple target neighbor MEC devices to Among the devices, the target neighbor MEC device with the smallest sum of the file transmission time for transmitting the to-be-requested file to the first MEC device and the historical task processing time of the deployed instance of the target application service is determined as the second MEC device.
  • the first MEC device may also determine any target neighbor MEC device as the second MEC device.
  • the first MEC device determines that the SLA condition of the task request cannot currently be satisfied, and returns a processing failure for the task request. information.
  • the routing object of the task request is determined based on the processing time of the MEC device for the task request and the file transmission time for transmitting the file to be requested between the MEC devices. Compared with the method that only considers the processing time of the task request by the MEC device, the reliability of the task processing is improved.
  • the routing object of the task request may also be determined by considering the time when the task request is transmitted between the MEC devices. For data-intensive edge applications that are very sensitive to latency, such as artificial intelligence (AI) processing, the processing time of MEC devices for task requests, the file distribution of UE's files to be requested, and the transmission of pending requests between MEC devices are considered comprehensively.
  • the file transfer time of the file and the time for transferring task requests between MEC devices to schedule task requests can provide deterministic task response delays and improve user experience.
  • Step 207 The first MEC device forwards the task request to the second MEC device.
  • the first MEC device may also add 1 to the forwarding times of the task request, and send the task request and the updated forwarding times to the second MEC device together.
  • the second MEC device After the second MEC device receives the task request, it performs the same operations as after the first MEC device receives the task request, and details are not described herein again in this embodiment of the present application.
  • the second MEC device may further send a response to the task request to the first MEC device after processing the task request through the second instance.
  • the first MEC device After receiving the response to the task request sent by the second MEC device, the first MEC device forwards the response to the task request to the target device.
  • the target device is a device that sends the task request to the first MEC device.
  • the target device may be a UE or other MEC device.
  • the routing path of the response to the task request sent by the UE is opposite to the routing path of the task request.
  • the UE sends the task request to the accessed MEC device, and finally the MEC device accessed by the UE also sends the task request to the UE.
  • the entire process of processing the task request by the MEC device is unaware of the UE, so there is no need to change the way the UE accesses the application service, and the compatibility with the UE is good.
  • Step 208 The first MEC device processes the task request through the first instance.
  • the first MEC device may also send a response to the task request to the target device.
  • the first MEC device directly sends a response to the task request to the UE.
  • the first MEC device sends a response to the task request to the MEC device that sent the task request, that is, the routing path of the response to the task request and the task request The routing path of the request is opposite, and finally the MEC device accessed by the UE sends a response to the task request to the UE.
  • the task processing methods provided in the embodiments of the present application can be applied to multiple scenarios, including but not limited to: AR, VR, and games, such as cloud game services based on KubeEdge/intelligent edge fabric (IEF); interactive live broadcast, For example, live streaming services based on KubeEdge/IEF; AI processing, such as face recognition and target tracking services based on KubeEdge/IEF.
  • AR virtual reality
  • VR virtual reality
  • games such as cloud game services based on KubeEdge/intelligent edge fabric (IEF)
  • IEF extensible-toprivity
  • interactive live broadcast For example, live streaming services based on KubeEdge/IEF
  • AI processing such as face recognition and target tracking services based on KubeEdge/IEF.
  • the state, load, task processing time and other factors of the MEC device are comprehensively considered to distribute the traffic multiple times. It can distribute UE's access traffic to other appropriate MEC devices, and improve the reliability and availability of end-to-end collaborative task processing. In scenarios such as a sudden increase of UEs in an area or a high load of local MEC devices, the traffic is distributed to other available MEC devices through multiple traffic distribution to ensure the availability and continuity of services.
  • route selection takes into account factors such as the maximum allowable response latency, network response latency between MEC devices, task processing time, and file transfer time to provide deterministic task response latency and improve user experience. .
  • factors such as the maximum allowable response latency, network response latency between MEC devices, task processing time, and file transfer time to provide deterministic task response latency and improve user experience.
  • multiple traffic distributions are used to meet application latency requirements and ensure application user experience. The method is realized by mutual perception between MEC devices, but the UE has no perception, so there is no need to change the way the UE accesses the application service, and the compatibility with the UE is good.
  • the MEC device when the MEC device itself does not meet the SLA conditions that should be satisfied in responding to the task request, it is not limited by the geographical location of the UE, and can flexibly select the MEC device that processes the task request among the neighboring MEC devices, making the task processing more flexible. high.
  • FIG. 3 is a schematic structural diagram of a task processing apparatus provided by an embodiment of the present application. The apparatus is applied to the first MEC device. As shown in Figure 3, the device 30 includes:
  • the receiving module 301 is configured to receive a task request from the UE, where the task request includes the application service identifier of the target application service requested by the UE to access and the service level agreement SLA information corresponding to the task request, and the SLA information is used to determine the SLA information that is satisfied in response to the task request. SLA conditions.
  • the first determining module 302 is configured to determine, in response to the first instance of the target application service deployed in the first MEC device not meeting the SLA condition, determining the second MEC device in the neighboring MEC devices of the first MEC device, and the second MEC device in the second MEC device. A second instance of the target application service is deployed and the second instance satisfies the SLA conditions.
  • the sending module 303 is configured to forward the task request to the second MEC device.
  • the SLA information corresponding to the task request indicates the maximum allowable response delay of the task request
  • the SLA conditions satisfied by the response task request include: the total delay of responding to the task request is not greater than the maximum allowable response delay
  • the device 30 further includes:
  • the acquiring module 304 is configured to acquire application information of the neighboring MEC device, where the application information includes the historical task processing time of the deployed application service instance.
  • the first determination module 302 includes: a first determination sub-module 3021, configured to determine a target neighbor MEC device among the neighbor MEC devices according to the application information of the neighbor MEC devices, wherein the target neighbor MEC device satisfies: the target application service is deployed with the target neighbor MEC device. instance, and the historical task processing time of the instance of the deployed target application service is less than the remaining allowable response delay, and the remaining allowable response delay is obtained based on the maximum allowable response delay.
  • the second determination sub-module 3022 is configured to determine the second MEC device in the target neighbor MEC device.
  • the application information further includes the load of the deployed application service instance
  • the second determination submodule 3022 is configured to: in response to the number of target neighbor MEC devices being greater than 1, according to the target application deployed in the multiple target neighbor MEC devices The historical task processing time of the instance of the service, the load of the instance of the target application service deployed in the multiple target neighbor MEC devices, and/or the distance between the multiple target neighbor MEC devices and the first MEC device, in the multiple target neighbor MEC devices A second MEC device is determined in the device.
  • the first determining submodule 3021 is configured to: determine the target neighbor MEC device among the neighbor MEC devices according to the application information of the neighbor MEC device and the network response delay between the neighbor MEC device and the first MEC device.
  • the target neighbor MEC device satisfies: an instance of the target application service is deployed, and the sum of the network response delay with the first MEC device and the historical task processing time of the deployed instance of the target application service is less than the remaining allowable response delay .
  • the second determination sub-module 3022 is configured to: in response to the number of target neighbor MEC devices being greater than 1, the first MEC device tracks the network response delay between the multiple target neighbor MEC devices and the first MEC device with the following steps.
  • the target neighbor MEC device with the smallest sum of historical task processing times of the deployed instances of the target application service is determined as the second MEC device.
  • the task request further includes requested file information
  • the requested file information includes a file identifier of the file to be requested
  • the application information further includes file information
  • the file information includes the file identifier of the local file and the file size of the local file.
  • the first determining submodule 3021 is configured to: determine the target neighbor MEC device among the neighbor MEC devices according to the application information of the neighbor MEC device, the request file information and the network bandwidth between the neighbor MEC device and the first MEC device.
  • the target neighbor MEC device satisfies: an instance of the target application service is deployed, the local file includes the file to be requested, and the file transfer time for transmitting the to-be-requested file to the first MEC device and the historical task processing time of the deployed instance of the target application service The sum is less than the remaining allowable response delay, and the file transmission time is determined based on the file size of the file to be requested and the network bandwidth between the neighbor MEC device and the first MEC device.
  • the second determination sub-module 3022 is configured to: in response to the number of target neighbor MEC devices being greater than 1, compare the file transfer time of the multiple target neighbor MEC devices for transmitting the file to be requested to the first MEC device with the deployed target The target neighbor MEC device with the smallest sum of historical task processing times of the instances of the application service is determined as the second MEC device.
  • the sending module 303 is further configured to broadcast and send the application information of the first MEC device.
  • the first determination module 302 is configured to: in response to the first instance not meeting the SLA condition, and the forwarded times of the task request received by the first MEC device is less than the forwarding times threshold, the first MEC device is in the neighbor MEC device. identify the second MEC device.
  • the apparatus 30 further includes a second determining module 305 .
  • the receiving module 301 is further configured to receive the MEC device list sent by the application controller, where the MEC device list includes communication address information and geographic location information of multiple MEC devices.
  • the second determining module 305 is configured to determine, among the multiple MEC devices, a MEC device whose distance from the first MEC device is less than the distance threshold as a neighbor MEC device of the first MEC device.
  • the receiving module 301 is further configured to, after forwarding the task request to the second MEC device, receive a response to the task request sent by the second MEC device.
  • the sending module 303 is further configured to forward the response to the task request to the target device.
  • the target device is a device that sends a task request to the first MEC device.
  • the receiving module 301 is configured to: in response to the UE not being in the service area corresponding to the first MEC device, receive the task request forwarded by the third MEC device. Or, in response to the UE being in the service area corresponding to the first MEC device, receiving the task request sent by the UE or the task request forwarded by the fourth MEC device.
  • the apparatus 30 further includes: a processing module 306, configured to process the task request through the first instance in response to the first instance satisfying the SLA condition.
  • the task processing device after receiving the task request sent by the UE through the receiving module, firstly judges whether it satisfies the SLA conditions that should be satisfied in response to the task request. After it is determined by the first determination module that it satisfies the SLA condition, the task request is processed; after it is determined by the first determination module that it does not meet the SLA condition, the task request is dispatched to the neighbor MEC device that satisfies the SLA condition by the sending module , the task request is processed by the neighbor MEC device that satisfies the SLA condition.
  • the response to the task request can meet the SLA conditions, and the reliability and availability of the task processing are improved.
  • the route selection for multiple traffic distribution is performed, and the UE's access traffic is distributed to other appropriate MEC devices. Improve the reliability and availability of end-to-end collaborative task processing. In scenarios such as a sudden increase of UEs in an area or a high load of local MEC devices, the traffic is distributed to other available MEC devices through multiple traffic distribution to ensure the availability and continuity of services. For latency-sensitive services, route selection takes into account factors such as the maximum allowable response latency, network response latency between MEC devices, task processing time, and file transfer time to provide deterministic task response latency and improve user experience. .
  • multiple traffic distributions are used to meet application latency requirements and ensure application user experience.
  • the method is realized by mutual perception between MEC devices, but the UE has no perception, so there is no need to change the way the UE accesses the application service, and the compatibility with the UE is good.
  • the MEC device itself does not meet the SLA conditions that should be satisfied in responding to the task request, it is not limited by the geographical location of the UE, and can flexibly select the MEC device that processes the task request among the neighboring MEC devices, making the task processing more flexible. high.
  • FIG. 7 is a block diagram of an MEC device provided by an embodiment of the present application. As shown in FIG. 7 , the MEC device 70 includes: a processor 701 and a memory 702 .
  • a memory 702 for storing a computer program, the computer program including program instructions
  • the processor 701 is configured to invoke the computer program to implement the task processing method described in the above method embodiments.
  • the MEC device 70 further includes a communication bus 703 and a communication interface 704 .
  • the processor 701 includes one or more processing cores, and the processor 701 executes various functional applications and data processing by running a computer program.
  • Memory 702 may be used to store computer programs.
  • the memory may store the operating system and application program elements required for at least one function.
  • the operating system may be an operating system such as a real-time operating system (Real Time eXecutive, RTX), LINUX, UNIX, WINDOWS, or OS X.
  • the communication interfaces 704 are used to communicate with other devices.
  • the communication interface 704 of the first MEC device may be used to send a task request to the second MEC device.
  • the memory 702 and the communication interface 704 are respectively connected to the processor 701 through the communication bus 703 .
  • the embodiment of the present application also provides a task processing system, where the system includes: a plurality of MEC devices.
  • the MEC equipment includes the device shown in any one of FIG. 3 to FIG. 6 or the MEC equipment shown in FIG. 7 .
  • Embodiments of the present application further provide a computer storage medium, where instructions are stored on the computer storage medium, and when the instructions are executed by a processor, the task processing method described in the foregoing method embodiments is implemented.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Computer Hardware Design (AREA)
  • General Engineering & Computer Science (AREA)
  • Information Transfer Between Computers (AREA)

Abstract

本申请公开了一种任务处理方法、装置及系统、计算机存储介质，属于通信技术领域。第一MEC设备接收来自UE的任务请求，该任务请求包括UE所请求访问的目标应用服务的应用服务标识以及任务请求对应的SLA信息，该SLA信息用于确定响应该任务请求所满足的SLA条件。响应于第一MEC设备中部署的目标应用服务的第一实例不满足SLA条件，第一MEC设备在第一MEC设备的邻居MEC设备中确定部署有目标应用服务的第二实例且第二实例满足SLA条件的第二MEC设备。第一MEC设备向第二MEC设备转发该任务请求。本申请通过MEC设备之间的相互感知对任务请求进行合理调度，提高了任务处理的可靠性和可用性。

Description

任务处理方法、装置及系统、计算机存储介质
本申请要求于2020年12月30日提交的申请号为202011613356.4、发明名称为“任务处理方法、装置及系统、计算机存储介质”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及通信技术领域，特别涉及一种任务处理方法、装置及系统、计算机存储介质。
背景技术
无线通信网络的建设是以数据中心(data center)为中心的。各数据中心使用分布式云技术进行管理，形成边缘云(edge cloud)加中心云(central cloud)的分级网络系统。数据中心为用户设备(user equipment，UE)提供应用服务(application server)的应用实例(以下简称为实例)部署在各边缘云上，使得应用服务的实例部署到更靠近UE的位置，降低了服务时延，提高了服务体验。
目前，UE访问应用服务的过程包括：UE向管理多接入边缘计算(multi-access edge computing，MEC)设备的应用控制器发送服务路由请求，该服务路由请求包括该UE所请求访问的应用服务的应用服务标识(identification，ID)。应用控制器根据该应用服务ID获取部署有该应用服务的实例的MEC设备，然后根据UE的地理位置信息从部署有该应用服务的实例的MEC设备中获取距离UE最近的MEC设备，并将该MEC设备的通信地址信息发送给UE。UE基于该MEC设备的通信地址信息接入该MEC设备，然后向该MEC设备发送任务请求，实现应用服务的访问。
但是，当某个区域内需要访问同一应用服务的UE的数量较多时，会出现多个UE访问同一个MEC设备上的该应用服务的实例的情况，导致该MEC设备上的该实例负载过大，进而导致UE的访问时延较高。因此目前MEC设备提供应用服务的可靠性较差。
发明内容
本申请提供了一种任务处理方法、装置及系统、计算机存储介质，可以解决目前MEC设备提供应用服务的可靠性较差的问题。
第一方面，提供了一种任务处理方法。该方法包括：第一MEC设备接收来自UE的任务请求，该任务请求包括UE所请求访问的目标应用服务的应用服务标识以及任务请求对应的SLA信息，该SLA信息用于确定响应该任务请求所满足的SLA条件。响应于第一MEC设备中部署的目标应用服务的第一实例不满足SLA条件，第一MEC设备在第一MEC设备的邻居MEC设备中确定第二MEC设备，该第二MEC设备中部署有目标应用服务的第二实例且第二实例满足SLA条件。第一MEC设备向第二MEC设备转发该任务请求。
本申请中，MEC设备接收到来自UE的任务请求后，在确定自身不满足响应该任务请求应满足的SLA条件后，将任务请求调度给满足该SLA条件的邻居MEC设备，由满足该SLA 条件的邻居MEC设备处理该任务请求。通过对任务请求的合理调度，使得对该任务请求的响应能够满足SLA条件，提高了任务处理的可靠性和可用性。
可选地，任务请求对应的SLA信息指示任务请求的最大允许响应时延，响应任务请求所满足的SLA条件，包括：响应任务请求的总时延不大于最大允许响应时延。
本申请中，基于任务请求的最大允许响应时延来选择处理该任务请求的MEC设备，可以保证的UE的访问时延，提高MEC设备提供应用服务的可靠性。
可选地，第一MEC设备还可以获取邻居MEC设备的应用信息，该应用信息包括部署的应用服务的实例的历史任务处理时间。则第一MEC设备在第一MEC设备的邻居MEC设备中确定第二MEC设备的实现过程，包括：第一MEC设备根据邻居MEC设备的应用信息，在邻居MEC设备中确定目标邻居MEC设备。其中，目标邻居MEC设备满足：部署有目标应用服务的实例，且部署的目标应用服务的实例的历史任务处理时间小于剩余允许响应时延，剩余允许响应时延基于最大允许响应时延得到。第一MEC设备在目标邻居MEC设备中确定第二MEC设备。
本申请中，基于MEC设备对任务请求的处理时间的考虑，确定任务请求的路由对象，该路由对象满足响应任务请求的SLA条件的概率较高，从而提高任务处理的可靠性。
可选地，应用信息还包括部署的应用服务的实例的负载，第一MEC设备在目标邻居MEC设备中确定第二MEC设备的实现过程，包括：响应于目标邻居MEC设备的数量大于1，第一MEC设备根据多个目标邻居MEC设备中部署的目标应用服务的实例的历史任务处理时间、多个目标邻居MEC设备中部署的目标应用服务的实例的负载和/或多个目标邻居MEC设备与第一MEC设备之间的距离，在多个目标邻居MEC设备中确定第二MEC设备。
本申请中，综合考虑MEC设备的状态、负载和任务处理时间等因素进行流量多次分发的路由选择，把UE的访问流量分发到合适的其它MEC设备，提升端边协同任务处理的可靠性和可用性。
可选地，第一MEC设备根据邻居MEC设备的应用信息，在邻居MEC设备中确定目标邻居MEC设备的实现过程，包括：第一MEC设备根据邻居MEC设备的应用信息以及邻居MEC设备与第一MEC设备之间的网络响应时延，在邻居MEC设备中确定目标邻居MEC设备。其中，目标邻居MEC设备满足：部署有目标应用服务的实例，且与第一MEC设备之间的网络响应时延跟部署的目标应用服务的实例的历史任务处理时间之和小于剩余允许响应时延。
相应地，第一MEC设备在目标邻居MEC设备中确定第二MEC设备的实现过程，可以包括：响应于目标邻居MEC设备的数量大于1，第一MEC设备将多个目标邻居MEC设备中，与第一MEC设备之间的网络响应时延跟部署的目标应用服务的实例的历史任务处理时间之和最小的目标邻居MEC设备确定为第二MEC设备。
本申请中，基于MEC设备对任务请求的处理时间以及MEC设备之间的网络响应时延的考虑，确定任务请求的路由对象，该路由对象能够满足响应任务请求的SLA条件的概率相较于只考虑MEC设备对任务请求的处理时间的方式更高，从而提高任务处理的可靠性。对于对时延非常敏感的计算密集型边缘应用，综合考虑MEC设备对任务请求的处理时间以及MEC设备之间的网络响应时延来调度任务请求，可以提供确定性的任务响应时延，提高用户体验。
可选地，任务请求还包括请求文件信息，请求文件信息包括待请求文件的文件标识，应 用信息还包括文件信息，文件信息包括本地文件的文件标识和本地文件的文件大小。第一MEC设备根据邻居MEC设备的应用信息，在邻居MEC设备中确定目标邻居MEC设备的实现过程，包括：第一MEC设备根据邻居MEC设备的应用信息、请求文件信息以及邻居MEC设备与第一MEC设备之间的网络带宽，在邻居MEC设备中确定目标邻居MEC设备。其中，目标邻居MEC设备满足：部署有目标应用服务的实例，本地文件包括待请求文件，且向第一MEC设备传输待请求文件的文件传输时间跟部署的目标应用服务的实例的历史任务处理时间之和小于剩余允许响应时延，文件传输时间基于待请求文件的文件大小以及邻居MEC设备与第一MEC设备之间的网络带宽确定。
相应地，第一MEC设备在目标邻居MEC设备中确定第二MEC设备的实现过程，可以包括：响应于目标邻居MEC设备的数量大于1，第一MEC设备将多个目标邻居MEC设备中，向第一MEC设备传输待请求文件的文件传输时间跟部署的目标应用服务的实例的历史任务处理时间之和最小的目标邻居MEC设备确定为第二MEC设备。
本申请中，基于MEC设备对任务请求的处理时间以及MEC设备之间传输待请求文件的文件传输时间，确定任务请求的路由对象，该路由对象能够满足响应任务请求的SLA条件的概率相较于只考虑MEC设备对任务请求的处理时间的方式更高，从而提高任务处理的可靠性。当任务请求包括待上传文件时，还可以考虑MEC设备之间传输任务请求的时间来确定任务请求的路由对象。对于对时延非常敏感的数据密集型边缘应用，综合考虑MEC设备对任务请求的处理时间、UE的待请求文件的文件分布、MEC设备之间传输待请求文件的文件传输时间以及MEC设备之间传输任务请求的时间来调度任务请求，可以提供确定性的任务响应时延，提高用户体验。
可选地，第一MEC设备还可以广播发送第一MEC设备的应用信息。
可选地，响应于第一MEC设备中部署的目标应用服务的第一实例不满足SLA条件，第一MEC设备在第一MEC设备的邻居MEC设备中确定第二MEC设备的实现过程，包括：响应于第一实例不满足SLA条件，且第一MEC设备接收到的任务请求的已转发次数小于转发次数阈值，第一MEC设备在邻居MEC设备中确定第二MEC设备。
可选地，第一MEC设备还可以接收应用控制器发送的MEC设备列表，该MEC设备列表包括多个MEC设备的通信地址信息和地理位置信息。第一MEC设备将多个MEC设备中，与第一MEC设备之间的距离小于距离阈值的MEC设备确定为第一MEC设备的邻居MEC设备。
可选地，在第一MEC设备向第二MEC设备转发任务请求之后，第一MEC设备还可以接收第二MEC设备发送的针对该任务请求的响应。第一MEC设备向目标设备转发针对该任务请求的响应。目标设备为向第一MEC设备发送任务请求的设备。该目标设备可以是UE，或者也可以是其它MEC设备。
本申请中，针对UE发送的任务请求的响应的路由路径与该任务请求的路由路径相反，UE向接入的MEC设备发送任务请求，最终也由UE接入的MEC设备向UE发送针对该任务请求的响应，MEC设备处理任务请求的整个过程UE是无感知的，因此无需改变UE访问应用服务的方式，对UE的兼容性较好。
可选地，第一MEC设备接收来自UE的任务请求的可能实现方式，包括：响应于UE不在第一MEC设备对应的服务区域内，第一MEC设备接收第三MEC设备转发的任务请求。 或者，响应于UE在第一MEC设备对应的服务区域内，第一MEC设备接收UE发送的任务请求或者第四MEC设备转发的任务请求。
本申请中，最终处理来自UE的任务请求的MEC设备对应的服务区域可以覆盖UE所在位置，也可以不覆盖UE所在位置，也即是，MEC设备在自身不满足响应任务请求应满足的SLA条件时，不受限于UE的地理位置，可以在邻居MEC设备中灵活选择处理该任务请求的MEC设备，使得任务处理的灵活性较高。
可选地，响应于第一实例满足SLA条件，第一MEC设备通过第一实例处理任务请求。
第二方面，提供了一种任务处理装置。所述装置包括多个功能模块，所述多个功能模块相互作用，实现上述第一方面及其各实施方式中的方法。所述多个功能模块可以基于软件、硬件或软件和硬件的结合实现，且所述多个功能模块可以基于具体实现进行任意组合或分割。
第三方面，提供了一种MEC设备，包括：处理器和存储器；
所述存储器，用于存储计算机程序，所述计算机程序包括程序指令；
所述处理器，用于调用所述计算机程序，实现上述第一方面及其各实施方式中的方法。
第四方面，提供了一种任务处理系统，所述系统包括：多个多接入边缘计算MEC设备，所述MEC设备包括如第二方面所述的装置或为如第三方面所述的MEC设备。
第五方面，提供了一种计算机存储介质，所述计算机存储介质上存储有指令，当所述指令被处理器执行时，实现上述第一方面及其各实施方式中的方法。
第六方面，提供了一种芯片，芯片包括可编程逻辑电路和/或程序指令，当芯片运行时，实现上述第一方面及其各实施方式中的方法。
本申请提供的技术方案带来的有益效果至少包括：
本申请中，在端边复杂网络、多层次边缘的系统环境下，综合考虑MEC设备的状态、负载和任务处理时间等因素进行流量多次分发的路由选择，把UE的访问流量分发到合适的其它MEC设备，提升端边协同任务处理的可靠性和可用性。在区域的UE突然大幅增加或局部MEC设备负载过高等场景下，通过多次流量分发，把流量分发到可用的其它MEC设备，保证业务的可用性和持续性。对于时延敏感的业务，综合考虑最大允许响应时延、MEC设备之间的网络响应时延、任务处理时间、文件传输时间等因素进行路由选择，提供确定性的任务响应时延，提升用户体验。对于计算密集型和数据密集的应用场景，通过多次流量分发，满足应用的时延要求，保证应用的用户体验。该方法通过MEC设备之间相互感知实现，而UE无感知，因此无需改变UE访问应用服务的方式，对UE的兼容性较好。另外，MEC设备在自身不满足响应任务请求应满足的SLA条件时，不受限于UE的地理位置，可以在邻居MEC设备中灵活选择处理该任务请求的MEC设备，使得任务处理的灵活性较高。
附图说明
图1是本申请实施例提供的一种通信网络的基础架构示意图；
图2是本申请实施例提供的一种任务处理方法的流程示意图；
图3是本申请实施例提供的一种任务处理装置的结构示意图；
图4是本申请实施例提供的另一种任务处理装置的结构示意图；
图5是本申请实施例提供的又一种任务处理装置的结构示意图；
图6是本申请实施例提供的再一种任务处理装置的结构示意图；
图7是本申请实施例提供的一种MEC设备的框图。
具体实施方式
为使本申请的目的、技术方案和优点更加清楚，下面将结合附图对本申请实施方式作进一步地详细描述。
图1是本申请实施例提供的一种通信网络的基础架构示意图。如图1所示，该通信网络包括：核心网(core network，CN)设备101、基站102A-102B(统称为基站102)以及MEC设备103A-103B(统称为MEC设备103)。其中，各个MEC设备103与对应的基站102相连，例如MEC设备103A与基站102A相连，MEC设备103B与基站102B相连。各个基站102与核心网设备101相连。图中各个设备的数量仅用作示例性说明，不作为对本申请实施例涉及的通信网络的限定。
MEC设备103是实现边缘云部署的基础设施平台。例如，MEC设备103可以由一个或多个服务器组成，或者由一个或多个服务器与其它网络设备(比如网关、交换机等设备)组成，或者为云计算平台，所组成的网络架构形成的网络资源用于给实例提供服务平台。每个MEC设备103有对应的服务区域，该服务区域可以是与MEC设备103相连的基站102对应的小区。例如，MEC设备103A对应的服务区域包括基站102A对应的小区，MEC设备103B对应的服务区域包括基站102B对应的小区。每个MEC设备103中部署有至少一个实例，该实例用于为该MEC设备103对应的服务区域内的UE提供应用服务。在本申请实施例中，应用服务指：通过各个类型的应用为UE提供的服务，比如某个开发商所运营的车辆网应用服务。而实例指的是一个应用服务部署在不同的MEC设备上的具体的应用。因此，一个应用服务可以对应多个实例。例如，本申请实施例中，第一MEC设备中部署有第一实例，第二MEC设备中部署有第二实例，其中，第一实例与第二实例为同一应用服务(目标应用服务)的实例。
如图1所示的通信网络可以扩展出多种变形。举例来说，可以对核心网进行控制面(control plane，CP)和用户面(user plane，UP)的划分，形成控制面与用户面分离(control and user plane separation，CUPS)通信网络。这种场景下，核心网设备101只包含控制面功能。而核心网的用户面由各网关(图中未示出)实现，即各网关属于核心网侧设备，执行核心网的用户面功能。网关与基站的数据连接由核心网控制面统一管控。
可选地，基站102为新无线接入基站(new radio gNodeB，NR gNB)或其它无线接入设备，本申请实施例对此不做限定。
本申请实施例中，应用服务的实例分布式地部署在基站附近的多个MEC设备上，用于为其相连基站的小区内的UE提供应用服务。举例来说，超高可靠低时延(ultra reliable and low latency communication，URLLC)应用的实例会被部署在多个靠近基站的MEC设备上。每个MEC设备上所部署的URLLC实例分别为UE提供URLLC服务。
MEC设备中可以记载有各应用服务的应用服务ID。应用服务ID用于唯一标识对应的应用服务，以将不同的应用服务区分开。应用服务ID可以是域名(domain name)的形式，或者，也可以由MEC设备统一为每个应用服务分配应用服务ID，比如应用服务ID可以为互联网协议(Internet Protocol，IP)地址。应用服务ID也可以是其它的形式，本申请实施例对此不做限定。
每个应用服务的实例还具有实例ID。实例ID用于唯一标识对应的实例，以将同一应用服务的不同实例区分开。即，同一应用服务的不同实例，其应用服务ID相同，但实例ID不同。实例ID可以是实例的IP地址的形式，也可以是其它的形式，本申请实施例对此不做限定。需要说明的是，当应用服务ID与实例ID均为IP地址形式时，发送端和/或接收端可通过报文中的不同字段对两者进行区分。比如，发送端与接收端互相约定报文的第一个字段为应用服务ID，第二个字段为实例ID，则发送端在向接收端发送报文时，在该报文的第一个字段设置的是应用服务的IP地址，在该报文的第二个字段设置的是实例的IP地址。接收端接收到该报文后，则能够从相应的字段中解析出应用服务的IP地址以及实例的IP地址。
请继续参见图1，该通信网络还可以包括应用控制器104。应用控制器104用于管理通信网络中的各MEC设备，进而对通信网络中的所有应用进行管控。应用控制器104为核心网侧的网络设备，应用控制器104连接核心网设备101，用于通过核心网设备101获取UE的地理位置信息。应用控制器104连接通信网络中的各个MEC设备，用于向各个MEC设备103发送控制信息。应用控制器104还连接各个基站侧的网络设备(比如基站)，用于根据用户侧的需求进行应用实例的控制决策等，例如决策UE接入哪个MEC设备102。
应用控制器104中存储有通信网络中的各个MEC设备103的信息，包括MEC设备103与基站102的连接关系以及MEC设备103的通信地址信息、地理位置信息和部署有哪些应用服务的实例等。该通信地址信息可以是IP地址。该地理位置信息可以采用城市名称和/或经纬度表示。应用控制器104还用于向MEC设备103提供其它MEC设备的通信地址信息和地理位置信息。应用控制器104还可以向MEC设备103提供其它MEC设备中所部署的应用服务的应用服务ID等信息。
本申请实施例中，当UE想要访问目标应用服务时，UE会向应用控制器104发送服务路由请求，该服务路由请求包括目标应用服务的应用服务ID。应用控制器104从核心网设备101处获取UE的地理位置信息。其中，核心网设备101可以根据UE连接的基站确定UE的地理位置。应用控制器104可以根据UE的地理位置信息从部署有目标应用服务的实例的MEC设备103中获取服务区域覆盖UE所在位置的MEC设备，该MEC设备可以是服务区域覆盖UE所在位置且距离UE最近的部署有目标应用服务的实例的MEC设备，或者，该MEC设备也可以是服务区域覆盖UE所在位置的部署有目标应用服务的实例的任一MEC设备。然后应用控制器104将获取的该MEC设备的通信地址信息发送给该UE。UE基于该MEC设备的通信地址信息接入该MEC设备，然后向该MEC设备发送任务请求。例如，在如图1所示的通信网络中，MEC设备103A和MEC设备103B中都部署有目标应用服务的实例，当请求访问目标应用服务的UE处于基站102A覆盖的小区时，应用控制器104会控制决策该UE接入MEC设备103A。
本申请实施例中，当MEC设备接收到来自UE的任务请求时，首先判断自身是否满足响 应该任务请求应满足的服务等级协议(service level agreement，SLA)条件。在确定自身满足该SLA条件后，处理该任务请求；在确定自身不满足该SLA条件后，将任务请求调度给满足该SLA条件的邻居MEC设备，由满足该SLA条件的邻居MEC设备处理该任务请求。通过对任务请求的合理调度，使得对该任务请求的响应能够满足SLA条件，提高了任务处理的可靠性和可用性。
可选地，图2是本申请实施例提供的一种任务处理方法的流程示意图。该方法可以应用于如图1所示的通信网络中。如图2所示，该方法包括：
步骤201、第一MEC设备接收应用控制器发送的MEC设备列表。
该MEC设备列表包括多个MEC设备的通信地址信息和地理位置信息。该通信地址信息可以是IP地址。该地理位置信息可以包括MEC设备所处的城市和/或所在位置的经纬度。该MEC设备列表还可以包括该多个MEC设备中所部署的应用服务的应用服务ID。
可选地，MEC设备列表中的多个MEC设备可以包括应用控制器管理的所有MEC设备，或者，也可以仅包括与第一MEC设备部署有至少一个相同应用服务的实例的MEC设备。
本申请实施例中，第一MEC设备中部署有目标应用服务的第一实例。该目标应用服务可以是任一应用服务。
步骤202、第一MEC设备将多个MEC设备中，与第一MEC设备之间的距离小于距离阈值的MEC设备确定为第一MEC设备的邻居MEC设备。
也即是，第一MEC设备的邻居MEC设备为与第一MEC设备之间的距离小于距离阈值的MEC设备。第一MEC设备在接收到MEC设备列表后，可以根据该第一MEC设备的地理位置信息以及MEC设备列表中各个MEC设备的地理位置信息，分别计算各个MEC设备到第一MEC设备的距离，并在确定邻居MEC设备后存储该邻居MEC设备的通信地址信息和地理位置信息。
可选地，第一MEC设备将多个MEC设备中，与第一MEC设备之间的距离小于距离阈值且部署有至少一个相同应用服务的实例的MEC设备确定为第一MEC设备的邻居MEC设备。由于第一MEC设备获取邻居MEC设备的目的是为了调度来自UE的任务请求，而该任务请求只能被调度至部署有该任务请求所请求访问的应用服务的实例的MEC设备，因此只有与第一MEC设备部署有至少一个相同应用服务的实例的MEC设备才有可能接收到被第一MEC设备调度的任务请求，因此设置第一MEC设备的邻居MEC设备满足与第一MEC设备部署有至少一个相同应用服务的实例的条件，在保证第一MEC设备对任务请求的调度有效性和全面性的同时，可以减少第一MEC设备存储邻居MEC设备的信息所占用的内容资源，还可以减少邻居MEC设备的数量以提高后续第一MEC设备对任务请求的调度效率。
步骤203、第一MEC设备获取该第一MEC设备的邻居MEC设备的应用信息。
第一MEC设备的邻居MEC设备的应用信息包括该邻居MEC设备中部署的应用服务的实例的历史任务处理时间。该历史任务处理时间可以包括最近N分钟内的任务请求最大处理时间和/或任务请求平均处理时间，N为正数。该历史任务处理时间还可以包括最近N分钟内的任务请求最小处理时间。
可选地，第一MEC设备的邻居MEC设备的应用信息还包括该邻居MEC设备中部署的应用服务的实例的负载，即该邻居MEC设备中部署的每个应用服务的实例对应的中央处理器(central processing unit，CPU)占用率。其中，MEC设备中运行一个应用服务的实例的CPU 资源通常是预先划分的。
第一MEC设备还可以统计该第一MEC设备的应用信息。
第一MEC设备在确定该第一MEC设备的邻居MEC设备并获取该邻居MEC设备的应用信息后，可以建立区域性的路由表。
第一种可能情况下，第一MEC建立的路由表可以包括如下内容：
本地MEC设备：
IP地址：xx.xx.xx.xx
地理位置：广州(经度/纬度)
APP列表：
APP1：
CPU占用率
历史任务处理时间
APP2：
CPU占用率
历史任务处理时间
邻居MEC设备1：
IP地址：xx.xx.xx.xx
地理位置：深圳(经度/纬度)
与第一MEC设备之间的距离：xxx公里
APP列表：
APP1：
CPU占用率
历史任务处理时间
APP3：
CPU占用率
历史任务处理时间
邻居MEC设备2：
…
邻居MEC设备3：
…
其中，APP列表中包括MEC设备中部署的各种类型的应用服务的实例的信息。
可选地，第一MEC设备可以向邻居MEC设备发送网络探测请求，以统计邻居MEC设备与该第一MEC设备之间的网络响应时延。邻居MEC设备与该第一MEC设备之间的网络响应时延可以指：第一MEC设备发送网络探测请求到接收到邻居MEC设备基于该网络探测请求回复的网络探测响应的总时长。该网络响应时延可以包括最近N分钟内的最大网络响应时延、平均网络响应时延和/或最小网络响应时延。第一MEC设备可以周期性地统计邻居MEC设备与该第一MEC设备之间的网络响应时延，以便实时更新网络响应时延。
第二种可能情况下，第一MEC建立的路由表可以包括如下内容：
本地MEC设备：
IP地址：xx.xx.xx.xx
地理位置：广州(经度/纬度)
APP列表：
APP1：
CPU占用率
历史任务处理时间
APP2：
CPU占用率
历史任务处理时间
邻居MEC设备1：
IP地址：xx.xx.xx.xx
地理位置：深圳(经度/纬度)
与第一MEC设备之间的距离：xxx公里
网络响应时延
APP列表：
APP1：
CPU占用率
历史任务处理时间
APP3：
CPU占用率
历史任务处理时间
邻居MEC设备2：
…
邻居MEC设备3：
…
可选地，第一MEC设备的邻居MEC设备的应用信息还包括该邻居MEC设备的文件信息，该文件信息包括本地文件的文件标识和本地文件的文件大小。该文件标识可以是文件名称。文件信息还可以包括本地文件的目录。
可选地，第一MEC设备可以向邻居MEC设备发送网络探测请求，以统计邻居MEC设备与该第一MEC设备之间的网络带宽。该网络带宽可以包括最近N分钟内的最大网络带宽、平均网络带宽和/或最小网络带宽。第一MEC设备可以周期性地统计邻居MEC设备与该第一MEC设备之间的网络带宽，以便实时更新网络带宽。
第三种可能情况下，第一MEC建立的路由表可以包括如下内容：
本地MEC设备：
IP地址：xx.xx.xx.xx
地理位置：广州(经度/纬度)
APP列表：
APP1：
CPU占用率
历史任务处理时间
APP2：
CPU占用率
历史任务处理时间
本地文件列表：
目录：
本地文件1：文件名称、文件大小
本地文件2：文件名称、文件大小
邻居MEC设备1：
IP地址：xx.xx.xx.xx
地理位置：深圳(经度/纬度)
与第一MEC设备之间的距离：xxx公里
网络带宽
APP列表：
APP1：
CPU占用率
历史任务处理时间
APP3：
CPU占用率
历史任务处理时间
本地文件列表：
目录：
本地文件1：文件名称、文件大小
本地文件2：文件名称、文件大小
邻居MEC设备2：
…
邻居MEC设备3：
…
在本申请的一个可选实施例中，步骤203的实现过程包括：第一MEC设备向邻居MEC设备发送应用信息获取请求；第一MEC设备接收邻居MEC设备基于该应用信息获取请求回复的该邻居MEC设备的应用信息。
可选地，第一MEC设备周期性地向邻居MEC设备发送应用信息获取请求，以实时更新邻居MEC设备的应用信息。第一MEC设备存储邻居MEC设备的最新的应用信息。
在本申请的另一个可选实施例中，步骤203的实现过程包括：第一MEC设备接收邻居MEC设备主动广播发送的应用信息。
可选地，第一MEC设备广播发送第一MEC设备的应用信息。当MEC设备之间通过无线连接，第一MEC设备广播发送第一MEC设备的应用信息，是指发送包含第一MEC设备的应用信息的报文且该报文是组寻址(group addressed)报文，该报文的目的地址可以是组播地址或广播地址，例如该报文的目的地址包括多个邻居MEC设备的IP地址。
可选地，通信网络中的各个MEC设备周期性地广播发送自身的应用消息，或者每当自身的应用信息发生更新后，广播发送更新后的应用信息。每个MEC设备存储邻居MEC设备的最新的应用信息。
步骤204、第一MEC设备接收来自UE的任务请求。
该任务请求包括UE所请求访问的目标应用服务的应用服务标识以及该任务请求对应的SLA信息，该SLA信息用于确定响应该任务请求所满足的SLA条件。
可选地，任务请求对应的SLA信息指示该任务请求的最大允许响应时延，则响应该任务请求所满足的SLA条件，包括：响应该任务请求的总时延不大于该最大允许响应时延。任务请求的最大允许响应时延，指从UE发送该任务请求起至该任务请求被处理完成的整个过程所允许耗费的最大时长。该任务请求被处理完成可以是UE接收到对该任务请求的响应。任务请求对应的SLA信息可以包括该任务请求的最大允许响应时延以及UE发送该任务请求的发送时刻。
可选地，该任务请求还包括待处理业务的相关信息，例如可以包括请求文件信息，该请求文件信息包括待请求文件的文件标识。待请求文件指UE请求下载的文件。当然，该任务请求还可以包括用于标识该任务请求的序列号等，本申请实施例对此不做限定。
在步骤204中，第一MEC设备接收来自UE的任务请求，可以是第一MEC设备接收UE直接发送的任务请求；或者也可以是第一MEC设备接收其它MEC设备转发的来自UE的任务请求。响应于UE不在第一MEC设备对应的服务区域内，第一MEC设备可以接收第三MEC设备转发的任务请求，第一MEC设备为第三MEC设备的邻居MEC设备。或者，响应于UE在第一MEC设备对应的服务区域内，第一MEC设备可以接收UE发送的任务请求或者第四MEC设备转发的任务请求，第一MEC设备为第四MEC设备的邻居MEC设备。
本申请实施例中，最终处理来自UE的任务请求的MEC设备对应的服务区域可以覆盖UE所在位置，也可以不覆盖UE所在位置，也即是，MEC设备在自身不满足响应任务请求应满足的SLA条件时，不受限于UE的地理位置，可以在邻居MEC设备中灵活选择处理该任务请求的MEC设备，使得任务处理的灵活性较高。
可选地，第一MEC设备还可以接收到与任务请求一起发送的转发次数，该转发次数为任务请求的已转发次数。例如，第一MEC设备接收到的与任务请求一起发送的转发次数为0，表示该任务请求是由UE直接发送给第一MEC设备的。又例如，第一MEC设备接收到的与任务请求一起发送的转发次数为1，表示该任务请求是经由一个MEC设备转发给第一MEC设备的。
步骤205、第一MEC设备判断该第一MEC设备中部署的目标应用服务的第一实例是否满足SLA条件；响应于第一MEC设备中部署的目标应用服务的第一实例不满足SLA条件，执行步骤206；或者，响应于第一MEC设备中部署的目标应用服务的第一实例满足SLA条件，执行步骤208。
可选地，任务请求中包括该任务请求的最大允许响应时延以及UE发送该任务请求的发送时刻。步骤205的实现过程包括：第一MEC设备计算该任务请求的剩余允许响应时延。其中，剩余允许响应时延＝最大允许响应时延-(当前时刻-UE发送该任务请求的发送时刻)。该当前时刻为计算该剩余允许响应时延的时刻，例如可以是判断第一实例是否满足SLA条件的时刻。响应于第一实例的历史任务处理时间小于该剩余允许响应时延，第一MEC设备确定第 一实例满足SLA条件；或者，响应于第一实例的历史任务处理时间大于或等于该第一剩余允许响应时延，第一MEC设备确定第一实例不满足SLA条件。
第一MEC设备在接收到来自UE的任务请求后，可以先计算该任务请求的剩余允许响应时延，其中，剩余允许响应时延＝最大允许响应时延-(当前时刻-UE发送该任务请求的发送时刻)，该当前时刻为计算该剩余允许响应时延的时刻，例如可以是第一MEC设备接收到该任务请求的接收时刻。然后根据该剩余允许响应时延，将该任务请求存储在待处理任务队列中，其中，待处理任务队列中的任务请求按照剩余允许响应时延由小至大的顺序从头部往尾部依次存储，即剩余允许响应时延小的任务请求放在待处理任务队列的前面，剩余允许响应时延大的任务请求放在待处理任务队列的后面。第一MEC设备从待处理任务队列的头部依次取出任务请求进行处理。
本申请实施例中，MEC设备可以优先处理剩余允许响应时延较小的任务请求，以使得尽量满足各个任务请求的时延需求，提高MEC设备处理任务的可靠性。
可选地，响应于第一MEC设备中部署的目标应用服务的第一实例不满足SLA条件，且第一MEC设备接收到的任务请求的已转发次数小于转发次数阈值，执行步骤206。当第一MEC设备中部署的目标应用服务的第一实例不满足SLA条件，且第一MEC设备接收到的任务请求的已转发次数达到转发次数阈值时，第一MEC设备确定当前无法满足该任务请求的SLA条件，并返回针对该任务请求的处理失败信息。例如，转发次数阈值可以取值为3、4或5。
本申请实施例中，通过设置转发次数阈值以限制任务请求的转发次数，避免任务请求被过多转发而额外造成任务请求和/或针对该任务请求的响应的较大传输时延，同时也限制了解析该任务请求的MEC设备的数量，避免占用过多MEC设备的处理资源。
步骤206、第一MEC设备在第一MEC设备的邻居MEC设备中确定第二MEC设备。
该第二MEC设备中部署有目标应用服务的第二实例且第二实例满足SLA条件。
结合参考步骤203中第一种可能情况下的路由表，步骤206的实现过程包括：第一MEC设备根据邻居MEC设备的应用信息，在邻居MEC设备中确定目标邻居MEC设备。第一MEC设备在目标邻居MEC设备中确定第二MEC设备。其中，目标邻居MEC设备满足：部署有目标应用服务的实例，且部署的目标应用服务的实例的历史任务处理时间小于剩余允许响应时延。该剩余允许响应时延基于任务请求的最大允许响应时延得到。例如，剩余允许响应时延＝最大允许响应时延-(当前时刻-UE发送该任务请求的发送时刻)。该当前时刻为计算该剩余允许响应时延的时刻。
在这种可能情况下，第一MEC设备在目标邻居MEC设备中确定第二MEC设备的实现过程，可以包括：响应于目标邻居MEC设备的数量大于1，第一MEC设备根据多个目标邻居MEC设备中部署的目标应用服务的实例的历史任务处理时间、多个目标邻居MEC设备中部署的目标应用服务的实例的负载和/或多个目标邻居MEC设备与第一MEC设备之间的距离，在该多个目标邻居MEC设备中确定第二MEC设备。例如，第一MEC设备可以将多个目标邻居MEC设备中部署的目标应用服务的实例的历史任务处理时间最短的目标邻居MEC设备确定为第二MEC设备。
或者，第一MEC设备也可以将任一目标邻居MEC设备确定为第二MEC设备。
可选地，当第一MEC设备在邻居MEC设备中找不到满足条件的目标邻居MEC设备时， 第一MEC设备确定当前无法满足该任务请求的SLA条件，并返回针对该任务请求的处理失败信息。
本申请实施例中，基于MEC设备对任务请求的处理时间的考虑，确定任务请求的路由对象，该路由对象满足响应任务请求的SLA条件的概率较高，从而提高任务处理的可靠性。
结合参考步骤203中第二种可能情况下的路由表，步骤206的实现过程包括：第一MEC设备根据邻居MEC设备的应用信息以及邻居MEC设备与第一MEC设备之间的网络响应时延，在邻居MEC设备中确定目标邻居MEC设备。第一MEC设备在目标邻居MEC设备中确定第二MEC设备。其中，目标邻居MEC设备满足：部署有目标应用服务的实例，且与第一MEC设备之间的网络响应时延跟部署的目标应用服务的实例的历史任务处理时间之和小于剩余允许响应时延。
在这种可能情况下，第一MEC设备在目标邻居MEC设备中确定第二MEC设备的实现过程，可以包括：响应于目标邻居MEC设备的数量大于1，第一MEC设备将多个目标邻居MEC设备中，与第一MEC设备之间的网络响应时延跟部署的目标应用服务的实例的历史任务处理时间之和最小的目标邻居MEC设备确定为第二MEC设备。
或者，第一MEC设备也可以将任一目标邻居MEC设备确定为第二MEC设备。
可选地，当第一MEC设备在邻居MEC设备中找不到满足条件的目标邻居MEC设备时，第一MEC设备确定当前无法满足该任务请求的SLA条件，并返回针对该任务请求的处理失败信息。
本申请实施例中，基于MEC设备对任务请求的处理时间以及MEC设备之间的网络响应时延的考虑，确定任务请求的路由对象，该路由对象能够满足响应任务请求的SLA条件的概率相较于只考虑MEC设备对任务请求的处理时间的方式更高，从而提高任务处理的可靠性。对于游戏、虚拟现实(virtual reality，VR)和增强现实(augmented reality，AR)等对时延非常敏感的计算密集型边缘应用，综合考虑MEC设备对任务请求的处理时间以及MEC设备之间的网络响应时延来调度任务请求，可以提供确定性的任务响应时延，提高用户体验。
结合参考步骤203中第三种可能情况下的路由表，任务请求包括请求文件信息，该请求文件信息包括待请求文件的文件标识，待请求文件的文件标识可以是目录、文件名称列表或文件描述。步骤206的实现过程包括：第一MEC设备根据邻居MEC设备的应用信息、请求文件信息以及邻居MEC设备与第一MEC设备之间的网络带宽，在邻居MEC设备中确定目标邻居MEC设备。第一MEC设备在目标邻居MEC设备中确定第二MEC设备。其中，目标邻居MEC设备满足：部署有目标应用服务的实例，本地文件包括待请求文件，且向第一MEC设备传输待请求文件的文件传输时间跟部署的目标应用服务的实例的历史任务处理时间之和小于剩余允许响应时延，该文件传输时间基于待请求文件的文件大小以及邻居MEC设备与第一MEC设备之间的网络带宽确定。邻居MEC设备向第一MEC设备传输待请求文件的文件传输时间等于所有待请求文件的文件大小除以该邻居MEC设备与第一MEC设备之间的网络带宽。
第一MEC设备在邻居MEC设备中确定目标邻居MEC设备的过程包括：首先，第一MEC设备基于步骤203中第三种可能情况下的路由表，获取包含所有待请求文件的邻居MEC设备以及各个待请求文件的大小。然后，第一MEC设备计算每个请求文件的文件传输时间，并累加所有请求文件的文件传输时间。最后，第一MEC设备将所有请求文件的文件传输时间 与部署的目标应用服务的实例的历史任务处理时间之和小于剩余允许响应时延的邻居MEC设备确定为目标邻居MEC设备。
在这种可能情况下，第一MEC设备在目标邻居MEC设备中确定第二MEC设备的实现过程，可以包括：响应于目标邻居MEC设备的数量大于1，第一MEC设备将多个目标邻居MEC设备中，向第一MEC设备传输待请求文件的文件传输时间跟部署的目标应用服务的实例的历史任务处理时间之和最小的目标邻居MEC设备确定为第二MEC设备。
或者，第一MEC设备也可以将任一目标邻居MEC设备确定为第二MEC设备。
可选地，当第一MEC设备在邻居MEC设备中找不到满足条件的目标邻居MEC设备时，第一MEC设备确定当前无法满足该任务请求的SLA条件，并返回针对该任务请求的处理失败信息。
本申请实施例中，基于MEC设备对任务请求的处理时间以及MEC设备之间传输待请求文件的文件传输时间，确定任务请求的路由对象，该路由对象能够满足响应任务请求的SLA条件的概率相较于只考虑MEC设备对任务请求的处理时间的方式更高，从而提高任务处理的可靠性。当任务请求包括待上传文件时，还可以考虑MEC设备之间传输任务请求的时间来确定任务请求的路由对象。对于人工智能(artificial intelligence，AI)处理等对时延非常敏感的数据密集型边缘应用，综合考虑MEC设备对任务请求的处理时间、UE的待请求文件的文件分布、MEC设备之间传输待请求文件的文件传输时间以及MEC设备之间传输任务请求的时间来调度任务请求，可以提供确定性的任务响应时延，提高用户体验。
步骤207、第一MEC设备向第二MEC设备转发任务请求。
可选地，第一MEC设备还可以将任务请求的转发次数加1，并向第二MEC设备一起发送该任务请求和更新后的转发次数。
第二MEC设备接收到任务请求后，执行与第一MEC设备接收到任务请求后相同的操作，本申请实施例在此不再赘述。响应于第二MEC设备中部署的目标应用服务的第二实例满足SLA条件，第二MEC设备在通过第二实例处理任务请求后，还可以向第一MEC设备发送针对该任务请求的响应。第一MEC设备接收到第二MEC设备发送的针对该任务请求的响应后，向目标设备转发针对该任务请求的响应。该目标设备为向第一MEC设备发送该任务请求的设备。该目标设备可以是UE或者是其它MEC设备。
本申请实施例中，针对UE发送的任务请求的响应的路由路径与该任务请求的路由路径相反，UE向接入的MEC设备发送任务请求，最终也由UE接入的MEC设备向UE发送针对该任务请求的响应，MEC设备处理任务请求的整个过程UE是无感知的，因此无需改变UE访问应用服务的方式，对UE的兼容性较好。
步骤208、第一MEC设备通过第一实例处理任务请求。
第一MEC设备通过第一实例处理完任务请求后，还可以向目标设备发送针对该任务请求的响应。当该任务请求由UE直接发送给第一MEC设备时，第一MEC设备直接向UE发送针对该任务请求的响应。当该任务请求由其它MEC设备转发给第一MEC设备时，第一MEC设备向给其发送任务请求的MEC设备发送针对该任务请求的响应，即针对该任务请求的响应的路由路径与该任务请求的路由路径相反，最终由UE接入的MEC设备向UE发送针对该任务请求的响应。
本申请实施例提供的任务处理方法可以应用于多个场景，包括但不限于：AR、VR和游 戏，例如基于KubeEdge/智能边缘平台(intelligent edge fabric，IEF)构建的云游戏服务；互动直播，例如基于KubeEdge/IEF构建的直播业务；AI处理，例如基于KubeEdge/IEF构建的人脸识别和目标跟踪等业务。
本申请实施例提供的任务处理方法的步骤先后顺序可以进行适当调整，步骤也可以根据情况进行相应增减。任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化的方法，都应涵盖在本申请的保护范围之内，因此不再赘述。
综上所述，在本申请实施例提供的任务处理方法中，在端边复杂网络、多层次边缘的系统环境下，综合考虑MEC设备的状态、负载和任务处理时间等因素进行流量多次分发的路由选择，把UE的访问流量分发到合适的其它MEC设备，提升端边协同任务处理的可靠性和可用性。在区域的UE突然大幅增加或局部MEC设备负载过高等场景下，通过多次流量分发，把流量分发到可用的其它MEC设备，保证业务的可用性和持续性。对于时延敏感的业务，综合考虑最大允许响应时延、MEC设备之间的网络响应时延、任务处理时间、文件传输时间等因素进行路由选择，提供确定性的任务响应时延，提升用户体验。对于计算密集型和数据密集的应用场景，通过多次流量分发，满足应用的时延要求，保证应用的用户体验。该方法通过MEC设备之间相互感知实现，而UE无感知，因此无需改变UE访问应用服务的方式，对UE的兼容性较好。另外，MEC设备在自身不满足响应任务请求应满足的SLA条件时，不受限于UE的地理位置，可以在邻居MEC设备中灵活选择处理该任务请求的MEC设备，使得任务处理的灵活性较高。
图3是本申请实施例提供的一种任务处理装置的结构示意图。该装置应用于第一MEC设备。如图3所示，装置30包括：
接收模块301，用于接收来自UE的任务请求，任务请求包括UE所请求访问的目标应用服务的应用服务标识以及任务请求对应的服务等级协议SLA信息，SLA信息用于确定响应任务请求所满足的SLA条件。
第一确定模块302，用于响应于第一MEC设备中部署的目标应用服务的第一实例不满足SLA条件，在第一MEC设备的邻居MEC设备中确定第二MEC设备，第二MEC设备中部署有目标应用服务的第二实例且第二实例满足SLA条件。
发送模块303，用于向第二MEC设备转发任务请求。
可选地，任务请求对应的SLA信息指示任务请求的最大允许响应时延，响应任务请求所满足的SLA条件，包括：响应任务请求的总时延不大于最大允许响应时延。
可选地，如图4所示，装置30还包括：
获取模块304，用于获取邻居MEC设备的应用信息，该应用信息包括部署的应用服务的实例的历史任务处理时间。第一确定模块302，包括：第一确定子模块3021，用于根据邻居MEC设备的应用信息，在邻居MEC设备中确定目标邻居MEC设备，其中，目标邻居MEC设备满足：部署有目标应用服务的实例，且部署的目标应用服务的实例的历史任务处理时间小于剩余允许响应时延，剩余允许响应时延基于最大允许响应时延得到。第二确定子模块3022，用于在目标邻居MEC设备中确定第二MEC设备。
可选地，应用信息还包括部署的应用服务的实例的负载，第二确定子模块3022，用于：响应于目标邻居MEC设备的数量大于1，根据多个目标邻居MEC设备中部署的目标应用服 务的实例的历史任务处理时间、多个目标邻居MEC设备中部署的目标应用服务的实例的负载和/或多个目标邻居MEC设备与第一MEC设备之间的距离，在多个目标邻居MEC设备中确定第二MEC设备。
可选地，第一确定子模块3021，用于：根据邻居MEC设备的应用信息以及邻居MEC设备与第一MEC设备之间的网络响应时延，在邻居MEC设备中确定目标邻居MEC设备。其中，目标邻居MEC设备满足：部署有目标应用服务的实例，且与第一MEC设备之间的网络响应时延跟部署的目标应用服务的实例的历史任务处理时间之和小于剩余允许响应时延。
相应地，第二确定子模块3022，用于：响应于目标邻居MEC设备的数量大于1，第一MEC设备将多个目标邻居MEC设备中，与第一MEC设备之间的网络响应时延跟部署的目标应用服务的实例的历史任务处理时间之和最小的目标邻居MEC设备确定为第二MEC设备。
可选地，任务请求还包括请求文件信息，请求文件信息包括待请求文件的文件标识，应用信息还包括文件信息，文件信息包括本地文件的文件标识和本地文件的文件大小。第一确定子模块3021，用于：根据邻居MEC设备的应用信息、请求文件信息以及邻居MEC设备与第一MEC设备之间的网络带宽，在邻居MEC设备中确定目标邻居MEC设备。其中，目标邻居MEC设备满足：部署有目标应用服务的实例，本地文件包括待请求文件，且向第一MEC设备传输待请求文件的文件传输时间跟部署的目标应用服务的实例的历史任务处理时间之和小于剩余允许响应时延，文件传输时间基于待请求文件的文件大小以及邻居MEC设备与第一MEC设备之间的网络带宽确定。
相应地，第二确定子模块3022，用于：响应于目标邻居MEC设备的数量大于1，将多个目标邻居MEC设备中，向第一MEC设备传输待请求文件的文件传输时间跟部署的目标应用服务的实例的历史任务处理时间之和最小的目标邻居MEC设备确定为第二MEC设备。
可选地，发送模块303，还用于广播发送第一MEC设备的应用信息。
可选地，第一确定模块302，用于：响应于第一实例不满足SLA条件，且第一MEC设备接收到的任务请求的已转发次数小于转发次数阈值，第一MEC设备在邻居MEC设备中确定第二MEC设备。
可选地，如图5所示，装置30还包括第二确定模块305。接收模块301，还用于接收应用控制器发送的MEC设备列表，MEC设备列表包括多个MEC设备的通信地址信息和地理位置信息。第二确定模块305，用于将多个MEC设备中，与第一MEC设备之间的距离小于距离阈值的MEC设备确定为第一MEC设备的邻居MEC设备。
可选地，接收模块301，还用于在向第二MEC设备转发任务请求之后，接收第二MEC设备发送的针对任务请求的响应。发送模块303，还用于向目标设备转发针对任务请求的响应。目标设备为向第一MEC设备发送任务请求的设备。
可选地，接收模块301，用于：响应于UE不在第一MEC设备对应的服务区域内，接收第三MEC设备转发的任务请求。或者，响应于UE在第一MEC设备对应的服务区域内，接收UE发送的任务请求或者第四MEC设备转发的任务请求。
可选地，如图6所示，装置30还包括：处理模块306，用于响应于第一实例满足SLA条件，通过第一实例处理任务请求。
综上所述，本申请实施例提供的任务处理装置，通过接收模块接收到UE发送的任务请 求后，首先判断自身是否满足响应该任务请求应满足的SLA条件。在通过第一确定模块确定自身满足该SLA条件后，处理该任务请求；在通过第一确定模块确定自身不满足该SLA条件后，通过发送模块将任务请求调度给满足该SLA条件的邻居MEC设备，由满足该SLA条件的邻居MEC设备处理该任务请求。通过对任务请求的合理调度，使得对该任务请求的响应能够满足SLA条件，提高了任务处理的可靠性和可用性。在端边复杂网络、多层次边缘的系统环境下，综合考虑MEC设备的状态、负载和任务处理时间等因素进行流量多次分发的路由选择，把UE的访问流量分发到合适的其它MEC设备，提升端边协同任务处理的可靠性和可用性。在区域的UE突然大幅增加或局部MEC设备负载过高等场景下，通过多次流量分发，把流量分发到可用的其它MEC设备，保证业务的可用性和持续性。对于时延敏感的业务，综合考虑最大允许响应时延、MEC设备之间的网络响应时延、任务处理时间、文件传输时间等因素进行路由选择，提供确定性的任务响应时延，提升用户体验。对于计算密集型和数据密集的应用场景，通过多次流量分发，满足应用的时延要求，保证应用的用户体验。该方法通过MEC设备之间相互感知实现，而UE无感知，因此无需改变UE访问应用服务的方式，对UE的兼容性较好。另外，MEC设备在自身不满足响应任务请求应满足的SLA条件时，不受限于UE的地理位置，可以在邻居MEC设备中灵活选择处理该任务请求的MEC设备，使得任务处理的灵活性较高。
关于上述实施例中的装置，其中各个模块执行操作的具体方式已经在有关该方法的实施例中进行了详细描述，此处将不做详细阐述说明。
图7是本申请实施例提供的一种MEC设备的框图。如图7所示，MEC设备70包括：处理器701和存储器702。
存储器702，用于存储计算机程序，所述计算机程序包括程序指令；
处理器701，用于调用所述计算机程序，实现如上述方法实施例所述的任务处理方法。
可选地，该MEC设备70还包括通信总线703和通信接口704。
其中，处理器701包括一个或者一个以上处理核心，处理器701通过运行计算机程序，执行各种功能应用以及数据处理。
存储器702可用于存储计算机程序。可选地，存储器可存储操作系统和至少一个功能所需的应用程序单元。操作系统可以是实时操作系统(Real Time eXecutive，RTX)、LINUX、UNIX、WINDOWS或OS X之类的操作系统。
通信接口704可以为多个，通信接口704用于与其它设备进行通信。例如在本申请实施例中，第一MEC设备的通信接口704可以用于向第二MEC设备发送任务请求。
存储器702与通信接口704分别通过通信总线703与处理器701连接。
本申请实施例还提供了一种任务处理系统，该系统包括：多个MEC设备。该MEC设备包括如图3至图6任一所示的装置或为如图7所示的MEC设备。
本申请实施例还提供了一种计算机存储介质，所述计算机存储介质上存储有指令，当所述指令被处理器执行时，实现如上述方法实施例所述的任务处理方法。
本领域普通技术人员可以理解实现上述实施例的全部或部分步骤可以通过硬件来完成，也可以通过程序来指令相关的硬件完成，所述的程序可以存储于一种计算机可读存储介质中，上述提到的存储介质可以是只读存储器，磁盘或光盘等。
在本申请实施例中，术语“第一”、“第二”和“第三”仅用于描述目的，而不能理解为指示或暗示相对重要性。
本申请中术语“和/或”，仅仅是一种描述关联对象的关联关系，表示可以存在三种关系，例如，A和/或B，可以表示：单独存在A，同时存在A和B，单独存在B这三种情况。另外，本文中字符“/”，一般表示前后关联对象是一种“或”的关系。
以上所述仅为本申请的可选实施例，并不用以限制本申请，凡在本申请的构思和原则之内，所作的任何修改、等同替换、改进等，均应包含在本申请的保护范围之内。

Claims (29)

  1. 一种任务处理方法，其特征在于，所述方法包括：
    第一多接入边缘计算MEC设备接收来自用户设备UE的任务请求，所述任务请求包括所述UE所请求访问的目标应用服务的应用服务标识以及所述任务请求对应的服务等级协议SLA信息，所述SLA信息用于确定响应所述任务请求所满足的SLA条件；
    响应于第一MEC设备中部署的所述目标应用服务的第一实例不满足所述SLA条件，所述第一MEC设备在所述第一MEC设备的邻居MEC设备中确定第二MEC设备，所述第二MEC设备中部署有所述目标应用服务的第二实例且所述第二实例满足所述SLA条件；
    所述第一MEC设备向所述第二MEC设备转发所述任务请求。
  2. 根据权利要求1所述的方法，其特征在于，所述任务请求对应的SLA信息指示所述任务请求的最大允许响应时延，所述响应所述任务请求所满足的SLA条件，包括：响应所述任务请求的总时延不大于所述最大允许响应时延。
  3. 根据权利要求2所述的方法，其特征在于，所述方法还包括：
    所述第一MEC设备获取所述邻居MEC设备的应用信息，所述应用信息包括部署的应用服务的实例的历史任务处理时间；
    所述第一MEC设备在所述第一MEC设备的邻居MEC设备中确定第二MEC设备，包括：
    所述第一MEC设备根据所述邻居MEC设备的应用信息，在所述邻居MEC设备中确定目标邻居MEC设备，其中，所述目标邻居MEC设备满足：部署有所述目标应用服务的实例，且部署的所述目标应用服务的实例的历史任务处理时间小于剩余允许响应时延，所述剩余允许响应时延基于所述最大允许响应时延得到；
    所述第一MEC设备在所述目标邻居MEC设备中确定所述第二MEC设备。
  4. 根据权利要求3所述的方法，其特征在于，所述应用信息还包括部署的应用服务的实例的负载，所述第一MEC设备在所述目标邻居MEC设备中确定所述第二MEC设备，包括：
    响应于所述目标邻居MEC设备的数量大于1，所述第一MEC设备根据多个所述目标邻居MEC设备中部署的所述目标应用服务的实例的历史任务处理时间、多个所述目标邻居MEC设备中部署的所述目标应用服务的实例的负载和/或多个所述目标邻居MEC设备与所述第一MEC设备之间的距离，在多个所述目标邻居MEC设备中确定所述第二MEC设备。
  5. 根据权利要求3所述的方法，其特征在于，所述第一MEC设备根据所述邻居MEC设备的应用信息，在所述邻居MEC设备中确定目标邻居MEC设备，包括：
    所述第一MEC设备根据所述邻居MEC设备的应用信息以及所述邻居MEC设备与所述第一MEC设备之间的网络响应时延，在所述邻居MEC设备中确定所述目标邻居MEC设备；
    其中，所述目标邻居MEC设备满足：部署有所述目标应用服务的实例，且与所述第一 MEC设备之间的网络响应时延跟部署的所述目标应用服务的实例的历史任务处理时间之和小于所述剩余允许响应时延。
  6. 根据权利要求5所述的方法，其特征在于，所述第一MEC设备在所述目标邻居MEC设备中确定所述第二MEC设备，包括：
    响应于所述目标邻居MEC设备的数量大于1，所述第一MEC设备将多个所述目标邻居MEC设备中，与所述第一MEC设备之间的网络响应时延跟部署的所述目标应用服务的实例的历史任务处理时间之和最小的目标邻居MEC设备确定为所述第二MEC设备。
  7. 根据权利要求3所述的方法，其特征在于，所述任务请求还包括请求文件信息，所述请求文件信息包括待请求文件的文件标识，所述应用信息还包括文件信息，所述文件信息包括本地文件的文件标识和所述本地文件的文件大小；所述第一MEC设备根据所述邻居MEC设备的应用信息，在所述邻居MEC设备中确定目标邻居MEC设备，包括：
    所述第一MEC设备根据所述邻居MEC设备的应用信息、所述请求文件信息以及所述邻居MEC设备与所述第一MEC设备之间的网络带宽，在所述邻居MEC设备中确定所述目标邻居MEC设备；
    其中，所述目标邻居MEC设备满足：部署有所述目标应用服务的实例，本地文件包括所述待请求文件，且向所述第一MEC设备传输所述待请求文件的文件传输时间跟部署的所述目标应用服务的实例的历史任务处理时间之和小于所述剩余允许响应时延，所述文件传输时间基于所述待请求文件的文件大小以及所述邻居MEC设备与所述第一MEC设备之间的网络带宽确定。
  8. 根据权利要求7所述的方法，其特征在于，所述第一MEC设备在所述目标邻居MEC设备中确定所述第二MEC设备，包括：
    响应于所述目标邻居MEC设备的数量大于1，所述第一MEC设备将多个所述目标邻居MEC设备中，向所述第一MEC设备传输所述待请求文件的文件传输时间跟部署的所述目标应用服务的实例的历史任务处理时间之和最小的目标邻居MEC设备确定为所述第二MEC设备。
  9. 根据权利要求3至8任一所述的方法，其特征在于，所述方法还包括：
    所述第一MEC设备广播发送所述第一MEC设备的应用信息。
  10. 根据权利要求1至9任一所述的方法，其特征在于，所述响应于第一MEC设备中部署的所述目标应用服务的第一实例不满足所述SLA条件，所述第一MEC设备在所述第一MEC设备的邻居MEC设备中确定第二MEC设备，包括：
    响应于所述第一实例不满足所述SLA条件，且所述第一MEC设备接收到的所述任务请求的已转发次数小于转发次数阈值，所述第一MEC设备在所述邻居MEC设备中确定所述第二MEC设备。
  11. 根据权利要求1至10任一所述的方法，其特征在于，所述方法还包括：
    所述第一MEC设备接收应用控制器发送的MEC设备列表，所述MEC设备列表包括多个MEC设备的通信地址信息和地理位置信息；
    所述第一MEC设备将所述多个MEC设备中，与所述第一MEC设备之间的距离小于距离阈值的MEC设备确定为所述第一MEC设备的邻居MEC设备。
  12. 根据权利要求1至11任一所述的方法，其特征在于，在所述第一MEC设备向所述第二MEC设备转发所述任务请求之后，所述方法还包括：
    所述第一MEC设备接收所述第二MEC设备发送的针对所述任务请求的响应；
    所述第一MEC设备向目标设备转发所述针对所述任务请求的响应，所述目标设备为向所述第一MEC设备发送所述任务请求的设备。
  13. 根据权利要求1至12任一所述的方法，其特征在于，所述第一MEC设备接收来自UE的任务请求，包括：
    响应于所述UE不在所述第一MEC设备对应的服务区域内，所述第一MEC设备接收第三MEC设备转发的所述任务请求；或者，
    响应于所述UE在所述第一MEC设备对应的服务区域内，所述第一MEC设备接收所述UE发送的所述任务请求或者第四MEC设备转发的所述任务请求。
  14. 一种任务处理装置，其特征在于，应用于第一多接入边缘计算MEC设备，所述装置包括：
    接收模块，用于接收来自用户设备UE的任务请求，所述任务请求包括所述UE所请求访问的目标应用服务的应用服务标识以及所述任务请求对应的服务等级协议SLA信息，所述SLA信息用于确定响应所述任务请求所满足的SLA条件；
    第一确定模块，用于响应于第一MEC设备中部署的所述目标应用服务的第一实例不满足所述SLA条件，在所述第一MEC设备的邻居MEC设备中确定第二MEC设备，所述第二MEC设备中部署有所述目标应用服务的第二实例且所述第二实例满足所述SLA条件；
    发送模块，用于向所述第二MEC设备转发所述任务请求。
  15. 根据权利要求14所述的装置，其特征在于，所述任务请求对应的SLA信息指示所述任务请求的最大允许响应时延，所述响应所述任务请求所满足的SLA条件，包括：响应所述任务请求的总时延不大于所述最大允许响应时延。
  16. 根据权利要求15所述的装置，其特征在于，所述装置还包括：
    获取模块，用于获取所述邻居MEC设备的应用信息，所述应用信息包括部署的应用服务的实例的历史任务处理时间；
    所述第一确定模块，包括：
    第一确定子模块，用于根据所述邻居MEC设备的应用信息，在所述邻居MEC设备中确定目标邻居MEC设备，其中，所述目标邻居MEC设备满足：部署有所述目标应用服务的实 例，且部署的所述目标应用服务的实例的历史任务处理时间小于剩余允许响应时延，所述剩余允许响应时延基于所述最大允许响应时延得到；
    第二确定子模块，用于在所述目标邻居MEC设备中确定所述第二MEC设备。
  17. 根据权利要求16所述的装置，其特征在于，所述应用信息还包括部署的应用服务的实例的负载，所述第二确定子模块，用于：
    响应于所述目标邻居MEC设备的数量大于1，根据多个所述目标邻居MEC设备中部署的所述目标应用服务的实例的历史任务处理时间、多个所述目标邻居MEC设备中部署的所述目标应用服务的实例的负载和/或多个所述目标邻居MEC设备与所述第一MEC设备之间的距离，在多个所述目标邻居MEC设备中确定所述第二MEC设备。
  18. 根据权利要求16所述的装置，其特征在于，所述第一确定子模块，用于：
    根据所述邻居MEC设备的应用信息以及所述邻居MEC设备与所述第一MEC设备之间的网络响应时延，在所述邻居MEC设备中确定所述目标邻居MEC设备；
    其中，所述目标邻居MEC设备满足：部署有所述目标应用服务的实例，且与所述第一MEC设备之间的网络响应时延跟部署的所述目标应用服务的实例的历史任务处理时间之和小于所述剩余允许响应时延。
  19. 根据权利要求18所述的装置，其特征在于，所述第二确定子模块，用于：
    响应于所述目标邻居MEC设备的数量大于1，所述第一MEC设备将多个所述目标邻居MEC设备中，与所述第一MEC设备之间的网络响应时延跟部署的所述目标应用服务的实例的历史任务处理时间之和最小的目标邻居MEC设备确定为所述第二MEC设备。
  20. 根据权利要求16所述的装置，其特征在于，所述任务请求还包括请求文件信息，所述请求文件信息包括待请求文件的文件标识，所述应用信息还包括文件信息，所述文件信息包括本地文件的文件标识和所述本地文件的文件大小；所述第一确定子模块，用于：
    根据所述邻居MEC设备的应用信息、所述请求文件信息以及所述邻居MEC设备与所述第一MEC设备之间的网络带宽，在所述邻居MEC设备中确定所述目标邻居MEC设备；
    其中，所述目标邻居MEC设备满足：部署有所述目标应用服务的实例，本地文件包括所述待请求文件，且向所述第一MEC设备传输所述待请求文件的文件传输时间跟部署的所述目标应用服务的实例的历史任务处理时间之和小于所述剩余允许响应时延，所述文件传输时间基于所述待请求文件的文件大小以及所述邻居MEC设备与所述第一MEC设备之间的网络带宽确定。
  21. 根据权利要求20所述的装置，其特征在于，所述第二确定子模块，用于：
    响应于所述目标邻居MEC设备的数量大于1，将多个所述目标邻居MEC设备中，向所述第一MEC设备传输所述待请求文件的文件传输时间跟部署的所述目标应用服务的实例的历史任务处理时间之和最小的目标邻居MEC设备确定为所述第二MEC设备。
  22. 根据权利要求16至21任一所述的装置，其特征在于，
    所述发送模块，还用于广播发送所述第一MEC设备的应用信息。
  23. 根据权利要求14至22任一所述的装置，其特征在于，所述第一确定模块，用于：
    响应于所述第一实例不满足所述SLA条件，且所述第一MEC设备接收到的所述任务请求的已转发次数小于转发次数阈值，所述第一MEC设备在所述邻居MEC设备中确定所述第二MEC设备。
  24. 根据权利要求14至23任一所述的装置，其特征在于，所述装置还包括第二确定模块；
    所述接收模块，还用于接收应用控制器发送的MEC设备列表，所述MEC设备列表包括多个MEC设备的通信地址信息和地理位置信息；
    所述第二确定模块，用于将所述多个MEC设备中，与所述第一MEC设备之间的距离小于距离阈值的MEC设备确定为所述第一MEC设备的邻居MEC设备。
  25. 根据权利要求14至24任一所述的装置，其特征在于，
    所述接收模块，还用于在向所述第二MEC设备转发所述任务请求之后，接收所述第二MEC设备发送的针对所述任务请求的响应；
    所述发送模块，还用于向目标设备转发所述针对所述任务请求的响应，所述目标设备为向所述第一MEC设备发送所述任务请求的设备。
  26. 根据权利要求14至25任一所述的装置，其特征在于，所述接收模块，用于：
    响应于所述UE不在所述第一MEC设备对应的服务区域内，接收第三MEC设备转发的所述任务请求；或者，
    响应于所述UE在所述第一MEC设备对应的服务区域内，接收所述UE发送的所述任务请求或者第四MEC设备转发的所述任务请求。
  27. 一种多接入边缘计算MEC设备，其特征在于，包括：处理器和存储器；
    所述存储器，用于存储计算机程序，所述计算机程序包括程序指令；
    所述处理器，用于调用所述计算机程序，实现如权利要求1至13任一所述的任务处理方法。
  28. 一种任务处理系统，其特征在于，所述系统包括：多个多接入边缘计算MEC设备，所述MEC设备包括如权利要求14至26任一所述的装置或为如权利要求27所述的MEC设备。
  29. 一种计算机存储介质，其特征在于，所述计算机存储介质上存储有指令，当所述指令被处理器执行时，实现如权利要求1至13任一所述的任务处理方法。
PCT/CN2021/125852 2020-12-30 2021-10-22 任务处理方法、装置及系统、计算机存储介质 WO2022142631A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202011613356.4 2020-12-30
CN202011613356.4A CN114697342A (zh) 2020-12-30 2020-12-30 任务处理方法、装置及系统、计算机存储介质

Publications (1)

Publication Number Publication Date
WO2022142631A1 true WO2022142631A1 (zh) 2022-07-07

Family

ID=82131632

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/125852 WO2022142631A1 (zh) 2020-12-30 2021-10-22 任务处理方法、装置及系统、计算机存储介质

Country Status (2)

Country Link
CN (1) CN114697342A (zh)
WO (1) WO2022142631A1 (zh)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108063773A (zh) * 2018-02-11 2018-05-22 清华大学 一种基于移动边缘计算的应用服务接入方法及装置
CN109495929A (zh) * 2017-09-12 2019-03-19 华为技术有限公司 一种业务处理方法、移动边缘计算设备及网络设备
US20200136978A1 (en) * 2018-10-30 2020-04-30 Verizon Patent And Licensing Inc. Method and system for predictive edge resources
US20200145337A1 (en) * 2019-12-20 2020-05-07 Brian Andrew Keating Automated platform resource management in edge computing environments

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109495929A (zh) * 2017-09-12 2019-03-19 华为技术有限公司 一种业务处理方法、移动边缘计算设备及网络设备
CN108063773A (zh) * 2018-02-11 2018-05-22 清华大学 一种基于移动边缘计算的应用服务接入方法及装置
US20200136978A1 (en) * 2018-10-30 2020-04-30 Verizon Patent And Licensing Inc. Method and system for predictive edge resources
US20200145337A1 (en) * 2019-12-20 2020-05-07 Brian Andrew Keating Automated platform resource management in edge computing environments

Also Published As

Publication number Publication date
CN114697342A (zh) 2022-07-01

Similar Documents

Publication Publication Date Title
JP7183416B2 (ja) 時間依存ネットワーキング通信方法及び装置
US11445039B2 (en) Method and apparatus for providing edge computing services
CN113596191B (zh) 一种数据处理方法、网元设备以及可读存储介质
CN110647391A (zh) 面向星地协同网络的边缘计算方法及系统
CN111010702B (zh) 时延敏感网络通信方法及其装置
CN113726843B (zh) 边缘云系统、数据传输方法、设备及存储介质
JP6121535B2 (ja) ワイヤレスローカルエリアネットワークにおけるサービスの差別化に基づく動的なアソシエーションの順序付けのためのシステムおよび方法
WO2022067829A1 (zh) 动态触发边缘应用服务器实例化的方法及装置
US20150058441A1 (en) Efficient content caching management method for wireless networks
WO2024104284A1 (zh) 一种基于nwdaf对计算资源的管理和决策的方法
JP2024513803A (ja) ネットワーク情報エクスポージャ方法、装置、電子機器、及びコンピュータプログラム
CN113747513A (zh) 一种服务质量管理方法和装置
WO2022105807A1 (zh) 服务实例部署方法、跨域互访通道建立方法及相关装置
CN108566294B (zh) 一种支持计算平面的通信网络系统
CN105656978A (zh) 一种资源共享方法及装置
WO2022142631A1 (zh) 任务处理方法、装置及系统、计算机存储介质
CN113810442A (zh) 资源预留的方法、装置、终端及节点设备
WO2023001083A1 (zh) 服务提供方法、系统、网关、设备和存储介质
CN115665262A (zh) 一种请求处理方法、装置、电子设备及存储介质
JP2021503747A (ja) カップリングリダイレクト方法および装置
CN113366814B (zh) 对边缘计算系统中的资源分配进行管理的方法
CN103069917B (zh) 用于建立数据路径、设备和通信系统的方法
WO2018188447A1 (zh) 一种ip地址配置方法及装置
WO2023024956A1 (zh) 网络设备管理方法、装置、系统及存储介质
CN115412530B (zh) 一种多集群场景下服务的域名解析方法及系统

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21913376

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21913376

Country of ref document: EP

Kind code of ref document: A1