CN114077498A - Method and system for selecting and transferring calculation load facing to mobile edge calculation - Google Patents

Method and system for selecting and transferring calculation load facing to mobile edge calculation Download PDF

Info

Publication number
CN114077498A
CN114077498A CN202111380847.3A CN202111380847A CN114077498A CN 114077498 A CN114077498 A CN 114077498A CN 202111380847 A CN202111380847 A CN 202111380847A CN 114077498 A CN114077498 A CN 114077498A
Authority
CN
China
Prior art keywords
sequence
power
container
working
obtaining
Prior art date
Legal status (The legal status 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 status listed.)
Granted
Application number
CN202111380847.3A
Other languages
Chinese (zh)
Other versions
CN114077498B (en
Inventor
陈明
殷知磊
张静静
程军强
梁辉
楚杨阳
曹洁
王博
周开来
彭伟伟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhengzhou University of Light Industry
Original Assignee
Zhengzhou University of Light Industry
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 Zhengzhou University of Light Industry filed Critical Zhengzhou University of Light Industry
Priority to CN202111380847.3A priority Critical patent/CN114077498B/en
Publication of CN114077498A publication Critical patent/CN114077498A/en
Application granted granted Critical
Publication of CN114077498B publication Critical patent/CN114077498B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
    • G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
    • G06F9/505—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals considering the load
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
    • G06F9/45533—Hypervisors; Virtual machine monitors
    • G06F9/45558—Hypervisor-specific management and integration aspects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
    • G06F9/5011—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resources being hardware resources other than CPUs, Servers and Terminals
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
    • G06F9/5011—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resources being hardware resources other than CPUs, Servers and Terminals
    • G06F9/5016—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resources being hardware resources other than CPUs, Servers and Terminals the resource being the memory
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5083—Techniques for rebalancing the load in a distributed system
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
    • G06F9/45533—Hypervisors; Virtual machine monitors
    • G06F9/45558—Hypervisor-specific management and integration aspects
    • G06F2009/4557—Distribution of virtual machine instances; Migration and load balancing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
    • G06F9/45533—Hypervisors; Virtual machine monitors
    • G06F9/45558—Hypervisor-specific management and integration aspects
    • G06F2009/45583—Memory management, e.g. access or allocation
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Landscapes

  • Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention relates to the technical field of artificial intelligence, in particular to a method and a system for selecting and transferring a calculation load facing to mobile edge calculation. The method collects power data and temperature data for an initial working vessel. And when the temperature data is abnormal, screening out the transferable working container according to the power data of other working containers. And judging the complementary correlation degree according to the correlation between the power data of the migratable working container before the container migration and the power data of the initial working container. And judging mutual exclusion association degree according to the power data of the migratable working container after the container migration and the power data of the initial working container. And screening out the optimal transferable working container according to the complementary relevance and the mutual exclusion relevance. The invention comprehensively analyzes the complementarity and mutual exclusion among the working containers and provides a scientific and effective container migration strategy.

Description

Method and system for selecting and transferring calculation load facing to mobile edge calculation
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a method and a system for selecting and transferring a calculation load facing to mobile edge calculation.
Background
In an edge calculation scenario with heavy load such as video analysis, the heat dissipation situation of devices such as roadside edge calculation of V2X is relatively severe, and when one device cannot meet the heat dissipation requirement, the temperature of the device is high, and power operation needs to be reduced, so that the calculation load of the current node needs to be transferred to the remaining calculation units.
At present, containerization development and projects are mature, and better convenience is provided for containerization computation load migration. The containerized migration means is similar to the virtualized virtual machine migration means, but the basic binary files of the software in operation are the same, so that only the input data and the temporary storage data in operation need to be migrated to another container, and a large amount of data required by operation does not need to be transferred, so that the virtual machine migration method has incomparable migration efficiency of the traditional virtualization.
The container migration strategy and technology are more, and the purpose of the container migration of the edge computing node is mainly to reduce the load of the equipment which is easily overheated. However, in the container migration policy, a problem that a container node receives a new load, which causes a bus bottleneck of internal hardware, other bottlenecks caused by device context switching waiting, and the like easily occurs, so that the power of the migrated container node does not reach an ideal target, that is, the migrated container node is not matched with the migrated container node. Unmatched container migration node pairs can greatly reduce the overall performance of the equipment and reduce the service quality.
Disclosure of Invention
In order to solve the above technical problems, an object of the present invention is to provide a method and a system for selecting and migrating a computation load facing to a moving edge computation, wherein the adopted technical solution is as follows:
the invention provides a calculation load selection and migration method facing mobile edge calculation, which is characterized by comprising the following steps:
obtaining a first power sequence and a first temperature sequence of an initial working container within a preset first sampling time period; when the elements in the first temperature sequence are larger than a preset migration temperature threshold value, obtaining second power sequences of other working containers in a preset second sampling time period; judging a container quality index according to the difference between the data in the second power sequence, and eliminating the other working containers corresponding to the container quality index which is lower than a preset quality index threshold value to obtain a plurality of transferable working containers;
obtaining a third power sequence of the migratable work container over the first sampling period; obtaining a complementary relevance degree of the initial working container and each transferable working container according to a first linear correlation of the first power sequence and the third power sequence;
transferring the load of the initial working container to the transferable working container to obtain a fourth power sequence and a second temperature sequence of the transferable working container; obtaining a heat dissipation coefficient sequence according to the fourth power sequence and the second temperature sequence; a linear sequence of load changes for the initial work container and the migratable work container; obtaining a second linear correlation of the linear sequence and the fourth power sequence; obtaining the mutual exclusion relevance of the initial working container and each transferable working container according to the fourth power sequence, the first power sequence, the second linear correlation and the heat dissipation coefficient sequence;
and screening out the optimal transferable working container according to the complementary relevance and the mutual exclusion relevance.
Further, the evaluating the container quality indicator according to the difference between the data in the second power sequence comprises:
acquiring a second power average value, a maximum second power and a minimum second power of the second power sequence, and acquiring the container quality index according to a container quality index formula; the container quality indicator formula comprises:
Figure BDA0003365409280000021
wherein C is the container quality index, alpha is a first fitting coefficient, PHIs the maximum second power, PLIs the minimum second power, PMeanFor the second power mean, trunc () is a truncation function.
Further, the obtaining of the complementary relevance of the initial working container and each of the migratable working containers according to the first linear correlation of the first power sequence and the third power sequence comprises:
obtaining a maximum third power in the third power sequence; obtaining a difference sequence of the third power sequence and the maximum third power; taking a Pearson correlation coefficient of the difference sequence and the first power sequence as the first linear correlation; and processing the first linear correlation by a truncation function to obtain the complementary correlation degree.
Further, the obtaining a heat dissipation coefficient sequence according to the fourth power sequence and the second temperature sequence includes:
obtaining the heat dissipation coefficient sequence according to a heat dissipation coefficient calculation formula; the heat dissipation coefficient calculation formula is as follows:
Figure BDA0003365409280000022
wherein E is the heat dissipation coefficient, P is the instantaneous fourth power in the fourth power sequence, TMaxIs a preset limit temperature, T is the secondInstantaneous second temperature of the two temperature sequence.
Further, the obtaining a second linear correlation of the linear sequence and the fourth power sequence comprises:
and taking the Pearson correlation coefficient of the linear sequence and the fourth power sequence as the second linear correlation.
Further, the obtaining the mutual exclusion relevance of the initial working container and each of the migratable working containers according to the fourth power sequence, the first power sequence, the second linear correlation and the heat dissipation coefficient sequence includes:
acquiring the mutual exclusion association degree according to the mutual exclusion association degree acquisition formula; the mutual exclusion association degree obtaining formula comprises:
Figure BDA0003365409280000031
wherein Z is the mutual exclusion relevance, PMeanBIs a fourth power mean value, P, of the fourth power sequenceMeanAIs the first power mean of the first power sequence, β is the second fitting coefficient, V is the second correlation, EBAnd for the heat dissipation coefficient sequence of the movable working container, trunc () is a truncation function.
Further, the screening out the optimal migratable work container according to the complementary relevance and the mutual exclusion relevance comprises:
and taking the product of the complementary relevance and the mutual exclusion relevance as a matching index, and screening out the optimal transferable working container according to the matching index.
Further, the screening the optimal migratable work container according to the matching index includes:
and matching the initial working container and the migratable working container by adopting a KM algorithm according to the matching index to obtain the optimal migratable working container.
The invention also provides a system for selecting and migrating the computational load facing the mobile edge computation, which comprises a memory, a processor and a computer program which is stored in the memory and can run on the processor, and is characterized in that the processor realizes any step of the method for selecting and migrating the computational load facing the mobile edge computation when executing the computer program.
The invention has the following beneficial effects:
1. according to the embodiment of the invention, the complementary correlation degree between the working containers is analyzed through the third power sequence of the transferable working container before the transfer and the first power sequence of the initial working container. And then analyzing the mutual exclusion relevance between the working containers according to the fourth power sequence of the migrated migratable working containers. The optimal transferable working container is screened out according to the complementary relevance and the mutual exclusion relevance, the relevance of the two working containers is comprehensively analyzed from the mutual exclusion and complementary relation, the most appropriate working container can be selected from the working containers of the equipment for transferring, the performance of the equipment is ensured, and the overall working efficiency is improved.
2. According to the embodiment of the invention, the quality of other working containers is evaluated according to the data of the second power sequence of other working containers in the second sampling time before the other working containers are migrated, and the migratable working containers are screened out, so that the working containers which are not suitable for being used as container migration nodes are greatly reduced, and the accuracy of the subsequent matching process is ensured.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a flowchart of a method for selecting and migrating a computation load facing a mobile edge computation according to an embodiment of the present invention.
Detailed Description
To further illustrate the technical means and effects of the present invention for achieving the predetermined objects, the following detailed description of the computing load selecting and migrating method and system for mobile edge computing according to the present invention will be provided with reference to the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" refers to not necessarily the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following describes a specific scheme of a method and a system for selecting and migrating a computation load facing to mobile edge computation in detail with reference to the accompanying drawings.
Referring to fig. 1, a flowchart of a method for selecting and migrating a computation load facing a mobile edge computation according to an embodiment of the present invention is shown, where the method includes:
step S1: obtaining a first power sequence and a first temperature sequence of an initial working container within a preset first sampling time period; when the elements in the first temperature sequence are larger than a preset migration temperature threshold value, obtaining second power sequences of other working containers in a preset second sampling time period; and judging the container quality index according to the difference between the data in the second power sequence, and eliminating other working containers corresponding to the container quality index which is lower than a preset quality index threshold value to obtain a plurality of transferable working containers.
For a container performing an edge calculation task, each unit is complex, and therefore in the embodiment of the present invention, the total power of the initial working container is used as analysis data to obtain a first power sequence of the initial working container within a preset first sampling time period. Meanwhile, the temperature of the current initial working container is recorded, the wafer temperature of the CPU is used as analysis data in the embodiment of the invention, and other heating area data related to the load, such as the temperature of a passive heat sink, a GPU, a memory module, an optical communication module and the like, can be selected in other embodiments.
In an embodiment of the present invention, the first sampling period is set to two minutes, and the sampling frequency is set to 1 second, i.e., the length of the first power sequence and the first temperature sequence is 120.
It should be noted that a general drive test unit in the working container needs to operate the deep neural network to implement the sensing processing on the image or video data, thereby implementing the edge analysis. In the processing process, the more the target data is, namely the greater the calculation load is, along with the increase of the target data, the working container is easy to be fully loaded at the moment, so that the temperature is increased, and the working container can adjust the temperature through a heat dissipation system in the full-loading process, so that the embodiment of the invention does not need to consider the content of load information and only analyzes the temperature change condition.
In actual work, there will be a limit temperature, after the temperature of working container reached limit temperature, in order to protect the hardware, thereby can reduce work efficiency and play the effect of cooling. Therefore, a migration temperature threshold needs to be set in a range smaller than the limit temperature, so that the initial working container can be timely migrated, and the working efficiency is prevented from being affected. That is, when the first temperature sequence of the initial working vessel reaches the migration temperature threshold, the initial working vessel is considered to be fully loaded at that time and vessel migration is required.
It should be noted that, when the temperature is higher than the migration temperature, although the operation can still be continued, the temperature problem is not solved, and finally the temperature is raised to the limit temperature, so as to realize overheat protection, and the temperature is reduced by controlling the power reduction mode, so as to affect the overall working performance, and therefore, it is necessary to find other suitable working container nodes for container migration, and to ensure the working efficiency.
In the embodiment of the present invention, the limit temperature is set to 95 degrees, and the migration temperature threshold is set to 80 degrees, and in other embodiments, the limit temperature may be specifically set according to the type of the work container and the type of the work task, which is not limited herein.
In order to avoid carrying out the useless work that the migration leads to in too much improper work container in the container migration process, need carry out primary screening to other work containers, select the work container that the quality is relatively poor, avoid at its extravagant efficiency on one's body in subsequent process, specifically include:
and acquiring a second power sequence of other working containers in a preset second sampling time period, namely acquiring total power data of other containers in the second sampling time period before the container migration time. The fluctuation of the second power sequence reflects the working stability of other current working containers, and when the second power sequence has obvious fluctuation, the power is adjusted according to strategies such as that the other working containers are also performing working tasks or performing container migration, and the like, so that the larger the fluctuation is, the more unsuitable the container migration is, namely, the larger the fluctuation is, the worse the container quality index is. The data fluctuation can be represented according to the difference between the data in the second power sequence, that is, the container quality index can be judged by using the difference between the data, and the method specifically includes:
acquiring a second power average value, a maximum second power and a minimum second power of the second power sequence, and acquiring a container quality index according to a container quality index formula; the container quality index formula comprises:
Figure BDA0003365409280000051
wherein C is a container quality index, alpha is a first fitting coefficient, PHIs the maximum second power, PLAt a minimum second power, PMeanFor the second power mean, trunc () is a truncation function. The ratio of the difference between the maximum second power and the minimum second power to the average value of the second power is used for expressing the data fluctuation in the container quality index formula; adjusting the value of the volatility by taking the first fitting coefficient as the weight of the volatility; the truncation function is used for standardizing the numerical value of the container quality index, and facilitates data analysis.
In the embodiment of the present invention, the second sampling period is set to one minute, and the sampling frequency is set to one second, i.e., the length of the second power sequence is 60. Alpha is set to 2, and the truncation function truncates a value less than 0 to 0, and in other embodiments, the specific setting may be analyzed according to specific data, and is not limited herein.
And eliminating other working containers corresponding to the container quality indexes lower than the preset quality index threshold value, reducing the data volume and obtaining a plurality of transferable working containers. It should be noted that the quality index threshold needs to be specifically set according to the model of the device, the type of the executed task, and the like, and is not limited herein.
Step S2: obtaining a third power sequence of the migratable working container within the first sampling time period; and obtaining the complementary relevance of the initial working container and each transferable working container according to the first linear correlation of the first power sequence and the three power sequences.
In order to analyze the matching relationship between the initial working container and the migratable working container, the length of the data needs to be uniform, that is, a third power sequence of the migratable working container in the first sampling time period is obtained. It should be noted that, before the container migration process, the less relevant the work data of the initial work container and the migratable work container is, the more different the work content and the work status between the two work containers are, i.e. the more suitable the container migration is. Therefore, the complementary association degree between the initial working container and each transferable working container is obtained according to the first linear correlation between the first power sequence and the third power sequence, that is, the greater the first linear correlation is, the greater the complementary association degree is, and the less suitable the two working containers are for migration, specifically comprising:
obtaining a maximum third power in the third power sequence; obtaining a difference sequence of the third power sequence and the maximum third power; taking the Pearson correlation coefficient of the difference sequence and the first power sequence as a first linear correlation; and processing the first linear correlation by a truncation function to obtain a complementary correlation degree. The complementary relation can be better embodied through the Pearson correlation coefficient of the difference sequence and the first power sequence.
Step S3: transferring the load of the initial working container to a transferable working container to obtain a fourth power sequence and a second temperature sequence of the transferable working container; obtaining a heat dissipation coefficient sequence according to the fourth power sequence and the second temperature sequence; a linear sequence of load changes for the initial work container and the migratable work container; obtaining a second linear correlation of the linear sequence and a fourth power sequence; and acquiring the mutual exclusion relevance of the initial working container and each transferable working container according to the fourth power sequence, the first power sequence, the second linear correlation and the heat dissipation coefficient sequence.
The load of the initial working container is the working data volume, and the load of the initial working container is transferred to the transferable working container. It should be noted that loads can be classified into a mixture of computation-intensive loads, storage-intensive loads and memory-intensive loads, and therefore, regardless of which type of load is the load at this time, as long as a bottleneck occurs in a certain type of load, the operation of the entire working container is subjected to a bottleneck, that is, an abnormality occurs in power variation. Therefore, a fourth power sequence of the movable working container after the container is moved needs to be obtained for analysis.
During the container migration process, the load in the migratable container will gradually increase along with the container migration process, the load of the initial working container will gradually decrease along with the container migration process, and a linear sequence of the load changes of the two containers can be obtained. For a good container migration strategy, the power in the migratable container will gradually increase with increasing load during the container migration process, and can also be considered as a linear relationship, i.e., in a good container migration strategy, the fourth power sequence should exhibit the same linear correlation with the linear sequence, and the pearson correlation coefficient of the linear sequence and the fourth power sequence is taken as the second linear correlation.
In order to ensure the stability of the transferable work container after the container is transferred, the temperature information of the transferable work container needs to be monitored, namely, the second temperature sequence is obtained. And ensuring the working state of the transferable working container according to the temperature sequence presented by the second temperature sequence. Obtaining a heat dissipation coefficient sequence according to the fourth power sequence and the second temperature sequence, namely the higher the fourth power is, the higher the risk of heat accumulation is, the worse the heat dissipation performance of the transportable container is, and the smaller the heat dissipation coefficient is; the smaller the second temperature is, the better the heat dissipation performance is, the larger the margin of the system is, and the larger the heat dissipation coefficient is. Specifically, obtaining the heat dissipation coefficient sequence includes:
fitting a heat dissipation coefficient calculation formula by a mathematical modeling method according to the relation between the power temperature and the heat dissipation coefficient, and obtaining a heat dissipation coefficient sequence according to the heat dissipation coefficient calculation formula; the calculation formula of the heat dissipation coefficient is as follows:
Figure BDA0003365409280000071
wherein E is the heat dissipation coefficient, P is the instantaneous fourth power in the fourth power sequence, TMaxT is the instantaneous second temperature of the second temperature sequence, at a predetermined limit temperature.
The heat dissipation coefficient may represent the heat dissipation capability of the working container, and in a good container migration policy, the heat dissipation coefficient of the migratable working container should be larger, and the second linear correlation should be more correlated, so the mutual exclusion association degree of the initial working container and each migratable working container is further analyzed by combining the data of the fourth power sequence and the first power sequence, and the larger the mutual exclusion association degree is, the more unsuitable the working containers are migrated, specifically including:
and according to the relation between the mutual exclusion relevancy and each coefficient, introducing a difference value between the fourth power mean value and the first power mean value to represent the power relevancy, fitting by a mathematical modeling method to obtain a mutual exclusion relevancy obtaining formula, and obtaining the mutual exclusion relevancy according to the mutual exclusion relevancy obtaining formula. The mutual exclusion association degree obtaining formula comprises the following steps:
Figure BDA0003365409280000072
wherein Z is the mutual exclusion association degree, PMeanBIs the fourth power mean value, P, of the fourth power sequenceMeanAIs the first power mean of the first power sequence, beta is the second fitting coefficient, V is the second correlation, EBFor the heat dissipation coefficient sequence of the transferable working container, trunc () is a truncation function.
In the mutual exclusion association degree obtaining formula, the difference between the fourth power average value and the first power average value represents the similarity degree of the two working containers after being started, if the difference exists, the mutual exclusion association degree is larger, which means that a rejection problem occurs, a bus bottleneck can be caused, and the power of the transferable working container can be reduced along with load transfer instead.
The meaning of the second fitting coefficient is to enlarge or reduce the difference between the fourth power mean value and the first power mean value, and the difference is used for adjusting data; conversely, when the second fitting coefficient is greater than one, the value range of the denominator can be ensured by amplifying the value, and the value of the second fitting coefficient can be set as required in a specific implementation situation, which is not limited too much, in the embodiment of the present invention, the value of the second fitting coefficient is 1.
The value range of the second correlation processed by the truncation function is [0,1], and when the value range is close to 1, the power change and the load change of the transferable working container are completely linear in the starting process after the container is transferred, so that the starting process of the transferable working container is smooth, and the bus bottleneck is not easy to appear; on the contrary, when the power is close to 0, the power change and the load change of the migratable work container are not linear, the power can have sudden abnormal changes, and irregular abnormal power is likely to cause the interior of the migratable work container to encounter a bus bottleneck, and cause the power to be unstable, thereby affecting the work efficiency.
Step S4: and screening out the optimal transferable working container according to the complementary relevance and the mutual exclusion relevance.
Therefore, the greater the complementary correlation degree and the mutual exclusion correlation degree, the less suitable the two working containers for migration, but the problem of overheating of the initial working container can still be solved, and therefore, a migratable working container with the minimum complementary correlation degree and the minimum mutual exclusion correlation degree needs to be found as an optimal migratable working container for container migration, which specifically includes:
and taking the product of the complementary relevance and the mutual exclusion relevance as a matching index, matching the initial working container and the migratable working container by adopting a KM algorithm according to the matching index, and taking the migratable working container with the minimum matching index as the optimal migratable working container. The container migration process may be performed according to the optimal migratable work container.
In summary, the embodiment of the present invention collects power data and temperature data of the initial working container. And when the temperature data is abnormal, screening out the transferable working container according to the power data of other working containers. And judging the complementary correlation degree according to the correlation between the power data of the migratable working container before the container migration and the power data of the initial working container. And judging mutual exclusion association degree according to the power data of the migratable working container after the container migration and the power data of the initial working container. And screening out the optimal transferable working container according to the complementary relevance and the mutual exclusion relevance. The embodiment of the invention comprehensively analyzes the complementarity and mutual exclusivity among the working containers and provides a scientific and effective container migration strategy.
The invention also provides a system for selecting and migrating the computational load facing the mobile edge computation, which comprises a memory, a processor and a computer program which is stored in the memory and can run on the processor, and is characterized in that the processor realizes any step of the method for selecting and migrating the computational load facing the mobile edge computation when executing the computer program.
It should be noted that: the precedence order of the above embodiments of the present invention is only for description, and does not represent the merits of the embodiments. And specific embodiments thereof have been described above. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.

Claims (9)

1. A method for selecting and migrating computation load facing to mobile edge computation is characterized by comprising the following steps:
obtaining a first power sequence and a first temperature sequence of an initial working container within a preset first sampling time period; when the elements in the first temperature sequence are larger than a preset migration temperature threshold value, obtaining second power sequences of other working containers in a preset second sampling time period; judging a container quality index according to the difference between the data in the second power sequence, and eliminating the other working containers corresponding to the container quality index which is lower than a preset quality index threshold value to obtain a plurality of transferable working containers;
obtaining a third power sequence of the migratable work container over the first sampling period; obtaining a complementary relevance degree of the initial working container and each transferable working container according to a first linear correlation of the first power sequence and the third power sequence;
transferring the load of the initial working container to the transferable working container to obtain a fourth power sequence and a second temperature sequence of the transferable working container; obtaining a heat dissipation coefficient sequence according to the fourth power sequence and the second temperature sequence; a linear sequence of load changes for the initial work container and the migratable work container; obtaining a second linear correlation of the linear sequence and the fourth power sequence; obtaining the mutual exclusion relevance of the initial working container and each transferable working container according to the fourth power sequence, the first power sequence, the second linear correlation and the heat dissipation coefficient sequence;
and screening out the optimal transferable working container according to the complementary relevance and the mutual exclusion relevance.
2. The method of claim 1, wherein the evaluating a container quality indicator according to the difference between data in the second power sequence comprises:
acquiring a second power average value, a maximum second power and a minimum second power of the second power sequence, and acquiring the container quality index according to a container quality index formula; the container quality indicator formula comprises:
Figure FDA0003365409270000011
wherein C is the container quality index, alpha is a first fitting coefficient, PHIs the maximum second power, PLIs the minimum second power, PMeanFor the second power mean, trunc () is a truncation function.
3. The method for selecting and migrating computational loads facing mobile edge computing according to claim 1, wherein the obtaining the complementary relevance of the initial work container and each of the migratable work containers according to the first linear correlation of the first power sequence and the third power sequence comprises:
obtaining a maximum third power in the third power sequence; obtaining a difference sequence of the third power sequence and the maximum third power; taking a Pearson correlation coefficient of the difference sequence and the first power sequence as the first linear correlation; and processing the first linear correlation by a truncation function to obtain the complementary correlation degree.
4. The method for selecting and migrating computational loads for mobile edge computing according to claim 1, wherein the obtaining a thermal dissipation factor sequence according to the fourth power sequence and the second temperature sequence comprises:
obtaining the heat dissipation coefficient sequence according to a heat dissipation coefficient calculation formula; the heat dissipation coefficient calculation formula is as follows:
Figure FDA0003365409270000021
wherein E is the heat dissipation coefficient, P is the instantaneous fourth power in the fourth power sequence, TMaxT is the instantaneous second temperature of the second temperature sequence.
5. The method of claim 1, wherein the obtaining the second linear correlation between the linear sequence and the fourth power sequence comprises:
and taking the Pearson correlation coefficient of the linear sequence and the fourth power sequence as the second linear correlation.
6. The method for selecting and migrating computational loads facing mobile edge computing according to claim 4, wherein the obtaining the mutual exclusion relevance of the initial working container and each of the migratable working containers according to the fourth power sequence, the first power sequence, the second linear correlation and the heat dissipation coefficient sequence comprises:
acquiring the mutual exclusion association degree according to the mutual exclusion association degree acquisition formula; the mutual exclusion association degree obtaining formula comprises:
Figure FDA0003365409270000022
wherein Z is the mutual exclusion relevance, PMeanBIs a fourth power mean value, P, of the fourth power sequenceMeanAIs the first power mean of the first power sequence, β is the second fitting coefficient, V is the second correlation, EBAnd for the heat dissipation coefficient sequence of the movable working container, trunc () is a truncation function.
7. The method for selecting and migrating computational loads facing mobile edge computing according to claim 1, wherein the step of screening out the optimal migratable work container according to the complementary correlation degree and the mutual exclusion correlation degree comprises:
and taking the product of the complementary relevance and the mutual exclusion relevance as a matching index, and screening out the optimal transferable working container according to the matching index.
8. The method for selecting and migrating computational loads facing mobile edge computing according to claim 6, wherein the screening out the optimal migratable work container according to the matching index comprises:
and matching the initial working container and the migratable working container by adopting a KM algorithm according to the matching index to obtain the optimal migratable working container.
9. A mobile edge computing oriented computing load selection and migration system comprising a memory, a processor and a computer program stored in said memory and executable on said processor, characterized in that said processor when executing said computer program implements the steps of the method according to any of claims 1 to 8.
CN202111380847.3A 2021-11-20 2021-11-20 Method and system for selecting and transferring calculation load facing to mobile edge calculation Active CN114077498B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111380847.3A CN114077498B (en) 2021-11-20 2021-11-20 Method and system for selecting and transferring calculation load facing to mobile edge calculation

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111380847.3A CN114077498B (en) 2021-11-20 2021-11-20 Method and system for selecting and transferring calculation load facing to mobile edge calculation

Publications (2)

Publication Number Publication Date
CN114077498A true CN114077498A (en) 2022-02-22
CN114077498B CN114077498B (en) 2023-03-28

Family

ID=80284029

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111380847.3A Active CN114077498B (en) 2021-11-20 2021-11-20 Method and system for selecting and transferring calculation load facing to mobile edge calculation

Country Status (1)

Country Link
CN (1) CN114077498B (en)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170366616A1 (en) * 2016-06-16 2017-12-21 Veniam, Inc. Systems and methods for managing containers in a network of moving things
CN109167671A (en) * 2018-07-11 2019-01-08 国网信通亿力科技有限责任公司 A kind of adapted communication system equally loaded dispatching algorithm towards quantum key distribution business
US20200145337A1 (en) * 2019-12-20 2020-05-07 Brian Andrew Keating Automated platform resource management in edge computing environments
US20200296155A1 (en) * 2020-03-27 2020-09-17 Intel Corporation Method, system and product to implement deterministic on-boarding and scheduling of virtualized workloads for edge computing
CN111694636A (en) * 2020-05-11 2020-09-22 国网江苏省电力有限公司南京供电分公司 Electric power Internet of things container migration method oriented to edge network load balancing
CN111880939A (en) * 2020-08-07 2020-11-03 曙光信息产业(北京)有限公司 Container dynamic migration method and device and electronic equipment
CN112181620A (en) * 2020-09-27 2021-01-05 郑州轻工业大学 Big data workflow scheduling method for sensing service capability of virtual machine in cloud environment
CN112910960A (en) * 2021-01-15 2021-06-04 郑州轻工业大学 Virtual network online migration method and device with time delay, resource and energy consumption perception

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170366616A1 (en) * 2016-06-16 2017-12-21 Veniam, Inc. Systems and methods for managing containers in a network of moving things
CN109167671A (en) * 2018-07-11 2019-01-08 国网信通亿力科技有限责任公司 A kind of adapted communication system equally loaded dispatching algorithm towards quantum key distribution business
US20200145337A1 (en) * 2019-12-20 2020-05-07 Brian Andrew Keating Automated platform resource management in edge computing environments
US20200296155A1 (en) * 2020-03-27 2020-09-17 Intel Corporation Method, system and product to implement deterministic on-boarding and scheduling of virtualized workloads for edge computing
CN111694636A (en) * 2020-05-11 2020-09-22 国网江苏省电力有限公司南京供电分公司 Electric power Internet of things container migration method oriented to edge network load balancing
CN111880939A (en) * 2020-08-07 2020-11-03 曙光信息产业(北京)有限公司 Container dynamic migration method and device and electronic equipment
CN112181620A (en) * 2020-09-27 2021-01-05 郑州轻工业大学 Big data workflow scheduling method for sensing service capability of virtual machine in cloud environment
CN112910960A (en) * 2021-01-15 2021-06-04 郑州轻工业大学 Virtual network online migration method and device with time delay, resource and energy consumption perception

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
ROBERTO MORABITO: "Virtualization on Internet of Things Edge Devices With Container Technologies: A Performance Evaluation", 《IEEE ACCESS 》 *
吕元琛: "容器云环境下容器调度策略的研究与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *

Also Published As

Publication number Publication date
CN114077498B (en) 2023-03-28

Similar Documents

Publication Publication Date Title
Wang et al. Towards efficient scheduling of federated mobile devices under computational and statistical heterogeneity
CN111190688B (en) Cloud data center-oriented Docker migration method and system
Huang et al. Stochastic configuration networks based adaptive storage replica management for power big data processing
RU2684416C2 (en) Loads scheduling in multi-battery devices
US9239740B2 (en) Program partitioning across client and cloud
US20080104608A1 (en) Starting up at least one virtual machine in a physical machine by a load balancer
CN109257429A (en) A kind of calculating unloading dispatching method based on deeply study
Raza et al. Autonomic performance prediction framework for data warehouse queries using lazy learning approach
CN113535409B (en) Server-free computing resource distribution system oriented to energy consumption optimization
EP3518118B1 (en) Non-homogenous synopsis for efficient partition pruning
CN109901932A (en) A kind of Server Consolidation method based on virtual machine
Khelifa et al. Combining task scheduling and data replication for SLA compliance and enhancement of provider profit in clouds
CA3161221A1 (en) System for monitoring and optimizing computing resource usage of cloud based computing application
CN114077498B (en) Method and system for selecting and transferring calculation load facing to mobile edge calculation
CN108241526B (en) Adaptive control strategy based on mode prediction
CN112148483A (en) Container migration method and related device
WO2024039549A1 (en) Energy based task shifting
Huang et al. Load balancing mechanism based on linear regression partition prediction in spark
CN106227600B (en) A kind of multidimensional virtual resource allocation method based on Energy-aware
Monshizadeh Naeen et al. Cloud data center cost management using virtual machine consolidation with an improved artificial feeding birds algorithm
CN115686160A (en) Chip operation method and device, SOC chip and storage medium
Bakhthemmat et al. Solving fully dynamic bin packing problem for virtual machine allocation in the cloud environment by the futuristic greedy algorithm
El Motaki et al. A prediction-based model for virtual machine live migration monitoring in a cloud datacenter
CN107967170A (en) Migrate computational methods, device, equipment and storage medium
Bahrami et al. The Reduce energy consumption in cloud computing data centres by optimizing virtual machines

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant