Fusion Communication Middle Platform for Edge-Cloud Collaborative Optimization Resource Scheduling Based on Pareto Improvement

Abstract

The present invention discloses a converged communication middleware for edge-cloud collaborative optimization of resource scheduling based on Pareto improvement, belonging to the technical field of task scheduling. The present invention supports the expansion of resources, terminals and services. Standard devices can be directly connected to the system, and users and terminals can be flexibly added through registration. At the same time, the system can meet the needs of subordinate or superior units to add service nodes, and realize the integration and comprehensive call of voice and data services with remote communication systems. Through edge-cloud collaborative computing, the control layer can allocate tasks to the most suitable computing resources according to the nature and requirements of the tasks, and at the same time utilize the high performance and large-scale computing power of cloud computing, so as to realize the intelligent allocation of resources and the efficient execution of tasks. This way of edge-cloud collaborative computing makes better use of the advantages of edge computing and cloud computing, and gives full play to the characteristics of fast response and low latency of the edge-cloud computing framework.

Description

Technical Field

[0001] The present invention relates to the technical field of task scheduling, and more particularly to a converged communication middleware for edge-cloud collaborative optimization of resource scheduling based on Pareto improvement. Background Art

[0002] In a process factory, the communication methods between processes generally adopt communication means such as telephones, broadcast intercoms, video monitoring, video conferencing, and intercoms. However, each terminal is independent of each other, and it is impossible to efficiently achieve interoperability between heterogeneous communication methods in daily communication and emergency situations.

[0003] In daily communication, business users do not have to care about how various devices are connected. Instead, as long as they issue target instructions in the most accustomed way according to business requirements, the communication middleware can provide an integrated solution for multimedia scheduling services including voice scheduling, video scheduling, data scheduling, etc., to meet the needs of efficient daily communication, instruction upload and download, and decision-making and command.

[0004] In the prior art, it is impossible to achieve horizontal interoperability between terminal devices such as telephones, monitors, and video conferencing, and it is impossible to automatically schedule edge servers. As a result, the utilization efficiency of terminal resources is low and the platform operation cost is high. Therefore, a converged communication middleware for edge-cloud collaborative optimization of resource scheduling based on Pareto improvement is proposed. Summary of the Invention

[0005] The technical problem to be solved by the present invention is: how to solve the problems in the prior art that edge servers cannot be automatically scheduled, the utilization efficiency of terminal resources is low, and the platform operation cost is high, and provide a converged communication middleware for edge-cloud collaborative optimization of resource scheduling based on Pareto improvement.

[0006] The present invention solves the above technical problems through the following technical solutions. The present invention includes a resource layer, a control layer, and an application layer;

[0007] The resource layer is used to integrate the completed business resources to implement the original business functions and provide the ability for comprehensive invocation;

[0008] The control layer is used to plan the optimal task scheduling scheme according to the tasks issued by the application layer by using the task scheduling algorithm in the edge-cloud collaborative computing framework, and then allocate the tasks to the corresponding edge servers or cloud servers through the resource layer;

[0009] The application layer is used to provide an entry for different tasks of various types of users to call the converged communication middleware.

[0010] Furthermore, in the resource layer, business subsystems of each department for a single requirement and other audio and video systems are integrated, and various Internet of Things devices, sensors, and edge servers are integrated. Through connection and collaboration, they interact with each business subsystem and perform protocol docking with softswitch devices, gateway devices at the terminal, or business subsystems in the resource layer, virtualize the accessed communication resources and uniformly identify them to achieve resource integration. They also connect different types of Internet of Things devices, access the Internet of Things devices to the edge cloud system, and achieve data transmission and protocol conversion.

[0011] Furthermore, in the resource layer, other audio and video systems include office telephone systems, intercom systems, conference systems, monitoring systems, individual soldier systems, SMS / email systems.

[0012] Furthermore, in the control layer, the task scheduling algorithm in the edge-cloud collaborative computing framework is the task scheduling algorithm based on Pareto improvement. This task scheduling algorithm comprehensively considers two objectives of QoS and ESS, obtains the optimal task scheduling scheme that can simultaneously improve QoS and ESS, and thus obtains the optimal task scheduling solution in the edge-cloud collaborative computing environment.

[0013] Furthermore, the specific processing process of the task scheduling algorithm based on Pareto improvement is as follows:

[0014] S11: Check whether a new task j is received. If so, add the new task j to the task set J; then check whether the task set J is empty. If it is not empty, sort the tasks in ascending order of the task arrival time.

[0015] S12: For each task j in the task set J: Traverse the edge server set E and the cloud service center c. For each edge server, check whether the task j can be executed on this edge server e; if it can, calculate the value of the QoS objective function Q on this edge server and select the task scheduling scheme that maximizes the Q value Continue to traverse the next edge server until all edge servers are checked to obtain the task scheduling scheme set

[0016] S13: Traverse the edge server set E and the cloud service center c again: For each edge server, check whether the task j can be executed on this edge server e; if it can, calculate the value of the ESS objective function S on this edge server and select the task scheduling scheme that maximizes the S value Continue to traverse the next edge server until all edge servers are checked to obtain the task scheduling scheme set

[0017] S14: and The intersection of and

[0018] is used as the final scheduling scheme Y, that is, the optimal task scheduling scheme;

[0018] S15: Continue to process the next task in the task set J until all tasks are processed.

[0019] Furthermore, in the step S12, the calculation formula of the QoS objective function Q is as follows:

[0020]

[0021] s.t. max(Q)

[0022]

[0023] where u(n) represents the nth user and j(n) represents the nth task; represents that user u has a request for task j; represents that the edge server e has the task resources required to run task j; represents the size of task j of the end user u; is the distance coefficient between the nodes where the user submits the task to the cloud service center, and d(u, c) is the distance between nodes u and c, μ 0 is the non-preference weight coefficient, μ 0 = -μ e μ e is the task execution preference weight coefficient associated with the edge server e, is the QoS coefficient of user u and task j.

[0024] Furthermore, in the step S13, the calculation formula of the ESS objective function S is as follows:

[0025]

[0026] s.t max(S)

[0027]

[0028] where, is the revenue index for scheduling the received tasks to the edge servers of the cloud service center, and respectively represent the average service energy consumption coefficients of the cloud service center and the edge server, and τ c and τ e respectively represent the task execution time equivalent values of the cloud service center and the edge server. The task execution time equivalent value τ and the average service energy consumption coefficient The product represents the system service consumption within the statistical period. is the ESS coefficient for user u and task j.

[0029] Furthermore, in the control layer, the edge-cloud collaborative computing modeling process of the task scheduling algorithm based on Pareto improvement is as follows:

[0030] S21: Define the edge-cloud collaborative computing model EC3M. EC3M is a six-tuple model, expressed as follows:

[0031] M EC3 =(U, J, E, c, O, θ)

[0032] where U is the set of users, including n(u) independent users, U = {u 0 , u 1 , …, u n(u)-1}, users do not interfere with each other, and various tasks submitted by users have time series correlation; J is the set of tasks, including n(j) service-oriented tasks, J = {j 0 , j 1 , …, j n(j)-1}, task j is expressed as j = {δ j , ∈ j , Δ j}, δ j is the computing resource required to execute task j, quantified by the CPU computing power required for each task, ∈ j is the storage resource required to execute task j, Δ j is the task type of task j: Δ j = 1 indicates that task j is a time-sensitive task, Δ j = 0 indicates that task j is a time-insensitive task; E is the set of edge servers, including n(e) geographically distributed edge servers, E = {e 0 , e 1 , …, e n(e)-1}, edge server e = {δ e , ∈ e}, δ e and ∈ e respectively represent the computing resource and storage resource of edge server e; c is the cloud service center, which has computing, storage, and network hardware resources and can load and run the task resources of all tasks; O = {max(Q), max(S)}, θ is the task scheduling algorithm based on Pareto improvement;

[0033] S22: Define the Quality of Service (QoS) of the user. QoS is used to focus on the service experience and quality of the user in the edge-cloud collaborative computing environment. The user tasks are locally executed on the edge server that receives the tasks. The shorter the response time of the user task request, the higher the QoS.

[0034] The QoS coefficient for user u and task j is calculated as follows:

[0035]

[0036] Among them, the QoS coefficient is related to the task size and task execution. The task size is quantified by the task execution time. represents the size of task j of the end user u. and respectively represent the proportions of locally executing task j on the receiving edge server, dispatching task j to other edge servers, and dispatching task j to the cloud service center. The corresponding task execution preference weight coefficients are and is the inter-node distance coefficient for local execution of the task on the receiving edge server. is the inter-node distance coefficient for the local edge server to dispatch the task to other edge servers; the distance degree value d(x, y) is related to the minimum bandwidth, cumulative delay, and reliability of the link between nodes x and y, and is equal to the cumulative delay divided by the product of reliability and minimum bandwidth.

[0037] S23: Define the System Service Effectiveness (ESS). ESS is used to focus on the system service revenue and system service consumption of the service provider in the edge-cloud collaborative computing. The ESS coefficient for user u and task j is calculated as follows:

[0038]

[0039] Among them, the ESS coefficient is related to the task size and task revenue. is the revenue index for executing the task on the local edge server. is the revenue index for the edge server to dispatch the received task to the neighboring edge server. is the revenue index for the edge server to dispatch the received task to the cloud service center.

[0040] Furthermore, in the application layer, three user interaction methods are provided for different user objects: providing a command and dispatch client for command and dispatch and management personnel; providing a mobile application client for mobile application scenarios; providing a standard API interface for third-party business systems.

[0041] The present invention has the following advantages compared with the prior art:

[0042] 1) Practicality and Efficiency

[0043] The system has simple and efficient operations, aiming to facilitate user operation and maintenance, while quickly responding to the needs of the decision-making and management levels. Through a unified operation platform and overall design, statistical information and decision-making support are provided in a timely manner to achieve the convenience of business processing and comprehensive management; in addition, the system construction cost is relatively low, protecting existing investments and avoiding resource waste.

[0044] 2) Reliability and Stability

[0045] The system adopts a distributed design, and each software module runs independently. Faults will not affect the operation of other modules. The operating system and database architecture are reliable and mature, and important data is backed up remotely. In terms of hardware, a redundant design is adopted, and the core switch, wired dispatching switch, and main server all have the function of dual-machine hot backup to ensure the stable operation of the system.

[0046] 3) Scalability and Flexibility

[0047] The system supports the expansion of resources, terminals, and services. Standard devices can be directly connected to the system, and users and terminals can be flexibly added through registration; at the same time, the system can meet the needs of subordinate or superior units to add service nodes, and realize the voice, data service interaction fusion and comprehensive call with remote communication systems; the system design conforms to the mainstream industry standards, has openness, and can be connected to other system platforms and share data.

[0048] 4) Efficiency

[0049] Through edge-cloud collaborative computing, the control layer can allocate tasks to the most suitable computing resources according to the nature and requirements of the tasks, while utilizing the high performance and large-scale computing power of cloud computing, thereby realizing the intelligent allocation of resources and the efficient execution of tasks; this edge-cloud collaborative computing method makes better use of the advantages of edge computing and cloud computing, and fully exerts the characteristics of fast response and low latency of the edge-cloud computing framework. Brief Description of the Drawings

[0050] Figure 1 It is a schematic diagram of the system architecture design of the integrated communication middleware platform in the first embodiment of the present invention;

[0051] Figure 2 It is a schematic flow diagram of the task scheduling algorithm based on Pareto improvement in the first embodiment of the present invention;

[0052] Figure 3 It is a schematic diagram of the integrated communication network design in the first embodiment of the present invention;

[0053] Figure 4It is a schematic diagram of the integration and docking of the intranet program-controlled switch and the converged communication middleware in the second embodiment of the present invention;

[0054] Figure 5 It is a schematic diagram of the integration and docking of the converged communication middleware and the venue audio in the second embodiment of the present invention;

[0055] Figure 6 It is a schematic diagram of the integration and docking of the converged communication middleware and the trunking system in the second embodiment of the present invention;

[0056] Figure 7 It is a schematic diagram of the integration and docking of the converged communication middleware and the SMS platform in the second embodiment of the present invention;

[0057] Figure 8 It is a schematic diagram of the docking of the converged communication middleware and the monitoring platform in the second embodiment of the present invention. Detailed implementation manners

[0058] The following makes a detailed description of the embodiments of the present invention. The embodiments are implemented on the premise of the technical solution of the present invention, and detailed implementation manners and specific operation processes are given. However, the protection scope of the present invention is not limited to the following embodiments.

[0059] Embodiment 1

[0060] This embodiment provides a technical solution: a converged communication middleware for edge-cloud collaborative optimization of resource scheduling based on Pareto improvement. The converged communication middleware created by the present invention is a cross-departmental, cross-regional, and multi-resource integrated communication and command system, which integrates resources such as Internet of Things devices, sensors, and edge servers to provide basic functions such as data collection and edge computing. By connecting and collaborating with the business service modules of the integrated subsystem, a comprehensive converged communication middleware that optimizes resource and task scheduling using an edge-cloud collaborative framework is constructed. According to the hierarchical structure of the converged communication middleware, its system architecture design is as Figure 1 shown:

[0061] The system (i.e., the converged communication middleware of the present invention) is generally divided into a resource layer, a control layer, and an application layer:

[0062] I. Resource layer

[0063] The resource layer is a key component of the converged communication middleware of the present invention. It integrates the business subsystems of each department facing single needs and other audio-visual systems, such as office telephone systems, intercom systems, conference systems, monitoring systems, single-soldier systems, SMS / email systems, etc. At the same time, the resource layer also integrates various important resources such as Internet of Things devices, sensors, and edge servers. The downstream terminals of the resource layer include Internet of Things devices, sensors, and edge servers, which undertake the tasks of data collection and edge computing in the resource layer and provide support for the upper-layer applications. The resource layer connects and collaborates with each business subsystem (i.e.,Figure 1 interact with the corresponding subsystems) and perform protocol docking with the softswitch devices, gateway devices or business subsystems in the resource layer of the terminal, virtualize and uniformly identify the accessed communication resources, and achieve resource integration. In addition, the resource layer is also responsible for connecting different types of Internet of Things devices, including gateway devices, edge routers, etc., connecting them to the edge cloud system, and realizing data transmission and protocol conversion to promote data circulation and interaction.

[0064] The converged communication middle platform uses service-based integration technology and utilizes the resource layer to integrate existing business systems instead of building new ones or replacing these systems. The resource layer integrates the completed business resources to achieve the original business functions and provides the ability for comprehensive invocation.

[0065] Through the functions of the resource layer, the control layer can integrate various devices into the system and connect different types of Internet of Things devices to the control layer. The resource layer acts as a bridge between the control layer and the terminal server, realizing seamless connection between the underlying resources and the upper-layer applications, thereby providing more flexible and efficient communication capabilities for the entire system. The role of the resource layer is to ensure that the platform can make full use of existing resources, achieve resource integration and optimized utilization, and improve the overall performance and functions of the system.

[0066] II. Control Layer

[0067] The control layer is the core of the converged communication middle platform, responsible for integrating and comprehensively invoking communication resources and implementing specific application business logics. It receives application layer instructions, specifically implements operation logics, and controls various resources by commanding each subsystem in the resource layer, such as functions like making calls, invoking monitoring images, and video conferencing.

[0068] Through cooperation with the access layer and the functions of the resource layer subsystems, the control layer realizes the integration and comprehensive invocation of communication resources and concretizes application business logics. In the edge cloud system, the control layer also needs to solve two key problems: First, dynamically formulate resource scheduling and pushing strategies according to task changes to optimize resource utilization; Second, while meeting user requirements (such as shortening task completion time), it is necessary to reduce the overall energy consumption of the system to balance the interests of users and service providers. To achieve these goals, the control layer coordinates the computing resources of the cloud and edge nodes and completes operations such as task offloading and service caching.

[0069] The control layer classifies various types of terminal resources and service resources accessed using differentiated icons, associates geographical location information, and presents them comprehensively based on the GIS interface, providing combined scheduling based on geographical location information. It can visually display various resources to be invoked with location as the key index, providing auxiliary decision-making for emergency event handling. The tasks issued by the application layer are planned into the best solutions through the scheduling optimization algorithm of the edge-cloud collaborative computing framework of the control layer, and then the control layer distributes the tasks to the corresponding edge servers or cloud servers through the resource layer. The task scheduling algorithm aims to comprehensively optimize task scheduling in the edge-cloud collaborative computing environment, balance user service requirements and system performance, and enable the entire system to operate efficiently.

[0070] In the edge-cloud collaborative computing framework, the design of the task scheduling algorithm is user- and service-provider-oriented, considering two objectives: quality of service (QoS) and system service effectiveness (ESS). Since optimizing a single objective may not necessarily guarantee that the other objective is also optimal, a trade-off needs to be made between the two. To this end, the algorithm introduces the concept of Pareto improvement in the field of economics, aiming to find a task scheduling solution that can improve both QoS and ESS simultaneously, thereby obtaining the optimal task scheduling solution in the edge-cloud collaborative computing environment.

[0071] When executing the Pareto improvement-based task scheduling algorithm on the edge server, first add the newly arrived tasks to the task set J. If multiple users initiate task requests within the same time period, sort the tasks in ascending order of task arrival time. The execution of the algorithm includes two stages:

[0072] The first stage is QoS objective optimization. Use the stochastic greedy approximation algorithm to generate m sets of task scheduling solutions, calculate the Q values in each set of task scheduling solutions, and select the task scheduling solution with the highest Q value. The target curve of QoS can be obtained from these solutions.

[0073] The second stage is ESS objective optimization. Similarly, use the stochastic greedy approximation algorithm to generate another m sets of task scheduling solutions, calculate the S values in each set of task scheduling solutions, and select the task scheduling solution with the highest S value, thereby obtaining the target curve of ESS.

[0074] Finally, gradually compare the task scheduling solutions obtained in the two stages through Pareto improvement, and select the tangent point or intersection point of the QoS and ESS target curves as the final optimal solution. The time complexity of this algorithm is O(n(j)×(log(n(j))+(n(e)+1)×m)).

[0075] The processing process of the above "stochastic greedy approximation algorithm" is the acquisition process of the following task scheduling solution set as follows.

[0076] As shown below Figure 2 The specific process of the above task scheduling algorithm is further described as follows:

[0077] The algorithm inputs are the edge server set E, the task set J, the cloud service center c, the resource parameters weight parameters delay parameter δ c and δ e The output is the task scheduling scheme Y.

[0078] First, check if a new task j is received. If so, add the new task j to the task set J. Then check if the task set J is empty. If not, sort the tasks in ascending order of the task arrival time.

[0079] Next, for each task j in the task set J: traverse the edge server set E and the cloud service center c. For each edge server, check if the task j can be executed on the edge server e (when and j ∈ J, e ∈ E); if it can, calculate the value of the QoS objective function Q on the edge server and select the task scheduling scheme that maximizes the value of Q Continue to traverse the next edge server until all edge servers are checked to obtain the task scheduling scheme set

[0080] Next, traverse the edge server set E and the cloud service center c again: for each edge server, check if the task j can be executed on the edge server e; if it can, calculate the value of the ESS objective function S on the edge server and select the task scheduling scheme that maximizes the value of S Continue to traverse the next edge server until all edge servers are checked to obtain the task scheduling scheme set Y j 2 [m];

[0081] Finally, take the and intersection as the final scheduling scheme Y. Then continue to process the next task in the task set J until all tasks are processed.

[0082] The calculation formula of QoS, that is, Q, is as follows:

[0083]

[0084] s.t. max(Q)

[0085]

[0086] The calculation formula of ESS, i.e., S, is as follows:

[0087]

[0088] s.t max(S)

[0089]

[0090] III. Application Layer

[0091] The application layer provides visual management and convenient access to the resources and services of the edge cloud, and provides an entry for various types of users to call the converged communication middleware for different tasks. For different user objects, three user interaction methods are provided: the command and dispatch client for command and dispatch and management personnel; the mobile application client for mobile application scenarios; and a standard API interface for third-party business systems, enabling third-party business systems to conveniently, quickly, and reliably call various multimedia services such as voice and video during their business use.

[0092] The system adopts an operation mode that combines "peace" and "war". In the normal operation state, it can meet the functional requirements such as daily business communication and system information monitoring. When an emergency event is triggered, it can enter the wartime operation mode. In the wartime mode, the instructions manually operated by users have the highest scheduling control authority, can start various emergency response processes, and can upgrade the event handling level at any time according to the development of the event, and report the event handling to the upper level. After the "war" operation ends, the data records can be retrospectively analyzed and statistically analyzed. In the "peace" time, most of the scheduling tasks are completed by the optimized scheduling algorithm within the edge-cloud collaborative computing framework, and the resource scheduling and push strategies are dynamically formulated according to the task changes, reducing the overall energy consumption of the system while efficiently meeting the requirements.

[0093] Converged Communication Networking

[0094] According to the requirements of the converged communication middleware for emergency communication guarantee, on the premise of meeting network transmission security, its system networking design is as Figure 3 shown below:

[0095] The converged communication middleware platform is deployed based on the enterprise intranet / Internet and the edge Internet of Things architecture, and can provide an integrated solution for multimedia dispatching services including voice dispatching, video dispatching, map dispatching, etc. The converged communication platform connects various existing devices such as PSTN, trunking systems, IP phones, video conferencing, video surveillance, 4G single soldiers, satellite phones, emergency command vehicles, etc. to the edge cloud system, deeply integrates various systems such as audio, video, and instant messaging, and provides SDK development interfaces to third-party systems. At the same time, it enables multiple users to conveniently and efficiently conduct visual multi-level command and dispatch and remote discussions through the network at different locations. The scheduling algorithm of the edge-cloud collaborative computing framework dynamically adjusts resource allocation and scheduling strategies according to the dynamic changes of task loads, maximizes resource utilization efficiency while ensuring that the quality of service meets user requirements, and reduces the overall system energy consumption. This mechanism takes into account both the user's service experience and the platform's operating costs, achieving a balance between efficiency and overhead. In case of emergencies, it can be manually switched to the wartime mode, and manual instructions have the highest authority in the dispatching platform, thus maximizing the efficiency of handling emergencies.

[0096] The edge-cloud collaborative computing framework is further described below.

[0097] The edge-cloud collaborative computing framework is a computing architecture designed to achieve the collaborative work of edge computing and cloud computing, with the goal of combining edge computing and cloud computing to achieve more efficient computing and data processing. The following introduces the edge-cloud collaborative computing modeling process of the task scheduling algorithm based on Pareto improvement:

[0098] Definition 1: Edge-Cloud Collaborative Computing Model (EC3M). EC3M is a six-tuple model, denoted as M EC3 :

[0099] M EC3 =(U, J, E, c, O, θ)

[0100] where U is the set of users, consisting of n(u) independent users, U = {u 0 , u 1 , …, u n(u)-1}. Users do not interfere with each other, and various tasks submitted by users have time-series correlations, so the number and type of tasks can be predicted. J is the set of tasks, consisting of n(j) service-oriented network applications (i.e., tasks), J = {j 0 , j 1 , …, j n(j)-1}. Task j is represented as j = {δ j , ∈ j , Δ j}, δ jis the computing resource required to execute task j, quantified by the CPU computing power (GHz / task) required for each task, ∈ j is the storage resource required to execute task j, Δ j is the task type of task j: Δ j = 1 indicates that task j is a time-sensitive task, Δ j = 0 indicates that task j is a non-time-sensitive task. Each task type contains various different tasks to meet different user requirements, and tasks can be further divided into multiple subtasks.

[0101] E is a set of edge servers, containing n(e) geographically distributed edge servers, E = {e 0 , e 1 , …, e n(e)-1}, and the edge server e = {δ e , ∈ e}, where δ e and ∈ e represent the computing resource and storage resource of edge server e respectively. Edge servers are restricted by hardware resources and can only load the task resources required for part of the tasks at a time. Edge servers can subdivide tasks and execute them locally or dispatch them to remote locations according to scheduling decisions.

[0102] c is the cloud service center, which has a large amount of hardware resources such as computing, storage, and network, and can load and run the task resources of all tasks. The cloud service center manages and monitors edge servers, effectively predicts user tasks, and pushes appropriate task resources to relevant edge servers. O is the optimization goal of edge-cloud collaborative computing, quantified by QoS and ESS (denoted as Q and S respectively), O = {max(Q), max(S)}. θ is the optimization algorithm for resource deployment and task scheduling, that is, the task scheduling algorithm described above in the present invention.

[0103] Definition 2: Quality of Service (QoS) of users. QoS mainly focuses on the service experience and quality of users in the edge-cloud collaborative computing environment. User tasks are executed locally on the edge server that receives the tasks. The shorter the response time of the user task request, the higher the QoS. The QoS coefficient is shown in the following formula:

[0104]

[0105] The QoS coefficient is related to the task size and task execution. The task size is quantified by the task execution time. represents the size of task j of the end user u, and represent the proportions of locally executing task j on the receiving edge server, dispatching task j to other edge servers, and dispatching task j to the cloud service center respectively. The corresponding task execution preference weight coefficients are and is the inter - node distance coefficient for local execution of tasks on the receiving edge server, is the inter - node distance coefficient for the local edge server to dispatch tasks to other edge servers. The distance degree value d(x,y) is related to the minimum bandwidth, cumulative delay, and reliability of the link between nodes x and y. The distance degree value is equal to the cumulative delay divided by the product of reliability and minimum bandwidth. The larger the minimum bandwidth, the smaller the cumulative delay, and the higher the reliability, the smaller the distance degree value.

[0106] The objective function of QoS is as follows:

[0107]

[0108] s.t. max(Q)

[0109]

[0110] where u(n) represents the nth user and j(n) represents the nth task; means that user u has a request for task j; means that edge server e has the task resources required to run task j; means the size of task j of the end - user u; is the inter - node distance coefficient for the user to submit the task to the cloud service center, and d(u,c) is the distance degree between nodes u and c, μ 0 is the non - preference weight coefficient, μ 0 =-μ e μ e is the task execution preference weight coefficient associated with edge server e, is the QoS coefficient of user u and task j. It can be seen from the above formula that the edge server has the resources required to run the task, and the larger the proportion of local task execution, the larger the Q value, that is, the higher the user service quality of edge - cloud collaborative computing.

[0111] Definition 3: System Service Effect (ESS): ESS mainly focuses on the system service revenue and system service consumption of service providers in edge - cloud collaborative computing. The system service effect coefficient is as follows:

[0112]

[0113] The system service revenue coefficient is related to the task size and task revenue; is the revenue index for task execution on the local edge server; is the revenue index for the edge server to dispatch the received task to neighboring edge servers; is the revenue index for the edge server to dispatch the received tasks to the cloud service center, represents the size of task j of end-user u, and as well as the same as above.

[0114] The objective function of ESS is as follows:

[0115]

[0116] s.t max(S)

[0117]

[0118] wherein, the system service consumption within the statistical period is the product of the task execution time equivalent value τ and the average service energy consumption coefficient of, is the revenue index of the edge server that schedules the received tasks to the cloud service center, and respectively represent the average service energy consumption coefficients of the cloud service center and the edge server, τ c and τ e respectively represent the task execution time equivalent values of the cloud service center and the edge server. The product of the task execution time equivalent value τ and the average service energy consumption coefficient represents the system service consumption within the statistical period. represents the size of task j of end-user u. The energy consumption coefficient depends on the hardware / software cost, as well as the system operation and maintenance cost. The former includes the hardware / software purchase cost and the depreciation cost, and the latter involves the device power consumption as well as the management and service cost. From the single quantity analysis, is much greater than is the ESS coefficient for user u and task j. The higher the system service revenue and the lower the system service consumption, the larger the S value, that is, the higher the system service effect of edge-cloud collaborative computing.

[0119] In the edge-cloud collaborative computing model, the objective optimization of QoS and ESS involves resource deployment and task scheduling. (1) It is necessary to predict the types and quantities of user tasks, reasonably push the task resources to the edge server, and efficiently utilize the computing, storage, network and other resources of the edge server. (2) It is also necessary to optimize the task scheduling, improve QoS, and enhance ESS through the collaborative processing of tasks by local edge servers, other edge servers and cloud service centers.

[0120] Example 2

[0121] The following further elaborates on the methods of accessing this system (converged communication middleware) for each type.

[0122] The task scheduling method based on Pareto improvement involves the ways for multiple terminals and edge server platforms to access the converged communication middleware, including voice terminals, video conferencing terminals, terminal systems, video surveillance, and business applications. Through these methods, terminal devices and edge servers can communicate and collaborate effectively with the converged communication middleware to optimize resource scheduling. To more clearly describe these methods and related architectures, the corresponding architecture diagrams are attached below, showing how terminal devices and edge servers are connected and interact with the converged communication middleware, as well as the data flow and communication paths between them.

[0123] As Figure 4 shown, it is a schematic diagram of the integrated connection between the intranet program-controlled switch and the converged communication middleware. Through system integration, the converged communication softswitch host device unifies all terminals on one interface, enabling quick calls to any terminal, and can also temporarily form groups for group calls and mass calls. In case of emergencies, it can initiate mass calls to various communication terminals in a specified area.

[0124] As Figure 5 shown, it is a schematic diagram of the integrated connection between the converged communication middleware and the venue audio system. The converged communication middleware realizes interconnection with the conference mixing console system of the command center by deploying a venue audio gateway, achieving seamless integration of local digital conference microphones and telephone systems, meeting the application requirements of video teleconferences in different industries, ensuring clear voice quality, and having a powerful telephone access control ability.

[0125] As Figure 6 shown, it is a schematic diagram of the integrated connection between the converged communication middleware and the trunking system. For the already deployed trunking intercom system, it provides a solution for wireless trunking gateway access, unifying trunking systems of different systems, different frequencies / channels, and different manufacturers into the IP network, realizing unified scheduling through a multimedia dispatching console, and achieving interconnection between different types of intercom systems and between the trunking intercom system and other communication terminals.

[0126] As Figure 7 shown, it is a schematic diagram of the integrated connection between the converged communication middleware and the SMS platform. It uses an SMS gateway to connect with the operator's SMS platform to realize the functions of sending group SMS and receiving SMS on the dispatching console. It supports single sending, group sending, forwarding, and resending of long SMS, as well as automatic SMS reminders. It supports common SMS templates for various emergencies. It can perform functions such as alarm linkage, group sending, and queuing.

[0127] As Figure 8As shown in the figure, it is a schematic diagram of the docking between the converged communication middleware platform and the monitoring platform, which refers to the docking with a third-party video monitoring platform. For a self-built video monitoring platform that can provide an external docking protocol or docking development interface and has the ability to dock with the converged communication middleware platform, the docking with the video access converged communication middleware platform should be achieved through the monitoring platform access method.

[0128] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A converged communication platform based on Pareto-improved edge-cloud collaborative optimization resource scheduling, characterized by: Includes resource layer, control layer and application layer; The resource layer is used to integrate the completed business resources to realize the original business functions and provide comprehensive calling capabilities; The control layer is used to plan the optimal task scheduling scheme based on the tasks issued by the application layer using the task scheduling algorithm in the edge-cloud collaborative computing framework, and then allocate the tasks to the corresponding edge servers or cloud servers through the resource layer; In the control layer, the task scheduling algorithm in the edge-cloud collaborative computing framework is a task scheduling algorithm based on the Pareto improvement. The task scheduling algorithm comprehensively considers the two objectives of QoS and ESS to obtain the optimal task scheduling solution that can improve both QoS and ESS at the same time, thereby obtaining the optimal task scheduling solution in the edge-cloud collaborative computing environment; The specific processing process of the task scheduling algorithm based on Pareto improvement is as follows: S11: Check whether a new task j is received. If yes, add the new task j to the task set J; then check whether the task set J is empty. If not, sort the tasks in ascending order of their arrival time; S12: For each task j in the task set J: traverse the edge server set E and the cloud service center c, and for each edge server, check whether task j can be executed on the edge server e; if so, calculate the value of the QoS objective function Q on the edge server and select the task scheduling scheme Y that maximizes the Q value j 1 ; Continue to traverse the next edge server until all edge servers have been checked and obtain the task scheduling solution set Y j 1 [m]; S13: Traverse the edge server set E and cloud service center c again: For each edge server, check whether task j can be executed on the edge server e; if so, calculate the value of the ESS objective function S on the edge server and select the task scheduling scheme Y that maximizes the S value j 2 ; Continue to traverse the next edge server until all edge servers have been checked and obtain the task scheduling solution set Y j 2 [m]; S14: Y j 1 [m] and Y j 2 The intersection of [m] is taken as the final scheduling scheme Y, i.e. the optimal task scheduling scheme; S15: Continue to process the next task in task set J until all tasks are processed; In step S12, the calculation formula of the QoS objective function Q is as follows: stmax(Q) Among them, u(n) represents the nth user, j(n) represents the nth task; Indicates that user u has a request for task j; It means that edge server e has the task resources required to run task j; represents the size of task j of end user u; is the distance coefficient between nodes from which the user submits tasks to the cloud service center, d(u,c) is the distance between nodes u and c, μ0 is the non-preference weight coefficient, μ0=-μ e , μ e is the task execution preference weight coefficient associated with edge server e, is the QoS coefficient of user u and task j; The application layer is used to provide an entry for various types of users to call the converged communication middle platform for different tasks.

2. According to claim 1, the Pareto-improved edge-cloud collaborative optimization resource scheduling fusion communication middle station is characterized in that: In the resource layer, the business subsystems of each department facing a single demand and other audio and video systems are integrated, and various IoT devices, sensors and edge servers are integrated. Through connection and collaboration, it interacts with each business subsystem and conducts protocol docking with the soft switching equipment, gateway equipment or business subsystem of the resource layer of the terminal, virtualizes and uniformly identifies the accessed communication resources to achieve resource integration; it also connects different types of IoT devices and connects the IoT devices to the edge cloud system to achieve data transmission and protocol conversion.

3. The fusion communication middle station based on Pareto-improved edge-cloud collaborative optimization resource scheduling according to claim 2 is characterized in that: In the resource layer, other audio and video systems include office telephone system, intercom system, conference system, monitoring system, individual system, and SMS / email system.

4. The fusion communication middle station based on Pareto-improved edge-cloud collaborative optimization resource scheduling according to claim 1 is characterized in that: In step S13, the calculation formula of the ESS objective function S is as follows: st max(S) in, is the revenue index of the edge server that dispatches the received tasks to the cloud service center, and Represent the average service energy consumption coefficients of the cloud service center and edge server, τ c and τ e They represent the task execution time equivalent value of the cloud service center and the edge server, respectively, and the task execution time equivalent value τ and the average service energy consumption coefficient The product of represents the system service consumption within the statistical period. is the ESS coefficient of user u and task j.

5. According to claim 4, the fusion communication middle station based on Pareto-improved edge-cloud collaborative optimization resource scheduling is characterized in that: In the control layer, the edge-cloud collaborative computing modeling process based on the Pareto-improved task scheduling algorithm is as follows: S21: Define the edge-cloud collaborative computing model EC3M. EC3M is a six-tuple model, which is expressed as follows: M EC3 =(U,J,E,c,O,θ) Where U is the user set, including n(u) independent users, U = {u0,u1,…,u n(u)-1 }, users do not interfere with each other, and the various tasks submitted by users have time series correlation; J is a task set, including n(j) service-oriented tasks, J = {j0,j1,…,j n(j)-1 }, task j is represented by j = {δ j ,∈ j ,Δ j }, δ j is the computational resource required to execute task j, quantified by the CPU computing power required for each task, ∈ j is the storage resource required to execute task j, Δ j is the task type of task j: Δ j =1 means task j is a time-sensitive task, Δ j = 0 means that task j is a time-insensitive task; E is the set of edge servers, including n(e) geographically distributed edge servers, E = {e0, e1, …, e n(e)-1 }, edge server e={δ e ,∈ e }, δ e and ∈ e represents the computing resources and storage resources of the edge server e respectively; c is the cloud service center, which has computing, storage, and network hardware resources and can load and run the task resources of all tasks; O = {max(Q), max(S)}, θ is the task scheduling algorithm based on Pareto improvement; S22: Define the user's quality of service QoS. QoS is used to focus on the user's service experience and quality in the edge-cloud collaborative computing environment. The user's task is executed locally on the edge server that receives the task. The shorter the response time of the user's task request, the higher the QoS. QoS coefficient for user u and task j The calculation method is as follows: Among them, the QoS coefficient is related to the task size and task execution. The task size is quantified by the task execution time. represents the size of task j of end user u, and They represent the proportion of receiving edge servers to execute task j locally, assigning task j to other edge servers, and assigning task j to the cloud service center. The corresponding task execution preference weight coefficients are and is the inter-node distance coefficient of the task executed locally on the receiving edge server, It is the node distance coefficient for the local edge server to dispatch tasks to other edge servers; the distance degree value d(x,y) is related to the minimum bandwidth, cumulative delay and reliability of the link between nodes x and y, and is equal to the cumulative delay divided by the product of reliability and minimum bandwidth; S23: Define the system service effect ESS, which is used to focus on the system service revenue and system service consumption of the service provider in edge-cloud collaborative computing. The ESS coefficient for user u and task j is The calculation method is as follows: Among them, the ESS coefficient is related to the task size and task benefit; is the profit index of executing tasks on the local edge server; is the profit index of the edge server dispatching the received tasks to the neighboring edge servers; It is the profit index of the edge server dispatching the received tasks to the cloud service center.

6. The converged communication platform based on Pareto-improved edge-cloud collaborative optimization resource scheduling according to claim 1 is characterized in that: In the application layer, three user interaction modes are provided for different user objects: a command and dispatch client is provided for command and dispatch and management personnel; a mobile application client is provided for mobile application scenarios; and a standard API interface is provided for third-party business systems.

Images