CN112104737B - Calculation migration method, mobile computing equipment and edge computing equipment - Google Patents
Calculation migration method, mobile computing equipment and edge computing equipment Download PDFInfo
- Publication number
- CN112104737B CN112104737B CN202010980258.8A CN202010980258A CN112104737B CN 112104737 B CN112104737 B CN 112104737B CN 202010980258 A CN202010980258 A CN 202010980258A CN 112104737 B CN112104737 B CN 112104737B
- Authority
- CN
- China
- Prior art keywords
- task
- computing node
- delay
- node
- edge
- 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.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
-
- 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
- G06F9/5088—Techniques for rebalancing the load in a distributed system involving task migration
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1004—Server selection for load balancing
- H04L67/1008—Server selection for load balancing based on parameters of servers, e.g. available memory or workload
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1004—Server selection for load balancing
- H04L67/101—Server selection for load balancing based on network conditions
-
- 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)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention provides a computation migration method, mobile computing equipment and edge computing equipment, wherein the method comprises the following steps: the mobile computing node receives the computing task and reads task information, and when the computing migration delay is within the maximum delay tolerable by the task, whether the local computing delay is within the maximum delay is judged; if yes, further judging whether the local calculation energy consumption is not greater than the transmission energy consumption, and if the local calculation energy consumption is less than or equal to the transmission energy consumption, performing calculation processing on the task locally by the mobile computing node; otherwise, the mobile computing node migrates the task to the edge computing node, and the edge computing node performs computing processing on the task. The invention considers the tolerable maximum delay of the calculation task, further reduces the task processing energy consumption of the mobile calculation node under the condition of meeting the maximum delay, not only provides strict delay guarantee of task granularity for the calculation task, but also reduces the processing energy consumption of the mobile calculation node and prolongs the life span of the mobile calculation node.
Description
Technical Field
The present invention relates to the field of communications technologies, and in particular, to a computation migration method, a mobile computing device, and an edge computing device.
Background
With the wide application of the internet of things in the fields of industry, agriculture, automatic driving and the like, the internet of things terminal and the intelligent terminal generate a huge amount of data at the edge of a network. In order to obtain the maximum value of the data, the data needs to be quickly calculated and processed. In particular, many internet of things data come from applications that have very stringent delay requirements, such as virtual reality, autonomous driving, and so on. Although cloud computing technology has enjoyed considerable success in providing high performance computing services for internet of things applications, cloud computing has difficulty meeting the ultra-low latency requirements of internet of things applications. The cloud computing resources are often far away from a data source, and a huge amount of data is transmitted to a remote cloud computing data center through a network, so that not only are a large amount of network bandwidth resources consumed and transmission energy consumed, but also the data is subjected to long network delay, and the ultra-low delay requirement of the application cannot be met.
Edge computing has been considered in recent years as an efficient computational paradigm for reducing network transmission delays for delay-sensitive applications. By deploying the computing nodes at the network edge, computing tasks of the internet of things terminal and the intelligent terminal are migrated to the network edge, network transmission delay and network equipment energy consumption can be reduced, and user experience of internet of things application is greatly improved.
Compute migration is an edge compute critical technique that considers when and in what way to migrate a compute task to which compute node. The existing computation migration method mainly reduces network delay by optimizing and configuring computation tasks between a remote cloud computing server and network edge computing nodes and between distributed homogeneous network edge nodes, and does not consider computation migration conditions between heterogeneous computing nodes under the same access network. The computing migration of the cooperation of heterogeneous edge computing nodes faces a severe technical challenge due to the heterogeneous computing capability of the edge nodes, the limited energy consumption of the mobile computing nodes, and the data volume and high dynamics of the application of the internet of things.
Disclosure of Invention
The invention aims to provide a computation migration method, mobile computing equipment and edge computing equipment, and aims to solve the technical problems of large delay and high energy consumption of mobile computing nodes in the process of processing a computation task.
The purpose of the invention can be realized by the following technical scheme:
a compute migration method comprising:
s1: a mobile computing node receives a computing task submitted by a user and reads task information of the task; wherein the user runs at least one mobile application; the task information comprises a task type, a task size and a maximum delay tolerable by the task;
s2: the mobile computing node judges whether the calculated migration delay exceeds the maximum delay, if so, the step S5 is executed, otherwise, the step S3 is executed; wherein the calculation migration delay is the sum of the wireless transmission delay of the task and the calculation delay of the task at the edge calculation node;
s3: the mobile computing node determining whether a local computing latency exceeds the maximum latency, if so, performing S6, otherwise performing S4;
s4: the mobile computing node judges whether the local computing energy consumption is not greater than the transmission energy consumption, if so, S5 is executed; otherwise, go to S6;
s5: the mobile computing node performs computing processing on the task and returns a processing result to the user;
s6: and the mobile computing node migrates the task to an edge computing node, and the edge computing node performs computing processing on the task and returns a processing result to the user.
Optionally, the delay d is calculated locally l The method specifically comprises the following steps: d l ＝(Q+S)/f l ；
Q is the size of a data block waiting for calculation in a cache queue of the mobile computing node, S is the size of the data block of the task, f l The calculated rate is the mobile computing node.
Optionally, local computing energy consumption E l The method specifically comprises the following steps: e l ＝α×S×f l ×f l ；
Where α is the locally calculated energy consumption factor.
Optionally, theRadio transmission delay d c Comprises the following steps: d c W/R; wherein, W is the data size of the task with bits as basic unit, and R is the wireless transmission rate.
Optionally, the wireless transmission rate R is specifically:
where B denotes the subcarrier bandwidth, N denotes the number of radio transmission subcarriers used by the mobile computing node, P n Representing the transmit power of the mobile computing node on subcarrier N (1. ltoreq. N. ltoreq.N), g n Representing the channel gain to noise ratio of the mobile computing node on subcarrier n.
Optionally, the computation delay d of the task at the edge computation node r The method comprises the following specific steps: d r ＝S/f r ；
Wherein f is r And calculating the maximum calculation rate available at the current moment for the edge calculation node.
Optionally, the transmission energy consumption is specifically: the transmission energy Ec is specifically:
optionally, the specific process of obtaining the maximum computation rate available at the current time of the edge computation node is:
the mobile computing node sends a message to the edge computing node to request to inquire the maximum computing rate available at the current moment of the edge computing node;
the edge computing node determines the maximum computing rate available at the current moment;
the edge computing node returns the maximum computing rate available at the current moment to the mobile computing node;
and the mobile computing node receives the maximum computing rate available at the current moment returned by the edge computing node.
The present invention also provides a mobile computing device comprising:
a processor, and a memory communicatively coupled to the processor;
wherein the memory has stored thereon instructions executable on the processor, the instructions when executed by the processor implementing the steps performed by the mobile computing node in the compute migration method.
The present invention also provides an edge computing device, comprising:
a processor, and a memory communicatively coupled to the processor;
the processor and the memory communicatively coupled to the processor can be virtualized into one or more virtual machines, the sum of the computational rates of all virtual machines running simultaneously on the device not exceeding the maximum rate of the processor; the memory has stored thereon instructions executable on the processor, the instructions when executed by the processor implementing the steps performed by the edge compute node in the compute migration method.
The invention provides a calculation migration method, mobile computing equipment and edge computing equipment, wherein the calculation migration method comprises the following steps: s1: a mobile computing node receives a computing task submitted by a user and reads task information of the task; wherein the user runs at least one mobile application; the task information comprises a task type, a task size and a maximum delay tolerable by the task; s2: the mobile computing node judges whether the calculated migration delay exceeds the maximum delay, if so, the step S5 is executed, otherwise, the step S3 is executed; wherein the calculation migration delay is the sum of the wireless transmission delay of the task and the calculation delay of the task at the edge calculation node; s3: the mobile computing node determining whether a local computing latency exceeds the maximum latency, if so, performing S6, otherwise performing S4; s4: the mobile computing node judges whether the local computing energy consumption is not greater than the transmission energy consumption, if so, S5 is executed; otherwise, go to S6; s5: the mobile computing node performs computing processing on the task and returns a processing result to the user; s6: and the mobile computing node migrates the task to an edge computing node, and the edge computing node performs computing processing on the task and returns a processing result to the user.
The computation migration method, the mobile computing device and the edge computing device provided by the invention consider the tolerable maximum delay of each requested computation task, further reduce the task processing energy consumption of the mobile computing node under the condition of meeting the maximum tolerable delay, provide strict delay guarantee of task granularity for the computation task, reduce the processing energy consumption of the mobile computing node and prolong the life span of the mobile computing node. According to the invention, the mobile computing node requests the usable maximum computing rate from the edge computing node, so that the computing delay of the migrated task at the edge computing node can be accurately known, the mode of acquiring the computing delay of the task at the edge computing node is simplified, and the rate and the efficiency of the computing migration decision process are improved.
Drawings
FIG. 1 is a schematic flow chart of a computation migration method according to the present invention;
fig. 2 is a schematic diagram of a query process of a maximum computation rate of an edge compute node in the compute migration method provided by the present invention;
FIG. 3 is a schematic structural diagram of a compute migration system to which the compute migration method provided by the present invention is applied;
FIG. 4 is a first schematic diagram of a compute migration apparatus corresponding to the compute migration method provided by the present invention;
FIG. 5 is a second schematic diagram of a compute migration apparatus corresponding to the compute migration method provided by the present invention;
FIG. 6 is a schematic diagram of a mobile computing device provided by the present invention;
FIG. 7 is a schematic structural diagram of an edge computing device provided by the present invention;
FIG. 8 is a graph comparing the energy consumption of a computing migration method provided by the present invention with the energy consumption of a mobile computing node for all local computing algorithms and all migration algorithms;
FIG. 9 is a graph comparing the mean delay of the compute migration method provided by the present invention with all local compute algorithms and all migration algorithms;
fig. 10 is a graph comparing the delay guarantee rates of the calculation migration method provided by the present invention with all local calculation algorithms and all migration algorithms.
Detailed Description
Embodiments of the present invention provide a computation migration method, a mobile computing device, and an edge computing device, so as to solve the technical problems of a large delay and a high energy consumption of a mobile computing node when processing a computation task.
To facilitate an understanding of the invention, the invention will now be described more fully with reference to the accompanying drawings. Preferred embodiments of the present invention are shown in the drawings. This invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.
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 terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.
Compute migration is an edge compute critical technique that considers when and in what way to migrate a compute task to which compute node. The existing computation migration method mainly reduces network delay by optimizing and configuring computation tasks between a remote cloud computing server and network edge computing nodes and between distributed homogeneous network edge nodes, and does not consider computation migration conditions between heterogeneous computing nodes in the same access network. In fact, with the increase of the computing power of the intelligent terminal, the intelligent terminal can also be used as a mobile computing node to perform computing tasks. Thus, an edge computing system may be comprised of multiple heterogeneous mobile computing nodes and edge computing server nodes.
Referring to fig. 1, the following is an embodiment of a method for computing migration according to the present invention, including:
s101: a mobile computing node receives a computing task submitted by a user and reads task information of the task; wherein the user runs at least one mobile application; the task information comprises a task type, a task size and a maximum delay tolerable by the task;
s102: the mobile computing node judges whether the computation migration delay exceeds the maximum delay, if so, S105 is executed, otherwise, S103 is executed; the calculation migration delay is the sum of the wireless transmission delay of the task and the calculation delay of the task at the edge calculation node;
s103: the mobile computing node judges whether the local computing delay exceeds the maximum delay, if so, S106 is executed, otherwise, S104 is executed;
s104: the mobile computing node judges whether the local computing energy consumption is not greater than the transmission energy consumption, if so, S105 is executed; otherwise, executing S106;
s105: the mobile computing node performs computing processing on the task and returns a processing result to the user;
s106: and the mobile computing node migrates the task to an edge computing node, and the edge computing node performs computing processing on the task and returns a processing result to the user.
In step S101, the mobile computing node receives a computing task submitted by a user running a mobile application, reads task information including a task type M, a data size S of the task using a data block as a basic unit, a data size W of the task using a bit as a basic unit, and a maximum tolerable delay D of the task, and proceeds to step S102.
In step S102, the mobile computing node determines whether the computation migration delay exceeds a maximum delay tolerable by the task. Let df denote the calculated migration delay, then d f By the formula d f ＝d c +d r Obtaining, where dc is the wireless transmission delay of the task, by formula d c Determined as W/R. Wherein, W is the data size of the task with bits as basic units, and R is the wireless transmission rate.
The radio transmission rate R is expressed as:
wherein B is the sub-carrier bandwidth, the natural number N represents the number of wireless transmission sub-carriers used by said mobile computing node, P n Representing the transmission power, g, of the mobile computing node on subcarrier N (1. ltoreq. N. ltoreq.N) n Representing the channel gain to noise ratio of the mobile computing node on subcarrier n.
It is worth noting that the transmission power of the mobile computing node on the used wireless transmission subcarriers needs to satisfy the following two conditions at the same time:
(1) the sum of the transmit powers of the mobile computing nodes on the used wireless transmission subcarriers does not exceed its maximum transmit power;
(2) the transmission power of the mobile computing node on any used wireless transmission sub-carrier is greater than zero, that is, the transmission power of the mobile computing node on the used wireless transmission sub-carrier needs to satisfy the following equation:
wherein P is a maximum transmit power of the mobile node; d is a radical of r Is the computation delay of the task at the edge compute node, d r Is given by the formula d r ＝S/f r And (4) determining.
Formula d r ＝S/f r In, f r Calculating the maximum available calculation rate of the node at the current moment for the edge; the maximum calculation rate available at the current moment of the edge calculation node is obtained through the maximum calculation rate query process of the edge calculation node. If d is satisfied f D, calculating that the migration delay does not exceed the maximum delay tolerable by the task, and executing the step S103; otherwise, step S105 is performed.
In step S103, the mobile computing node determines whether the local computing delay exceeds a maximum delay tolerable by the task. Let d l Representing localityCalculating the delay, d l Is given by the formula d l ＝(Q+S)/f l Determining; wherein Q is the size of the data block waiting for calculation in the buffer queue of the mobile computing node, f i A computed rate for the mobile computing node. If d is satisfied l D, if the local computation delay does not exceed the maximum delay tolerable by the task, the step S104 is proceeded to; otherwise, the mobile computing node migrates the task to an edge computing node, and executes step S106.
It is understood that the mobile computing node migrates the task to the edge computing node, i.e. the mobile computing node sends the task to the edge computing node through subcarriers 1 to N at the transmit power of step S102 or step S104.
Step S104: the mobile computing node determines whether the local computing energy consumption is not greater than the transmission energy consumption. Let E i Representing local computational energy consumption, by E l ＝α×S×f l ×f l Determining, wherein α is a locally calculated energy consumption factor.
Let E o Representing the transmission energy consumption, which is the product of the sum of the transmit powers of the sub-carriers used by the mobile computing node for transmitting the task and the radio transmission delay of the task, i.e.,
In step S105, the mobile computing node performs computing processing on the task to obtain a processing result, and the mobile computing node, which is the computing node executing the computing processing, returns the processing result to the user.
In step S106, the mobile computing node migrates the task to an edge computing node, the edge computing node performs computing processing on the task to obtain a processing result, and the computing node that performs the computing processing, that is, the edge computing node, returns the processing result to the user.
Referring to fig. 2, a specific process of obtaining the maximum computation rate available at the current time of the edge computing node is as follows:
step S201: the mobile computing node sends a message to the edge computing node to request to inquire the maximum computing rate available at the current moment of the edge computing node;
step S202: the edge computing node determines the maximum computing rate available at the current moment;
step S203: the edge computing node returns the maximum computing rate available at the current moment to the mobile computing node;
step S204: and the mobile computing node receives the maximum computing rate available at the current moment returned by the edge computing node.
In this embodiment, the specific process of the mobile computing node querying the maximum computation rate of the edge computing node is as follows: mobile computing node sends request (i.e., send) to edge computing node<Request>Information) inquiring the maximum computation rate available for the edge computing node; after receiving the request of the mobile computing node, the edge computing node determines the available maximum computing rate at the current moment; the edge computing node responds (i.e., sends) to the mobile computing node<Response>Message) to return the maximum computation rate f available at its current moment k (ii) a And the mobile computing node receives the maximum computing rate available at the current moment returned by the edge computing node.
Specifically, the step of determining the maximum computation rate available at the current time by the edge computation node specifically includes the following steps:
first, the edge computing node determines that a set of virtual machines VM ═ V that can run at the current time is { V ═ V } 1 ，V 2 ，...，V m ，...，V M And setting the calculation rate corresponding to the virtual machine set as f ═ f 1 ，f 2 ，...，f m ，...，f M In which f m Representing virtual machine V m And (M is more than or equal to 1 and less than or equal to M), the sum of the computing rates of the virtual machine set which can run at the current moment meets the following conditions:
wherein，f CPU Calculating a rate for the CPU of the edge compute node;
secondly, selecting a virtual machine set which can run at the current moment and satisfies f k A virtual machine Vk of max f (1 ≦ k ≦ M), where max (·) represents a maximum function, then virtual machine V k Having a maximum calculation rate, calculating its calculation rate f k As the maximum computation rate available at the current time of the edge computing node.
Referring to fig. 3, in general, consider that there are several mobile computing nodes and an edge computing node in a specific wireless network, the mobile computing nodes are randomly distributed in the coverage area of the wireless network, the edge computing node is located in the core of the wireless network, that is, the edge computing node is located near a base station, and both the mobile computing nodes and the edge computing node have computing processing capabilities. In general, the computing processing power of the edge computing node is higher than that of the mobile computing node, which is physically closer to the user running the mobile application.
In this embodiment, in order to meet the delay requirement of the computing task requested by the user, the mobile computing node locally processes the requested computing task or migrates to the edge computing node for computing according to its own computing resource state, wireless network state, and computing resource state of the edge computing node.
Referring to fig. 1 to 3, when any one of the mobile computing nodes receives a computing task request submitted by a user running a mobile application, the mobile computing node performs the following steps in the computing migration method: reading task information, when the calculation migration delay is within the maximum delay range tolerable by the task, further judging whether the local calculation delay is within the maximum delay range tolerable by the task, and when the local calculation delay is also within the maximum delay range tolerable by the task, the local calculation and the calculation migrated to the edge calculation node can both meet the maximum delay requirement of the task.
Then, further judging whether the local computing energy consumption is not greater than the transmission energy consumption, when the local computing energy consumption is less than or equal to the transmission energy consumption, the mobile computing node determines to locally perform computing processing on the task, and the mobile computing node puts the task into a local CPU computing queue; otherwise, the mobile computing node decides to migrate the task to an edge computing node, the mobile computing node sends the task to the edge computing node through the wireless base station, and the edge computing node schedules a virtual machine to process the task.
It will be appreciated that when the computational migration delay does not meet the delay requirement of the task, the mobile computing node decides to process the task locally; and when the calculation migration delay can meet the delay requirement of the task and the local calculation delay cannot meet the delay requirement of the task, the mobile computing node migrates the task to the edge computing node for processing.
The calculation migration method provided by the embodiment of the invention considers the tolerable maximum delay of each requested calculation task, further reduces the task processing energy consumption of the mobile calculation node under the condition of meeting the maximum tolerable delay, not only provides strict delay guarantee of task granularity for the calculation task, but also reduces the processing energy consumption of the mobile calculation node and prolongs the life span of the mobile calculation node. According to the invention, the mobile computing node requests the usable maximum computing rate from the edge computing node, so that the computing delay of the migrated task at the edge computing node can be accurately known, the mode of acquiring the computing delay of the task at the edge computing node is simplified, and the rate and the efficiency of the computing migration decision process are improved.
Referring to fig. 4, a computing migration apparatus provided by the present invention may be disposed on a mobile computing node, the apparatus including:
the computing task sensing module 11 is configured to receive a task submitted by a user running a mobile application program, and acquire task information;
a local computation delay and energy consumption perception module 12, configured to compute local computation delay experienced by the task in local computation and generated local computation energy consumption;
a wireless transmission rate and power allocation module 13, configured to obtain a wireless transmission rate, a transmission power of the apparatus on a used wireless transmission subcarrier, a wireless transmission delay, and transmission energy consumption;
a migration computation delay sensing module 14, configured to initiate a maximum computation rate query process for an edge computing node, obtain a maximum computation rate available for the edge computing node, and obtain a computation delay of the task at the edge computing node;
a computation migration decision module 15, configured to determine whether to locally process the requested computation task or to migrate to an edge compute node for computation;
the local computation scheduling module 16 is used for locally processing the computation tasks and returning processing results;
and the calculation migration execution module 17 migrates the task to the edge calculation node.
The working process of the device is as follows: the calculation task sensing module 11 receives a calculation task request submitted by a user, reads task information, and transmits the information to the calculation migration decision module 15. The calculation migration decision module 15 obtains the local calculation delay and the local calculation energy consumption from the local calculation delay and energy consumption sensing module 12, obtains the transmission delay from the wireless transmission rate and power allocation module 13, and obtains the calculation delay of the task at the edge calculation node from the migration calculation delay sensing module 14; then, it is decided whether to locally process the requested computing task or to migrate to an edge computing node for computing processing. If the calculation task is processed locally, the task is sent to a local calculation scheduling module 16 for processing; if the task is migrated to the edge computing node, the task is sent to the edge computing node through the computation migration execution module 17.
Specifically, the computation migration decision module 15 sends a message to the migration computation delay sensing module 14 to obtain the computation delay of the edge computation node, and after the migration computation delay sensing module 14 receives the message sent by the computation migration decision module 15, the migration computation delay sensing module initiates a maximum computation rate query process of the edge computation node to obtain the maximum computation rate available for the edge computation node, and then computes the computation delay of the task at the edge computation node, and returns the computation delay to the computation migration decision module 15.
Referring to fig. 5, the computation migration apparatus provided in the present invention may be disposed on an edge compute node, and the apparatus includes:
a computation rate query response module 31, configured to respond to a query of a maximum computation rate available to the edge computing node requested by the mobile computing node, and return the maximum computation rate available to the edge computing node at the current time;
the resource management module 32 is configured to monitor usage of computing resources of the edge computing node, and determine a set of virtual machines that can be run at the current time;
a calculation migration task sensing module 33, configured to sense a calculation task migrated from a mobile computing node;
a computation migration module 34, configured to determine how to perform a computation task migrated from the mobile computing node;
and the scheduling module 35 is used for processing the migrated computing tasks.
The working process of the device is as follows: when the computation rate query response module 31 receives an edge computing node maximum available computation rate query request sent by a mobile computing node, the computation rate query response module 31 queries a currently operable virtual machine set to the resource management module 32, the resource management module 32 returns the currently operable virtual machine set and a corresponding computation rate to the computation rate query response module 31, and the computation rate query response module 31 selects a virtual machine with a maximum computation rate from the currently operable virtual machine set, and returns the computation rate of the virtual machine as the maximum computation rate available at the current time of the edge computing node to the mobile computing node. When the calculation migration task sensing module 33 receives a calculation task sent by a mobile computing node, the task information is sent to the calculation migration module 34, the calculation migration module 34 obtains a set of virtual machines that can be operated at the current time from the resource management module 32, then selects one virtual machine from the set, and sends the information of the virtual machine and the information of the task to the scheduling module 35. After receiving the virtual machine information and the computation task information of the computation migration module 34, the scheduling module 35 enables the virtual machine to execute the computation task, and returns a processing result.
Referring to fig. 6, the present invention further provides a mobile computing device, including:
a processor, and a memory communicatively coupled to the processor;
wherein the memory has stored thereon instructions executable on the processor, the instructions when executed by the processor implementing the steps performed by the mobile computing node in the compute migration method.
In particular, the apparatus may be provided on a mobile computing node, the apparatus comprising: a processor 21, and a memory 22 communicatively coupled to the processor; the memory 22 has stored thereon a compute migration program operable on the processor 21, which when executed by the processor 21 implements the steps performed by the mobile computing node in the compute migration method.
Referring to fig. 7, the present invention further provides an edge computing apparatus, including:
a processor, and a memory communicatively coupled to the processor;
the processor and the memory communicatively coupled to the processor can be virtualized into one or more virtual machines, the sum of the computational rates of all virtual machines running simultaneously on the device not exceeding the maximum rate of the processor; the memory has stored thereon instructions executable on the processor, the instructions when executed by the processor implementing the steps performed by the edge compute node in the compute migration method.
Specifically, the apparatus may be disposed on an edge computing node, and the apparatus includes: a processor 41, and a memory 42 coupled to the processor; the processor 41 and the processor-coupled memory 42 can be virtualized as one or more virtual machines, the sum of the computing rates of all virtual machines running simultaneously on the device not exceeding the maximum rate of the processor 41; the memory 42 has stored thereon a compute migration program operable on the processor 41; the compute migration program when executed by the processor 41 will implement the steps of the edge compute nodes in the compute migration methods of embodiments 1 and 2 as described above.
The effect of the present invention can be further illustrated by the following simulation results:
1. simulation conditions
Matlab is adopted to evaluate the delay guarantee of the calculation migration method and the energy consumption performance of the mobile calculation node. Adopting a calculation migration system model shown in fig. 3, wherein the system is provided with 4 local calculation nodes and 1 edge calculation node, the calculation rates of the local calculation nodes are all fl ═ 0.6GHz, and the CPU calculation rate fcpu of the edge calculation node is 2 GHz; the mobile computing nodes are communicated with the edge computing nodes through a wireless network, the maximum transmitting power of each mobile computing node is 0.1W, the number of available subcarriers is 16, and the bandwidth of each subcarrier is 0.3125 MHz; setting a local computing energy consumption factor of the mobile computing node to be 10-28; of said mobile computing nodes, 3 mobile computing nodes receive tasks having an average size of 250KB in bit basis, an average data block size of 475MHz, and a maximum tolerable delay of 1 ms; the average size of the tasks received by the other mobile computing node in basic units of bits is 420KB, the average data block size is 798MHz, and the maximum delay tolerable for the task is 10 ms.
Simulation effects the abscissa task request rate in fig. 8-10 is the average number of tasks of computing requests received per millisecond by each mobile computing node; simulation results all of the local computing algorithms described in FIGS. 8-10 process at each mobile computing node the computing task request received by that node; the all-migration algorithm is that all the mobile computing nodes migrate the received computing tasks to the edge computing nodes for processing.
2. Comparison of simulation results
Fig. 8 is a comparison graph of the energy consumption of the mobile node of the computation migration method provided by the present invention and all local computation algorithms and all migration algorithms, and as shown in fig. 8, under various task request rates, the energy consumption of the mobile node of the computation migration method provided by the present invention is lower than the energy consumption of all local computation algorithms, but higher than the energy consumption of all migration algorithms.
FIG. 9 is a comparison graph of the mean delay of the computation migration method provided by the present invention and all local computation algorithms and all migration algorithms, as shown in FIG. 9, the mean delay of the computation migration method provided by the present invention is lower than the mean delay of all local computation algorithms under various task request rates; with the rise of the task request rate, the adoption of all the migration algorithms can lead the available resources of the edge computing nodes to be less and less, so that the average delay of the algorithm rises rapidly with the rise of the task request rate; and the calculation migration algorithm provided by the invention can still keep lower average delay under the condition that the task request rate rises, and particularly, when the task request rate is more than 0.7tasks/ms, the calculation migration method provided by the invention can still provide average delay of less than 1.5ms, and the delay guarantee effect is very obvious.
Fig. 10 is a diagram comparing the delay guarantee rates of the computation migration method, all local computation algorithms and all migration algorithms provided in the present invention, where the delay guarantee rate is the ratio of the number of tasks whose delay is guaranteed to the number of total tasks that arrive at the system. As shown in fig. 10, under various task request rates, the delay guarantee rates of the computation migration method provided by the present invention are all higher than the delay guarantee rates of all local computation algorithms, and therefore are significantly better than all local computation algorithms; under the condition of a low task request rate, the computing resources of the edge nodes are sufficient, and the delay of all the requested tasks can be guaranteed by transferring all the computing tasks to the edge nodes.
Therefore, under the condition of lower task request rate, the delay guarantee rate of all migration algorithms is higher than that of the calculation migration method provided by the invention; with the increase of the task request rate, all the migration algorithms are adopted, so that the available resources of the edge nodes are less and less, and even the congestion of the edge nodes is caused, and therefore, with the increase of the task request rate, the delay guarantee rate of all the migration algorithms is rapidly reduced.
The delay guarantee rate of the computational migration method provided by the invention is slowly reduced along with the rise of the task request rate, and particularly, when the task request rate is more than 0.7tasks/ms, the computational migration method provided by the invention still can provide a delay guarantee rate higher than 85%, so that the computational migration method provided by the invention has obvious effects on the aspects of delay guarantee stability and delay guarantee rate.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In the embodiments provided in the present application, it should be understood that the disclosed system, apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit may be implemented in the form of hardware, or may also be implemented in the form of a software functional unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk, and various media capable of storing program codes.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (4)
1. A method of computing migration, comprising:
s1: a mobile computing node receives a computing task submitted by a user and reads task information of the task; wherein the user runs at least one mobile application, and the task information comprises a task type, a task size and a maximum delay tolerable for the task;
s2: the mobile computing node judges whether the calculated migration delay exceeds the maximum delay, if so, the step S5 is executed, otherwise, the step S3 is executed; wherein the calculation migration delay is the sum of the wireless transmission delay of the task and the calculation delay of the task at the edge calculation node;
wireless transmission delay d of said task c Comprises the following steps: d c W/R; wherein W is data of the task with bit as basic unitThe size, R, is the wireless transmission rate,
b denotes the sub-carrier bandwidth, N denotes the number of radio transmission sub-carriers used by the mobile computing node, P n Representing the transmission power of the mobile computing node on subcarrier n, g n Representing the channel gain noise ratio of the mobile computing node on a subcarrier N, wherein N is more than or equal to 1 and less than or equal to N;
computing delay d of the task at the edge computing node r Comprises the following steps: d is a radical of r ＝S/f r (ii) a Wherein S is the data block size of the task, f r Calculating the maximum available calculation rate of the node at the current moment for the edge;
s3: the mobile computing node judges whether the local computing delay exceeds the maximum delay, if so, executes S6, otherwise executes S4;
the local computation delay d l Comprises the following steps: d l ＝(Q+S)/f l (ii) a Q is the size of a data block waiting for calculation in a cache queue of the mobile computing node, S is the size of the data block of the task, f l Computing a rate for the mobile computing node;
s4: the mobile computing node judges whether the local computing energy consumption is not greater than the transmission energy consumption, if so, S5 is executed; otherwise, go to S6;
said local computing energy consumption E l Comprises the following steps: e l ＝α×S×f l ×f l (ii) a Wherein α is a local computational energy consumption factor;
the transmission energy Ec is:
s5: the mobile computing node performs computing processing on the task and returns a processing result to the user;
s6: and the mobile computing node migrates the task to an edge computing node, and the edge computing node performs computing processing on the task and returns a processing result to the user.
2. The computation migration method according to claim 1, wherein the specific process of obtaining the maximum computation rate available at the current time of the edge computation node is:
the mobile computing node sends a message to the edge computing node to request to inquire the maximum computing rate available at the current moment of the edge computing node;
the edge computing node determines the maximum computing rate available at the current moment;
the edge computing node returns the maximum computing rate available at the current moment to the mobile computing node;
and the mobile computing node receives the maximum computing rate available at the current moment returned by the edge computing node.
3. A mobile computing device, comprising:
a processor, and a memory communicatively coupled to the processor;
wherein the memory has stored thereon instructions executable on the processor, the instructions when executed by the processor implementing the steps performed by the mobile computing node in the compute migration method of claim 1 or 2.
4. An edge computing device, comprising:
a processor, and a memory communicatively coupled to the processor;
the processor and the memory communicatively coupled to the processor can be virtualized into one or more virtual machines, the sum of the computational rates of all virtual machines running simultaneously on the device not exceeding the maximum rate of the processor; the memory has stored thereon instructions executable on the processor, the instructions when executed by the processor implementing the steps performed by the edge compute node in the compute migration method of claim 1 or 2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010980258.8A CN112104737B (en) | 2020-09-17 | 2020-09-17 | Calculation migration method, mobile computing equipment and edge computing equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010980258.8A CN112104737B (en) | 2020-09-17 | 2020-09-17 | Calculation migration method, mobile computing equipment and edge computing equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112104737A CN112104737A (en) | 2020-12-18 |
| CN112104737B true CN112104737B (en) | 2022-08-30 |
Family
ID=73760232
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010980258.8A Active CN112104737B (en) | 2020-09-17 | 2020-09-17 | Calculation migration method, mobile computing equipment and edge computing equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112104737B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113241854B (en) * | 2021-06-10 | 2022-12-06 | 南方电网科学研究院有限责任公司 | Remote updating method and system for power utilization state of user |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109413676A (en) * | 2018-12-11 | 2019-03-01 | 西北大学 | Combine the edge calculations moving method of lower uplink in a kind of ultra dense heterogeneous network |
| WO2019204898A1 (en) * | 2018-04-26 | 2019-10-31 | 10518590 Canada Inc. | Workload scheduling in a distributed computing environment based on an applied computational value |
| EP3605329A1 (en) * | 2018-07-31 | 2020-02-05 | Commissariat à l'énergie atomique et aux énergies alternatives | Connected cache empowered edge cloud computing offloading |
| CN111491332A (en) * | 2020-04-20 | 2020-08-04 | 中国石油大学(北京) | Dynamic service migration method and device |
-
2020
- 2020-09-17 CN CN202010980258.8A patent/CN112104737B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019204898A1 (en) * | 2018-04-26 | 2019-10-31 | 10518590 Canada Inc. | Workload scheduling in a distributed computing environment based on an applied computational value |
| EP3605329A1 (en) * | 2018-07-31 | 2020-02-05 | Commissariat à l'énergie atomique et aux énergies alternatives | Connected cache empowered edge cloud computing offloading |
| CN109413676A (en) * | 2018-12-11 | 2019-03-01 | 西北大学 | Combine the edge calculations moving method of lower uplink in a kind of ultra dense heterogeneous network |
| CN111491332A (en) * | 2020-04-20 | 2020-08-04 | 中国石油大学(北京) | Dynamic service migration method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112104737A (en) | 2020-12-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110418416B (en) | Resource allocation method based on multi-agent reinforcement learning in mobile edge computing system | |
| Cheng et al. | Energy-efficient joint offloading and wireless resource allocation strategy in multi-MEC server systems | |
| Liu et al. | Latency and reliability-aware task offloading and resource allocation for mobile edge computing | |
| CN107995660B (en) | Joint task scheduling and resource allocation method supporting D2D-edge server unloading | |
| Mao et al. | Power-delay tradeoff in multi-user mobile-edge computing systems | |
| CN108809695B (en) | Distributed uplink unloading strategy facing mobile edge calculation | |
| Bozorgchenani et al. | Centralized and distributed architectures for energy and delay efficient fog network-based edge computing services | |
| CN109343904B (en) | Lyapunov optimization-based fog calculation dynamic unloading method | |
| CN112188551B (en) | Computation migration method, computation terminal equipment and edge server equipment | |
| CN109151864B (en) | Migration decision and resource optimal allocation method for mobile edge computing ultra-dense network | |
| Dai et al. | Joint offloading and resource allocation in vehicular edge computing and networks | |
| Nath et al. | Multi-user multi-channel computation offloading and resource allocation for mobile edge computing | |
| CN111511028B (en) | Multi-user resource allocation method, device, system and storage medium | |
| Messaoudi et al. | On using edge computing for computation offloading in mobile network | |
| CN111475274A (en) | Cloud collaborative multi-task scheduling method and device | |
| CN110149401B (en) | Method and system for optimizing edge calculation task | |
| CN114697333B (en) | Edge computing method for energy queue equalization | |
| CN114138373A (en) | Edge calculation task unloading method based on reinforcement learning | |
| CN111263401A (en) | Multi-user cooperative computing unloading method based on mobile edge computing | |
| CN114449529A (en) | Resource allocation optimization method and device based on mobile edge calculation and storage medium | |
| Shu et al. | Joint offloading strategy based on quantum particle swarm optimization for MEC-enabled vehicular networks | |
| CN110780986B (en) | Internet of things task scheduling method and system based on mobile edge computing | |
| CN112788698A (en) | Data processing method and device and terminal equipment | |
| CN112104737B (en) | Calculation migration method, mobile computing equipment and edge computing equipment | |
| Yang et al. | Optimal task scheduling in communication-constrained mobile edge computing systems for wireless virtual reality |
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 |