CN108429784B - Energy efficiency priority cloud resource allocation and scheduling method - Google Patents
Energy efficiency priority cloud resource allocation and scheduling method Download PDFInfo
- Publication number
- CN108429784B CN108429784B CN201810041238.7A CN201810041238A CN108429784B CN 108429784 B CN108429784 B CN 108429784B CN 201810041238 A CN201810041238 A CN 201810041238A CN 108429784 B CN108429784 B CN 108429784B
- Authority
- CN
- China
- Prior art keywords
- server
- virtual machine
- servers
- virtual machines
- migration
- 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.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 85
- 238000013468 resource allocation Methods 0.000 title claims abstract description 20
- 238000013508 migration Methods 0.000 claims abstract description 38
- 230000005012 migration Effects 0.000 claims abstract description 38
- 238000005265 energy consumption Methods 0.000 claims abstract description 28
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 21
- 238000005457 optimization Methods 0.000 claims description 24
- 230000008569 process Effects 0.000 claims description 20
- 238000004088 simulation Methods 0.000 abstract description 9
- 230000006870 function Effects 0.000 description 11
- 238000011156 evaluation Methods 0.000 description 8
- 238000007726 management method Methods 0.000 description 8
- 230000008447 perception Effects 0.000 description 6
- 230000006978 adaptation Effects 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 4
- 230000014509 gene expression Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000017525 heat dissipation Effects 0.000 description 2
- 238000012856 packing Methods 0.000 description 2
- 239000013598 vector Substances 0.000 description 2
- 230000007423 decrease Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/104—Peer-to-peer [P2P] networks
- H04L67/1074—Peer-to-peer [P2P] networks for supporting data block transmission mechanisms
-
- 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
- H04L67/50—Network services
- H04L67/60—Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources
- H04L67/63—Routing a service request depending on the request content or context
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Computer And Data Communications (AREA)
- Power Sources (AREA)
Abstract
The invention discloses a cloud resource allocation and scheduling method with priority on energy efficiency. The virtual machine allocation method reduces the number of working servers by using an extended boxing algorithm; the dynamic virtual machine migration method dynamically integrates the virtual machines into the minimum number of cloud servers by using a linear integer programming method, so that the number of working servers is further reduced, and the energy efficiency of a computing center is improved. The migration method and the distribution method work cooperatively, and the total energy consumption of the cloud computing center can be effectively reduced. Simulation research results show that the energy efficiency priority cloud resource allocation and scheduling method can effectively improve the energy efficiency of the cloud computing system.
Description
Technical Field
The invention belongs to the technical field of information, relates to allocation and scheduling of server resources in a cloud computing center, and particularly relates to a cloud resource allocation and scheduling method with priority on energy efficiency.
Background
In modern cloud computing centers, high energy consumption is a key issue for increasing operating costs. According to literature statistics, the cost of power and heat dissipation has increased 4 times over the last 10 years, and over 50% of cloud computing centers determine power and heat dissipation costs as key factors limiting server deployment. Therefore, it is imperative to reduce the energy consumption of cloud computing centers.
In fact, the cloud center consumes about 70% of the peak power when the idle servers consume power, and waste caused by the idle servers consuming power is considered to be a main cause of low energy efficiency. On the other hand, when partial virtual machines on the servers finish tasks, the running virtual machines are dynamically integrated into the minimum number of cloud servers, and then the idle servers are placed into a low-power-consumption mode, so that the energy efficiency of the cloud computing center can be improved. The energy-efficient cloud resource allocation and scheduling algorithm can reasonably allocate resources and dynamically adjust the active servers to reduce the number of working servers, and the idle servers are set to be in a low-power-consumption mode, so that the efficiency can be effectively improved, and the method is widely concerned by researchers.
At present, the resource allocation and scheduling method has a virtual machine allocation method with optimal server load and a virtual machine allocation method only adopting energy efficiency priority. The optimal server load virtual machine distribution method mainly considers the balance of server loads of a cloud computing center; the virtual machine allocation method that only prioritizes energy efficiency is limited to the allocation method that allocates virtual machines on servers with optimal energy efficiency, but does not perform integrated optimal scheduling on virtual machines carried by the servers after the virtual machines are finished, so the energy-saving effect of the method is not ideal.
Disclosure of Invention
The invention aims to provide a cloud resource allocation and scheduling method with priority on energy efficiency. The virtual machine allocation method improves the energy efficiency of the cloud computing center by reducing the number of working servers and setting idle servers to be in a sleep mode by using an extended boxing algorithm; the dynamic virtual machine migration method optimizes the servers finishing the service work by using a linear integer programming method, and further reduces the number of the working servers. The migration method and the distribution method work cooperatively, and the total energy consumption of the cloud computing center can be effectively reduced.
In order to realize the task, the invention adopts the following technical scheme:
a method for energy efficiency priority cloud resource allocation and scheduling comprises the following steps:
the method comprises the steps that firstly, a client packs an application program into a virtual machine, generates a virtual machine request and sends the virtual machine request to a cloud computing center; the cloud computing center establishes a linear constraint condition of the allocation process according to the virtual machine request;
the linear constraint conditions are as follows:
in the above formula, the two-dimensional variable xijRepresenting virtual machines VMiWhether to work on server j, if work on server xijIs set to 1, otherwise xijSetting to 0; n is the total number of the requested virtual machines; m is the total number of the cloud computing center servers; p is a radical ofiIs a virtual machine VMiMaximum power consumption of (1); x is the number ofijRepresenting virtual machines VMiWhether to work on server j; pj,maxIs the upper power limit for server j; e.g. of the typejIs a decision variable; pj,currentIs the current power of server j; e.g. of the typejFor decision variables, if server j provides a running environment for the virtual machine, then ejIs set to 1, otherwise ejSetting to 0;
secondly, the cloud computing center allocates the virtual machines to different servers according to the linear constraint conditions;
and step three, dynamically integrating the running virtual machines into the minimum number of servers according to the task completion condition after the virtual machines are distributed, and then putting the idle servers into a low power consumption mode or closing the idle servers.
Further, the specific process of the second step includes:
step 2.1, constructing an optimization objective function distributed by the virtual machine with priority on energy efficiency:
in the above formula, N represents the number of servers;
2.2, converting the virtual machine allocation problem into an optimization problem of multiple constraint conditions;
and 2.3, solving the optimization problem, and realizing reasonable distribution of the virtual machines according to the solution result.
Further, the multi-constraint optimization problem P described in step 2.2 is represented as:
in the above formula, piRepresenting virtual machines VMiMaximum power consumption of, xijRepresenting two-dimensional variables if virtual machine VMiWorking on server j, then xijIs set to 1, otherwise xijSetting to 0; n represents the total number of requesting virtual machines.
Further, the specific process of the third step includes:
step 3.1, deducing energy consumption which can be saved by migrating the virtual machine;
step 3.2, acquiring a target function of the virtual machine integrated migration;
step 3.3, obtaining the constrained conditions in the virtual machine migration process;
step 3.4, converting the problem of dynamically migrating and adjusting the virtual machine into an optimization problem with multiple constraint conditions;
and 3.5, solving the problem optimization problem by using an integer linear programming ILP algorithm.
Further, the energy consumption f that can be saved as shown in step 3.1 is expressed as:
in the above formula, m' represents the number of servers in a non-idle state, and m is satisfied' < m, wherein m is the total number of the cloud computing center servers; pi,idleIs the power consumption of the server in an idle state; y isiIs an identification of whether the server is idle or not; p'kPower consumption required to migrate virtual machine k; three-dimensional variable zijkRepresents the migration of virtual machine k from server i to server j; q. q.siRepresenting the number of virtual machines running on server i.
Further, the objective function in step 3.2 is to solve the maximum value of the energy consumption f in step 3.1.
Further, the constraints described in step 3.3 are expressed as follows:
in the above formula, zijkRepresents the migration of virtual machine k from server i to server j; z is a radical ofjlk'Represents the migration of virtual machine k' from server j to server l; m' represents the number of servers in a non-idle state; p is a radical ofkPower consumption required to migrate virtual machine k; pj,maxIs the upper power limit for server j; pj,currentIs the current power of server j; y isiIndicating whether the server is idle or not, if the server is in an idle state due to virtual machine migration, yiPut 1, otherwise yiSetting 0;
representing a lower limit on the number of servers in operation.
The utility model provides a cloud resource allocation and scheduling system that efficiency is preferred, includes virtual machine allocation management module, energy consumption evaluation module and the energy perception scheduler that connects gradually, wherein:
the virtual machine allocation management module is used for processing the request of the client and scheduling the virtual machine;
the energy consumption evaluation module is used for calculating the power consumption of the virtual machine and the server;
the energy-aware scheduler is used for placing a virtual machine in the cloud computing center.
Further, the energy-aware scheduler includes an allocation module and a migration module, wherein the allocation module is configured to place the virtual machine at an initial location, and the migration module minimizes the number of servers in use or in an active state by dynamically integrating the virtual machine.
Compared with the prior art, the invention has the following technical characteristics:
1. the invention designs a virtual machine allocation method with preferential energy efficiency based on an extended boxing algorithm, and solves the problem of extended boxing by adopting a space-limited online algorithm. The method comprehensively considers the first-time adaptation strategy and the optimal adaptation strategy, and has reasonable calculation complexity and good performance.
2. The virtual machine running on the server gradually leaves the system along with the completion of the task, the load of the server changes, and at the moment, the virtual machine allocation scheme initially designed according to the energy efficiency priority target cannot continuously meet the minimum energy consumption requirement. The dynamic virtual machine migration method designed based on the integer linear optimization algorithm can dynamically integrate the running virtual machines into the minimum number of cloud servers, so that the energy efficiency of the cloud computing center is further improved.
3. The invention adopts a dynamic virtual machine migration method and a virtual machine allocation method with priority on energy efficiency to work in a coordinated mode in stages, so that the total energy consumption of the cloud computing center can be effectively reduced.
Drawings
FIG. 1 is a system model designed by the present invention;
FIG. 2 is a flow chart of a space-constrained online algorithm for solving an extended binning problem;
FIG. 3 is a flowchart of a stealth enumeration method algorithm;
FIG. 4 is a graph comparing energy consumption of the process of the present invention with other processes;
FIG. 5 is a graph comparing the energy savings of the method of the present invention with other methods.
Detailed Description
The technical solution of the present invention is described in further detail below with reference to the accompanying drawings. The energy efficiency priority cloud resource allocation and scheduling method comprises the following steps:
the method comprises the steps that firstly, a client packs an application program into a virtual machine, generates a virtual machine request and sends the virtual machine request to a cloud computing center; the cloud computing center establishes a linear constraint condition of the allocation process according to the virtual machine request; the cloud computing center comprises a plurality of servers.
The linear constraint condition of the virtual machine request distribution process comprises that one server can only run a certain number of virtual machines at most; one virtual machine can only run on one server; the server can run more virtual machines only if the remaining resources are sufficient. The constraint mathematical expression is as follows:
(1) the cloud computing center admits the service request according to a service quality protocol agreed with a user in advance, arranges virtual machines for the service request and distributes each virtual machine to a server:
in the above formula, the two-dimensional variable xijRepresenting virtual machines VMiWhether to work on server j, if work on server xijIs set to 1, otherwise xijSetting to 0; n is the total number of the requested virtual machines; and m is the total number of the servers of the cloud computing center.
(2) The constraint relationship between the upper power limit of the server j and the current power is as follows:
in the above formula, piIs a virtual machine VMiThe maximum power consumption of (1) is an inherent parameter of the virtual machine; x is the number ofijRepresenting virtual machines VMiWhether to work on server j; pj,maxIs the upper power limit for server j; e.g. of the typejFor decision variables, if server j provides a running environment for the virtual machine, then ejIs set to 1, otherwise ejSetting to 0; that is, e is any virtual machine that runs on server jjIs set to 1, indicating a serverWhen j is in a normal working state and the server j is in a state without running the virtual machine, ejIs set to 0; pj,currentIs the current power of server j;
(3) the number of the virtual machines running on the server changes continuously, the power of the server changes correspondingly, and when P isj,max>Pj,currentAnd P isj,currentNot equal to 0, the lower limit of the number of servers in operation is
Because the number of servers in working operation needs to be greater than or equal to the lower limit of the number of servers, the following inequality is obtained:
secondly, the cloud computing center distributes the virtual machines to different servers according to the linear constraint conditions
In the invention, in order to better realize the distribution of the virtual machines, a virtual machine distribution method based on an expansion box algorithm of a plurality of constraint strips is provided. The theoretical basis of the virtual machine allocation algorithm of the invention is to expand the packing algorithm, and the aim is to pack some articles (virtual machines) into a group of boxes (servers for providing operating environment for the virtual machines) with power consumption as a characteristic parameter. The method comprises the following specific steps:
step 2.1, an optimization objective function distributed by the virtual machine with priority on energy efficiency is constructed
Assuming that the number of servers in a cloud computing center is m, a key decision variable e is defined for each server jjIf server j provides a running environment for the virtual machine, then ejIs set to 1, otherwise ejIs set to 0. Based on an extended packing algorithm as a theoretical basis, loading virtual machines into a server with power consumption as a characteristic parameter, arranging all the virtual machines on the server, and ensuring the number N of the servers to be minimum, thereby establishing an objective function:
2.2, converting the virtual machine allocation problem into an optimization problem of multiple constraint conditions;
let the upper power limit of each server of the cloud computing center be the same, denoted as Pj,max{ j ═ 1,2, … m }; total number of requested virtual machines is n, virtual machine VMiIs characterized by a lifetime tiAnd maximum power consumption pi(ii) a Defining a two-dimensional variable xijIf the virtual machine VMiWorking on server j, then xijIs set to 1, otherwise xijIs set to 0. Pj,maxRepresents the upper power limit of server j; pj,currentRepresents the current power of server j; pj,idleRepresenting power consumption, P, when the server is idlej,currentAnd Pj,idleThe relationship between:
the optimization process suffers from the following linear constraints: a server can only run a certain number of virtual machines at most; one virtual machine can only run on one server; and only when the remaining resources are enough, the server can run more virtual machines, and the equations (1) to (3) given in the step two are used as the constraint conditions of the objective function in the scheme.
Per virtual machine VMiThe method has the advantages that proper servers are selected according to the requirements, computing resources are matched according to the requirements, and cloud resource allocation with priority on energy efficiency is achieved. The problem of reducing the number of running servers and improving the energy efficiency is converted into an optimization problem of an expression (4) with expressions (1), (2) and (3) as constraint conditions, namely:
and 2.3, solving the optimization problem, and realizing reasonable distribution of the virtual machines according to the solution result.
The problem P belongs to a multi-constraint condition expansion boxing problem, and the multi-constraint condition expansion boxing problem can be solved by using a space limited online algorithm. In the embodiment, the problem of expanding boxing is solved by adopting a space-limited online algorithm, the first-time adaptation strategy and the optimal adaptation strategy are comprehensively considered by the algorithm, the calculation complexity is reasonable, and the performance is good. The flow chart of the space-limited online algorithm for solving the extended container is shown in FIG. 2:
through the calculation process of fig. 2, a virtual machine placement matrix X is solved:
if xijIf 1, the corresponding virtual machine VM is setiIs arranged on the server.
And step three, according to the task completion condition after the virtual machines are distributed, the invention provides a dynamic virtual machine migration method based on integer linear programming, the virtual machines running on the servers gradually leave the system along with the completion of the tasks, at the moment, real-time information of the resource utilization rate of the servers is collected, the running virtual machines are dynamically integrated into the servers with the least number, and then the idle servers are placed into a low-power-consumption mode.
Step 3.1, deducing energy consumption which can be saved by migrating the virtual machine;
if a group of m' servers in a non-idle state exists, the operating power of the servers is less than the upper power limit, and the virtual machines running on the servers may be migrated. Clearly, the problem of reducing m' is the NP-hard problem. Establishing a savable energy consumption function f through virtual machine migration as follows:
in the above formula, m 'represents the number of servers in a non-idle state, and m' is less than m; pi,idleIs the power consumption of the server in an idle state; y isiIs an identification of whether the server is idle or not; p'kPower consumption required to migrate virtual machine k; three-dimensional variable zijkRepresents the migration of virtual machine k from server i to server j; q. q.siRepresenting the number of virtual machines running on server i;
step 3.2, acquiring a target function of the virtual machine integrated migration;
by solving the maximum value of the energy consumption function f which can be saved, the objective function of the virtual machine integration migration can be obtained, and the calculation formula is as follows:
step 3.3, obtaining the constrained conditions in the virtual machine migration process;
the virtual machine migration process is subject to condition constraints, and the constraint conditions to be met comprise that the virtual machine is migrated from one server i to another different server j (i ≠ j); the virtual machine cannot be migrated multiple times; the power consumed by the migration of the virtual machine cannot be too large; and the number of the servers which work and run after the virtual machine migration is completed is more than or equal to the lower limit of the number of the running servers. The concrete expression formula of the constraint condition is as follows:
wherein: z is a radical ofijkRepresents the migration of virtual machine k from server i to server j; z is a radical ofjlk'Represents the migration of virtual machine k' from server j to server l; m' represents the number of servers in a non-idle state; p is a radical ofkPower consumption required to migrate virtual machine k; pj,maxIs the upper power limit for server j; pj,currentIs the current power of server j; y isiIndicating whether the server is idle or not, if the server is in an idle state due to virtual machine migration, yiPut 1, otherwise yiSetting 0;
representing a lower limit on the number of servers in operation.
Step 3.4, converting the migration adjustment virtual machine problem into an optimization problem with multiple constraint conditions; namely, the optimization problem of the formula (7) with the constraint condition of the formula (8). Can be expressed as:
and 3.5, solving the problem optimization problem by utilizing an Integer Linear Programming (ILP) algorithm, migrating the virtual machines in the running state to the minimum number of servers according to the solving result, and then putting the idle servers into a low power consumption mode.
The problem P can be solved by using an Interactive Linear and General optimization solver LINGO (LINGO) from LINDO systems Inc., or by using a stealth enumeration method.
Solving a problem P by using a LINGO optimization solver: inputting an objective function starting from MAX in a first row of an editing window; secondly, inputting 'ST' in a second row, wherein the constraint condition is represented; thirdly, constraint conditions are input from the third row; and finally, ending the input of the mathematical model by using END. Note that the variable declaration is made after the END, the int variable represents that the variable is an 0/1 integer variable, the gin variable represents that the variable is an integer variable, and a feasible solution vector can be obtained after operation
Solving feasible solutions in problem P using stealth enumeration
The flow of the algorithm is shown in fig. 3.
Obtaining feasible solution vectors
Then, if z isijkWhen 1, then VMkMigrating from server i to server j; otherwise, the VM is maintainedkWorking on the origin server.
The dynamic virtual machine migration method based on integer linear programming can minimize the number of working hosts, maximize the number of idle hosts, realize optimal migration and effectively reduce energy consumption of a cloud computing center.
The invention also provides a system for realizing the method, which comprises a virtual machine allocation management module, an energy consumption evaluation module and an energy perception scheduler, wherein the virtual machine allocation management module, the energy consumption evaluation module and the energy perception scheduler are sequentially connected, and the method comprises the following steps:
the virtual machine allocation management module is used for processing the request of the client and scheduling the virtual machine;
the energy consumption evaluation module is positioned between the virtual machine distribution management module and the energy perception scheduler and used for calculating the power consumption of the virtual machine and the server;
the energy perception scheduler is used for placing a virtual machine in the cloud computing center; specifically, the energy perception scheduler is composed of two functional modules, namely an allocation module and a migration module, wherein the allocation module is used for placing the virtual machine at an initial position through the virtual machine allocation method provided by the invention; the migration module minimizes the number of servers in use or active state by dynamically integrating virtual machines.
In the model, an application program of a cloud user is equivalent to a virtual machine, and material resources provided by a device operator are equivalent to a server, and the specific implementation process is as follows: the client packs the application program into a virtual machine, generates a virtual machine request, and sends the virtual machine request to a virtual machine allocation management module, and the virtual machine allocation management module processes the client request; and meanwhile, the energy consumption evaluation module calculates the power consumption of the virtual machines and the servers by using an energy evaluation tool, and prepares for reasonably distributing the virtual machines. According to the result of the energy consumption evaluation module, an allocation module in the energy-aware scheduler places the virtual request of the cloud user at the optimal server position. When a sufficient number of services leave, a migration module in the energy-aware scheduler dynamically adjusts the integrated virtual machine, releases the server as much as possible, and turns it into a sleep mode or shuts it down.
The effect of the invention can be further illustrated by the following simulation experiment:
1. simulation conditions
Carrying out simulation experiments according to a system model, setting that the power requirements of the virtual machines are randomly and uniformly distributed in [0,1] in the simulation process, the arrival time of the virtual machines accords with the poisson process, the processing time of the virtual machines follows exponential distribution, and the specific parameter settings are shown in the following table:
| parameter(s) | Numerical value |
| Number m of |
10 |
| Number of virtual machines n | {100,200,300} |
| Maximum power P of each |
20 |
| Average load virtual machine number of each cloud server | 40 |
2. Emulated content
Simulation 1: the cloud resource allocation and scheduling simulation is performed in the simulated system by respectively adopting the method of the present invention, the random allocation method and the energy efficiency allocation method only, and fig. 4 is obtained. Fig. 4 shows that, compared with the random allocation method and the energy-efficient allocation method only, the method of the present invention with normalized results of the random allocation method can save energy consumption required for resource allocation and scheduling in a cloud computing environment.
Energy consumption cost reduction in systems with the number of virtual machines of 100, 200 and 300 is respectively 17.9%, 15.9% and 12% by only adopting an energy efficiency priority allocation method; when the energy efficiency priority cloud resource allocation and scheduling method is adopted, the energy consumption cost is respectively reduced by 24.8%, 21.6% and 15.3%.
Simulation 2: cloud resource allocation and scheduling simulation are performed in the simulated system by respectively adopting the method of the invention and only adopting an energy efficiency allocation method, and a graph 5 is obtained. Fig. 5 shows that under different load conditions of the server, the method of the present invention saves more energy than the method using only the energy efficiency distribution method. However, as the load of the server increases, the energy consumption saved by the two methods gradually decreases, and because the configuration optimization possibility of the virtual machine is reduced under high load, the benefit brought by the algorithm is reduced.
Claims (4)
1. A method for energy-efficiency-first cloud resource allocation and scheduling is characterized by comprising the following steps:
the method comprises the steps that firstly, a client packs an application program into a virtual machine, generates a virtual machine request and sends the virtual machine request to a cloud computing center; the cloud computing center establishes a linear constraint condition of the allocation process according to the virtual machine request;
the linear constraint conditions are as follows:
in the above formula, the two-dimensional variable xijRepresenting virtual machines VMiWhether to work on server j, if work on server xijIs set to 1, otherwise xijSetting to 0; n is the total number of the requested virtual machines; m is the total number of the cloud computing center servers; p is a radical ofiIs a virtual machine VMiMaximum power consumption of (1); x is the number ofijRepresenting virtual machines VMiWhether to work on server j; pj,maxIs the upper power limit for server j; e.g. of the typejIs a decision variable; pj,currentIs the current power of server j; e.g. of the typejFor decision variables, if server j provides a running environment for the virtual machine, then ejIs set to 1, otherwise ejSetting to 0;
secondly, the cloud computing center allocates the virtual machines to different servers according to the linear constraint conditions;
step three, dynamically integrating the running virtual machines into the minimum number of servers according to the task completion condition after the virtual machines are distributed, and then putting the idle servers into a low power consumption mode or closing the idle servers;
the specific process of the second step comprises the following steps:
step 2.1, constructing an optimization objective function distributed by the virtual machine with priority on energy efficiency:
in the above formula, N represents the number of servers;
2.2, converting the virtual machine allocation problem into an optimization problem of multiple constraint conditions;
step 2.3, solving the optimization problem, and realizing reasonable distribution of the virtual machines according to the solution result;
the multi-constraint optimization problem P described in step 2.2 is represented as:
in the above formula, piRepresenting virtual machines VMiMaximum power consumption of, xijRepresenting two-dimensional variables if virtual machine VMiWorking on server j, then xijIs set to 1, otherwise xijSetting to 0; n represents the total number of requesting virtual machines;
the specific process of the third step comprises the following steps:
step 3.1, deducing energy consumption which can be saved by migrating the virtual machine;
step 3.2, acquiring a target function of the virtual machine integrated migration;
step 3.3, obtaining the constrained conditions in the virtual machine migration process;
step 3.4, converting the problem of dynamically migrating and adjusting the virtual machine into an optimization problem with multiple constraint conditions;
and 3.5, solving the problem optimization problem by using an integer linear programming ILP algorithm.
2. The energy efficiency priority cloud resource allocation and scheduling method according to claim 1, wherein the energy consumption f that can be saved in step 3.1 is represented as:
in the above formula, m ' represents the number of servers in a non-idle state, and m ' is satisfied '<m is the total number of the cloud computing center servers; pi,idleIs the power consumption of the server in an idle state; y isiIs an identification of whether the server is idle or not; p'kPower consumption required to migrate virtual machine k; three-dimensional variable zijkRepresents the migration of virtual machine k from server i to server j; q. q.siRepresenting the number of virtual machines running on server i.
3. The energy efficiency priority cloud resource allocation and scheduling method according to claim 1, wherein the objective function in step 3.2 is to solve the maximum value of the energy consumption f in step 3.1.
4. The energy-efficiency-first cloud resource allocation and scheduling method according to claim 1, wherein the constraint condition in step 3.3 is expressed as follows:
in the above formula, zijkRepresents the migration of virtual machine k from server i to server j; z is a radical ofjlk'Representing migration of virtual machine k' from server j to server jA server l; m' represents the number of servers in a non-idle state; p is a radical ofkPower consumption required to migrate virtual machine k; pj,maxIs the upper power limit for server j; pj,currentIs the current power of server j; y isiIndicating whether the server is idle or not, if the server is in an idle state due to virtual machine migration, yiPut 1, otherwise yiSetting 0;
representing a lower limit on the number of servers in operation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810041238.7A CN108429784B (en) | 2018-01-16 | 2018-01-16 | Energy efficiency priority cloud resource allocation and scheduling method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810041238.7A CN108429784B (en) | 2018-01-16 | 2018-01-16 | Energy efficiency priority cloud resource allocation and scheduling method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108429784A CN108429784A (en) | 2018-08-21 |
| CN108429784B true CN108429784B (en) | 2020-09-29 |
Family
ID=63155926
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810041238.7A Expired - Fee Related CN108429784B (en) | 2018-01-16 | 2018-01-16 | Energy efficiency priority cloud resource allocation and scheduling method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108429784B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111401616B (en) * | 2020-03-09 | 2022-03-08 | 西安建筑科技大学 | Double-layer scheduling method for precast concrete component in supply chain environment |
| CN111481921B (en) * | 2020-04-10 | 2023-06-13 | 网易（杭州）网络有限公司 | Scheduling method, device, equipment and storage medium of cloud game instance |
| CN117763570B (en) * | 2024-02-22 | 2024-05-07 | 福建银数信息技术有限公司 | Cloud computing-based secure resource pool management method and system |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102404412A (en) * | 2011-12-28 | 2012-04-04 | 北京邮电大学 | Energy saving method and system for cloud compute data center |
| CN102609295A (en) * | 2011-10-18 | 2012-07-25 | 华中科技大学 | Dynamic operation scheduling system of virtual machine |
| US8490091B2 (en) * | 2011-03-28 | 2013-07-16 | International Business Machines Corporation | Virtual machine placement to improve memory utilization |
| CN103488539A (en) * | 2013-09-23 | 2014-01-01 | 北京交通大学 | Data center energy saving method based on central processing unit (CPU) dynamic frequency modulation technology |
| CN103576829A (en) * | 2012-08-01 | 2014-02-12 | 复旦大学 | Hybrid genetic algorithm based dynamic cloud-computing virtual machine scheduling method |
| CN104298339A (en) * | 2014-10-11 | 2015-01-21 | 东北大学 | Server integration method oriented to minimum energy consumption |
| CN105786619A (en) * | 2016-02-24 | 2016-07-20 | 中国联合网络通信集团有限公司 | Virtual machine distribution method and device |
| CN106598733A (en) * | 2016-12-08 | 2017-04-26 | 南京航空航天大学 | Three-dimensional virtual resource scheduling method of cloud computing energy consumption key |
| CN106775987A (en) * | 2016-12-30 | 2017-05-31 | 南京理工大学 | A kind of dispatching method of virtual machine for improving resource efficiency safely in IaaS cloud |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8468230B2 (en) * | 2007-10-18 | 2013-06-18 | Fujitsu Limited | Method, apparatus and recording medium for migrating a virtual machine |
| US9594582B2 (en) * | 2009-01-05 | 2017-03-14 | International Business Machines Corporation | Detection and management of dynamic migration of virtual environments |
| US8214829B2 (en) * | 2009-01-15 | 2012-07-03 | International Business Machines Corporation | Techniques for placing applications in heterogeneous virtualized systems while minimizing power and migration cost |
| US8825863B2 (en) * | 2011-09-20 | 2014-09-02 | International Business Machines Corporation | Virtual machine placement within a server farm |
-
2018
- 2018-01-16 CN CN201810041238.7A patent/CN108429784B/en not_active Expired - Fee Related
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8490091B2 (en) * | 2011-03-28 | 2013-07-16 | International Business Machines Corporation | Virtual machine placement to improve memory utilization |
| CN102609295A (en) * | 2011-10-18 | 2012-07-25 | 华中科技大学 | Dynamic operation scheduling system of virtual machine |
| CN102404412A (en) * | 2011-12-28 | 2012-04-04 | 北京邮电大学 | Energy saving method and system for cloud compute data center |
| CN103576829A (en) * | 2012-08-01 | 2014-02-12 | 复旦大学 | Hybrid genetic algorithm based dynamic cloud-computing virtual machine scheduling method |
| CN103488539A (en) * | 2013-09-23 | 2014-01-01 | 北京交通大学 | Data center energy saving method based on central processing unit (CPU) dynamic frequency modulation technology |
| CN104298339A (en) * | 2014-10-11 | 2015-01-21 | 东北大学 | Server integration method oriented to minimum energy consumption |
| CN105786619A (en) * | 2016-02-24 | 2016-07-20 | 中国联合网络通信集团有限公司 | Virtual machine distribution method and device |
| CN106598733A (en) * | 2016-12-08 | 2017-04-26 | 南京航空航天大学 | Three-dimensional virtual resource scheduling method of cloud computing energy consumption key |
| CN106775987A (en) * | 2016-12-30 | 2017-05-31 | 南京理工大学 | A kind of dispatching method of virtual machine for improving resource efficiency safely in IaaS cloud |
Non-Patent Citations (3)
| Title |
|---|
| 云数据中心中虚拟机放置和实时迁移研究;马飞;《中国博士论文全文数据库 信息科技辑》;20131015;第2.4节 * |
| 基于数据中心的两阶段虚拟机能效优化部署算法;张小庆;《计算机应用》;20141110;第1-2节 * |
| 张小庆.基于数据中心的两阶段虚拟机能效优化部署算法.《计算机应用》.2014, * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108429784A (en) | 2018-08-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105893148B (en) | A kind of accidental task low energy consumption dispatching method based on RM strategy | |
| Zhou et al. | A novel virtual machine deployment algorithm with energy efficiency in cloud computing | |
| CN108429784B (en) | Energy efficiency priority cloud resource allocation and scheduling method | |
| CN108182115A (en) | A kind of virtual machine load-balancing method under cloud environment | |
| CN105446816B (en) | A kind of energy optimization dispatching method towards heterogeneous platform | |
| CN102508714A (en) | Green-computer-based virtual machine scheduling method for cloud computing | |
| CN104331321A (en) | Cloud computing task scheduling method based on tabu search and load balancing | |
| CN101938416A (en) | Cloud computing resource scheduling method based on dynamic reconfiguration virtual resources | |
| Mao et al. | A multi-resource task scheduling algorithm for energy-performance trade-offs in green clouds | |
| CN104601664A (en) | Cloud computing platform resource management and virtual machine dispatching control system | |
| CN108572873A (en) | A kind of load-balancing method and device solving the problems, such as Spark data skews | |
| CN105426241A (en) | Cloud computing data center based unified resource scheduling energy-saving method | |
| CN102981890A (en) | Computing task and virtual machine deploying method within a virtual data center | |
| CN107861796A (en) | A kind of dispatching method of virtual machine for supporting cloud data center energy optimization | |
| CN104023042B (en) | Cloud platform resource scheduling method | |
| CN107357641A (en) | Method for scheduling task in a kind of cloud computing | |
| CN107168770A (en) | A kind of cloud data center workflow schedule of low energy consumption and resource provision method | |
| CN107395731A (en) | A kind of adjusting method and device of the container cluster based on service orchestration | |
| CN103257896B (en) | A kind of Max-D job scheduling method under cloud environment | |
| CN105260230A (en) | Resource scheduling method for data center virtual machine based on segmented service level agreement | |
| CN112416516A (en) | Cloud data center resource scheduling method for resource utility improvement | |
| CN109976911A (en) | A kind of adaptive resource dispatching method | |
| CN105847385B (en) | A kind of cloud computing platform dispatching method of virtual machine based on operation duration | |
| He et al. | Energy-efficient framework for virtual machine consolidation in cloud data centers | |
| CN111242801A (en) | Power system regulation and control cloud power grid operation analysis platform |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200929 |