WO2018130163A1 - Scheduling method and device for mobile cloud computing platform - Google Patents

Abstract

Disclosed is a scheduling method for a mobile cloud computing platform, comprising the following steps: monitoring resource occupancy of cloud system programs in a group of cloud systems when running the group of cloud systems; if the current total resource occupancy is higher than a preset upper threshold, dynamically changing, according to resource request amounts occupied by all of the plurality of cloud system programs, an execution sequence of the group of cloud system programs, so that execution priorities of the cloud system programs with relatively low resource occupancy are higher than those of the cloud system programs with relatively high resource occupancy; executing the cloud system programs with relatively high resource occupancy during idle periods at execution intervals of the plurality of cloud system programs with relatively low resource occupancy; and if the current total resource occupancy is lower than a preset lower threshold, restoring or maintaining an original execution sequence. Also disclosed is a scheduling device for a mobile cloud computing platform, comprising a monitoring module, an execution priority scheduling module, an idle work scheduling module, and an execution priority restoring module.

Description

Scheduling method and device for mobile cloud computing platform

Related application

The present application claims priority to Chinese Patent Application No. 201710020920.3, filed on Jan. 11, 2017, entitled "Scheduling Method and Apparatus for Mobile Cloud Computing Platform."

Technical field

The invention belongs to the field of cloud computing and mobile communication, and in particular relates to a scheduling method and device of a mobile cloud computing platform.

Background technique

Intelligent mobile terminals, such as smart phones, tablets, and portable computers, have become the network terminal devices that people can't keep in every day. At present, the smart phone terminal product has a short update cycle, and consumers tend to replace the new version of the mobile terminal product on average every year. However, these updated products usually stack performance parameters such as CPU computing power, memory size, and storage capacity, while other hardware (such as display screens and network adapters) are less updated. Too frequent replacement of electronic products puts economic pressure on consumers and is not conducive to energy conservation and environmental protection. Moreover, for some users who are pursuing "sufficient use", their smart terminals often have excess performance and waste.

On the other hand, with the development of mobile communication technology, the fifth generation mobile communication technology (5G) is about to be commercialized, and the intelligent mobile terminal will enjoy a more high-speed and convenient network interconnection environment. In the 5G era, there will be 50 billion connected Internet of Everything services in the world, and people's computing power and quality of service requirements for intelligent terminals are getting higher and higher. Mobile cloud computing will become one of the key technologies for 5G network innovation services. Mobile cloud computing is a new mode of delivery and use of IT resources or information services. It is the product of introducing cloud computing into the mobile Internet.

In addition, in the smart mobile terminal market, there is also a need for enterprise applications and security services. There are also users who need to use a variety of operating systems or custom systems. For example, if a user uses a Windows system while on a business trip, and a communication mobile phone uses an Android system, if you can use multiple operating systems on the same mobile terminal, it will bring great convenient. It also brings the problem of allocating resources for a large number of these cloud operating systems in the cloud server.

Summary of the invention

The present invention solves the above problems by providing a scheduling method and apparatus for a mobile cloud computing platform. The mobile cloud computing platform integrates the cloud terminal resource to provide an operating environment for the plurality of cloud systems, and each group of the cloud system internally runs a plurality of cloud system programs in sequence, and the cloud system program is distributed to the corresponding mobile terminal through the network to provide Cloud computing services.

The solution of the present invention is a scheduling method of a mobile cloud computing platform, which includes the following steps: monitoring the resource occupancy of each cloud system program during running a group of cloud systems; if the current total resource occupancy is higher than When the preset upper threshold is preset, the execution order of the group of cloud system programs is dynamically changed according to the resource request amount occupied by each of the plurality of cloud system programs, so that the execution of the cloud system program with a small resource occupancy is prioritized The execution priority of the cloud system program that is higher than the resource usage is large; the cloud system program with a large resource occupancy is executed during the idle period of the execution interval of the multiple cloud system programs with less resource occupation; if the current resource When the total occupancy is lower than the preset lower threshold, the original execution order of the cloud system program is restored or maintained.

Further, the resource occupancy includes occupying a length of a central processing unit time slice, occupying a usage of the random access memory, occupying a usage of the data storage, or occupying a usage of the network resource.

Preferably, the upper threshold value and/or the lower limit threshold value can be adjusted in real time according to the amount of cloud terminal resources currently fused by the mobile cloud computing platform.

Further, the scheduling method of the mobile cloud computing platform further includes: recovering resources of the cloud system program that the mobile terminal logs out, and allocating to the remaining running cloud systems in the cloud system group where the log-out cloud system program is located. program.

Further, the scheduling method of the mobile cloud computing platform further includes preferentially allocating resources of the logged out cloud system program to a cloud system program with a larger resource occupancy.

Further, the scheduling method of the mobile cloud computing platform further includes allocating a part of the cloud terminal resources fused by the mobile cloud computing platform to one or more accessed mobile terminals.

Another aspect of the present invention provides a scheduling apparatus for a mobile cloud computing platform, including: a monitoring module, configured to monitor a resource occupancy of each of the cloud system programs during running of a group of cloud systems; and execute a priority scheduling module, When determining whether the current total resource occupation amount is higher than a preset upper threshold, dynamically changing the execution order of the group cloud system program according to the resource request amount occupied by each of the plurality of cloud system programs, so that The execution priority of the cloud system program with a smaller resource usage is higher than the execution priority of the cloud system program with a large resource occupancy; the idle work scheduling module performs the execution interval of multiple cloud system programs with a small resource occupancy. The idle system performs a cloud system program with a large amount of resources; and the priority recovery module is configured to restore or maintain the original execution order of the cloud system program when the current total resource occupancy is lower than a preset lower threshold.

In still another aspect, the scheduling device of the mobile cloud computing platform can include a storage and a processor, wherein the processor executes program commands stored in the memory to perform the scheduling method of the cloud computing platform described above.

The invention has the beneficial effects of providing a scheduling mechanism of the mobile cloud computing platform, optimizing and rationally allocating resources, and providing a stable system running environment for the mobile cloud computing terminal to call.

DRAWINGS

1 is a schematic diagram showing the operation mode of a mobile cloud computing platform according to the present invention;

2 is a schematic diagram showing a scheduling process of a mobile cloud computing platform according to the present invention;

3 is a flow chart showing a scheduling process of a mobile cloud computing platform in accordance with the present invention.

detailed description

FIG. 1 is an application platform of the present invention and a working mode thereof, wherein the cloud computing terminal and the cloud computing server are networked through a communication device, and the networking manner adopted may be a wired connection, a wireless connection, or any combination thereof. It should be understood that the cloud computing terminal, the cloud computing server, and the communication device may each be plural in number. Cloud computing terminals can operate in both basic and cloud computing modes. Maintain basic communication functions such as calls, text messages, etc. in the basic mode. In cloud mode, you can access the cloud computing server through a high-speed network, and load rich system resources, such as logging in to use a variety of customized entertainment systems and office systems. The cloud computing terminal can be embodied as a smart phone, a tablet computer, a portable game machine, or the like having a network communication function. Specifically, the cloud computing terminal and the cloud computing server may be connected by using a mobile communication connection (for example, 2G, 3G, 4G, or 5G communication mode), a wide area network, or an Ethernet network. Communication devices include network adapters, network switches, routers, modems, fiber/twin-pair lines, wireless APs, antennas, mobile communication base stations, and the like.

The cloud computing server may include a cloud cluster architecture, a convergence scheduling module, and a transmission management module, wherein the fusion scheduling module includes a processor, a memory, and an external scheduler, and the transmission management module includes a network manager and a high speed network interface. The cloud cluster architecture integrates the cloud terminal resources to provide a running environment for the multiple groups of cloud systems, and each group of cloud systems internally runs multiple cloud system programs in sequence, and the cloud system programs process the communication of the corresponding cloud computing terminals through the network manager to provide Cloud computing services. The processor may cause the external scheduler to schedule resources from various cloud computing devices in the cloud cluster network or from the cloud computing terminal according to the current running condition of the cloud computing server. In an example, when the storage capacity of the cloud computing server is tight, the processor may analyze the idle NAS resource connected to the cloud computing server in the cloud cluster network, and the available storage content of the NAS is addressed to the memory through the scheduler. The cloud computing server is used internally or for use by one or more cloud computing terminals. In another example, when the memory capacity of the cloud computing server is tight, the processor may analyze the idle memory resource of the cloud computing terminal connected to the cloud computing server, and integrate the available idle memory resources of the cloud computing terminal by using an external scheduler. In the memory pool in the cloud computing server.

2 is a schematic diagram of a scheduling process of a mobile cloud computing platform according to the present invention. The cloud computing platform may have a resource allocation module for adjusting the allocation amount of the software and hardware resources according to the request amount and the request priority of the cloud computing service layer and the transmission management unit, and according to the network throughput of the communication transmission unit. The resource allocation module includes: a memory fusion unit, configured to fuse memory resources of other cloud computing devices by using a cloud computing server; and a storage fusion unit, configured to integrate storage resources of other cloud computing devices by using a cloud computing server.

The cloud computing platform further includes an internal scheduler, which performs the following steps: monitoring the resource occupancy of each of the cloud system programs during running of the cloud system; if the current total resource occupancy is higher than a preset upper threshold, Dynamically changing the execution order of the group of cloud system programs according to the resource request amount occupied by each of the plurality of cloud system programs, so that the execution priority of the cloud system program with less resource occupancy is higher than the resource occupancy ratio Execution priority of a large cloud system program; executing a cloud system program with a large resource occupancy during an idle period of an execution interval of a plurality of cloud system programs having a small resource occupancy; if the current total resource occupancy is lower than a preset When the lower threshold is reached, the original execution order of the cloud system program is restored or maintained.

3 is a flow chart of a mobile cloud computing platform scheduling method in accordance with a preferred embodiment of the present invention.

The method specifically includes the following steps:

S101. Pre-allocate cloud system resources and group cloud systems required by the same environment according to requirements of user registration customization;

S102. Monitor resource usage of each of the cloud system programs during running of a group of cloud systems;

S103. Determine whether the current total resource occupation is higher than a preset upper threshold.

S104. Dynamically change an execution order of the group of cloud system programs according to a resource request quantity occupied by each of the multiple cloud system programs, so that a cloud system program with a smaller resource occupancy has a higher execution priority than resource usage. The execution priority of a large number of cloud system programs;

S105. Perform a cloud system program with a large resource occupation in an idle period of an execution interval of multiple cloud system programs with a small resource occupancy;

S106. Continue to monitor resource usage of each of the cloud system programs;

S107. When it is determined that the current total resource occupancy is lower than a preset lower threshold, the original execution sequence of the cloud system program is restored or maintained in step S108.

The above scheduling method will be described below by a specific example.

When the cloud computing server is working, the plurality of cloud computing client terminals send execution request information of the cloud system program to the cloud computing server. The cloud computing server will receive each cloud system program request information sent by multiple terminals. After receiving the request information of the plurality of cloud system programs, the resource allocation management unit 31 writes the execution order of the plurality of cloud system programs into the resource allocation management storage area. If there is still an unexecuted cloud system program in the execution sequence of the resource allocation management storage area, the newly requested cloud system program will write to the end of the sequence, that is, the newly requested cloud system program has the lowest execution priority.

For example, the execution order stored in the resource allocation management storage area includes nine cloud system programs to be executed, including an A program of the first client terminal, a B program of the second client terminal, a C program of the third client terminal, and the like. . After the first client terminal issues the execution request information of the D program to the cloud computing server, the resource allocation management unit writes the D program into the table of the execution order, and the execution order of the D program is 10, that is, the execution priority level is the lowest. Then, the resource allocation management unit determines whether the current usage rate of the cloud computing server CPU is higher than a threshold, and if it is higher than the threshold, dynamically changes the execution order of each cloud system program. At this time, the resource allocation management unit determines the resource request amount requested by each cloud system program, that is, the length of the CPU time slice, and changes the execution order of each cloud system program according to the length of the occupied time slice. The principle of changing the execution order is that the amount of resource requests is large, that is, the cloud system program that takes a long time slice has a low priority level, and the resource request amount is small, that is, the cloud system program occupying a short time slice has a high priority.

According to the above change principle, if the D program of the first client terminal occupies the shortest time slice, only 5 milliseconds, and its execution priority is the highest, the execution order can be set to "1". The B program occupation time slice of the second client terminal is 10 milliseconds, which is only longer than the time slice occupied by the A program, and the execution priority level is high, and the execution order can be set to “2”. The A program of the first client terminal occupies the longest time slice, and has the lowest execution priority level, and the execution order can be set to "10". The C program of the third client terminal takes a longer time slice and the execution priority is lower, and the execution order is set to "9".

By analogy, the resource allocation management unit of the cloud computing server changes the priority level of execution of each cloud system program according to the length of the time slice. Of course, if the usage rate of the CPU is lower than the threshold, the resource allocation management unit does not change the execution order of the cloud system program, that is, the execution order of the multiple cloud system programs is arranged in the chronological order of the cloud system program request. Then, according to the execution order stored in the resource allocation management storage area, the cloud system program is sequentially executed from the high priority level to the low priority level. When the CPU executes the cloud system program, it is a short time slice allocated to the cloud system program, in which a process or thread of the cloud system program is executed, and when the process or thread is executed, the CPU has a free time slice. A process or thread that can execute the next cloud system program.

After a process or thread is executed, the resource allocation management unit needs to reorder the execution order of the resource allocation management storage area, delete the executed cloud system program from the execution sequence, and update each cloud system program. The remaining length of the occupied time slice, and the execution order of each cloud system program is arranged according to the remaining length of the occupied time slice. Of course, it is also determined whether all the cloud system programs are executed. If the cloud system program is still not executed, the new cloud system program is received, or the CPU executes the cloud system program according to the updated execution order.

It can be seen that the resource allocation management unit dynamically changes the execution order of multiple cloud system programs in real time. When a thread or process of a cloud system program is executed, the execution sequence is updated immediately, and the next cloud system program is executed. a process or thread. In this way, multiple cloud system programs are alternately executed, and the cloud system program with a small resource request has a high priority, a short execution time, and a fast response time of the client terminal.

It can be seen that even if a cloud computing client terminal sends a cloud system program request with a large resource request to the cloud computing server, the cloud system program with a small amount of other resource requests requested later cannot be executed in time, thereby avoiding In the post-requested cloud system program response time is too long, effectively avoiding the "dead" phenomenon of the client terminal.

The execution time of the cloud system program with a small amount of resource requests is often short, and the cloud system program processes that execute the small amount of resource requests often have more idle time slices to execute cloud system programs with large resource requests. Therefore, a cloud system program that preferentially executes a small resource request does not wait too long or cannot execute a cloud system program that causes a large amount of resource requests. When the CPU executes the cloud system program, data exchange with the RAM and the ROM is required, and data needs to be transmitted to the client terminal through the network interface. Therefore, a certain cloud system program occupies a long CPU time slice, and the RAM usage occupied by the cloud system is There are also many ROM resources and network resources.

The above is only a preferred embodiment of the present invention, and the present invention is not limited to the above embodiment, and as long as it achieves the technical effects of the present invention by the same means, it should fall within the scope of protection of the present invention. Various modifications and changes may be made to the technical solutions and/or embodiments within the scope of the invention.

While the principles of the present invention have been described in detail, the preferred embodiments of the present invention The details of the embodiments are not intended to limit the scope of the invention, and any obvious changes, such as equivalent transformations, simple substitutions, etc., based on the technical solutions of the present invention, fall within the scope of the present invention. Within the scope of protection.

Claims (8)

A scheduling method for a mobile cloud computing platform, the mobile cloud computing platform converging cloud terminal resources to provide an operating environment for a plurality of cloud systems, and each group of cloud systems internally simulates running multiple cloud system programs in sequence, and the cloud system program passes The network is assigned to a corresponding mobile terminal to provide a cloud computing service, characterized in that the scheduling method comprises: Monitor the resource footprint of each of the cloud system programs while running a set of cloud systems; If the current total resource occupancy is higher than a preset upper threshold, dynamically changing the execution order of the cloud system program according to the resource request amount occupied by each of the multiple cloud system programs, so that the resource occupancy is The execution priority of a smaller cloud system program is higher than the execution priority of a cloud system program with a large resource occupancy; Performing a cloud system program with a large resource occupancy in an idle period of an execution interval of a plurality of cloud system programs having a small resource occupancy; If the current total resource occupancy is lower than the preset lower threshold, the original execution order of the cloud system program is restored or maintained. The scheduling method of the mobile cloud computing platform according to claim 1, wherein the resource occupancy comprises occupying a length of a central processing unit time slice, occupying a usage of a random access memory, occupying a usage of a data storage device, or occupying network resources. Usage amount. The scheduling method of the mobile cloud computing platform according to claim 1, wherein the upper threshold and/or the lower threshold can be adjusted in real time according to the amount of cloud terminal resources currently fused by the mobile cloud computing platform. The method for scheduling a mobile cloud computing platform according to claim 1, further comprising: The resources of the cloud system program that the mobile terminal logs out are recovered and distributed to the remaining running cloud system programs in the cloud system group in which the log-out cloud system program is located. The scheduling method of the mobile cloud computing platform according to claim 4, further comprising: Priority is given to allocating resources of the cloud system program to a cloud system program with a larger resource usage. The method for scheduling a mobile cloud computing platform according to claim 1, further comprising: A portion of the cloud terminal resources fused by the mobile cloud computing platform are distributed to one or more accessed mobile terminals. A scheduling device for a mobile cloud computing platform, the mobile cloud computing platform converging cloud terminal resources to provide an operating environment for a plurality of cloud systems, and each group of cloud systems internally simulates running multiple cloud system programs in sequence, and the cloud system program passes The network is assigned to a corresponding mobile terminal to provide a cloud computing service, characterized in that the scheduling device comprises: a monitoring module, configured to monitor a resource occupancy of each of the cloud system programs during running of a group of cloud systems; The priority scheduling module is configured to dynamically change the group cloud system according to the resource request amount occupied by each of the plurality of cloud system programs, when the current resource occupancy total is higher than a preset upper threshold. The execution order of the program is such that the execution priority of the cloud system program with a smaller resource occupancy is higher than the execution priority of the cloud system program with a large resource occupancy; The idle work scheduling module executes a cloud system program with a large resource occupancy in an idle period of an execution interval of a plurality of cloud system programs with a small resource occupancy; The priority recovery module is configured to restore or maintain the original execution order of the cloud system program when the current resource usage total is lower than a preset lower threshold. A scheduling device for a mobile cloud computing platform, comprising: a storage and a processor, wherein the processor executes program commands stored in the memory to perform any of claims 1 to 6 The method described is required.

Images