CN107423109B - Anonymous random variable-based virtual machine energy-saving scheduling method - Google Patents
Anonymous random variable-based virtual machine energy-saving scheduling method Download PDFInfo
- Publication number
- CN107423109B CN107423109B CN201710371693.9A CN201710371693A CN107423109B CN 107423109 B CN107423109 B CN 107423109B CN 201710371693 A CN201710371693 A CN 201710371693A CN 107423109 B CN107423109 B CN 107423109B
- Authority
- CN
- China
- Prior art keywords
- host
- load
- hosts
- virtual machine
- low
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
- G06F2009/4557—Distribution of virtual machine instances; Migration and load balancing
-
- 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
Abstract
The invention relates to an anonymous random variable-based virtual machine energy-saving scheduling method, which comprises the following steps: s1: reading physical host data in the cluster, and dividing the physical host data into an overload host, a normal host, a low-load host and a dormant host; s2: traversing all the overloaded hosts, selecting a virtual machine from each overloaded host, and trying to migrate the virtual machine to other hosts in the cluster; s3: traversing all normal hosts, predicting the overload probability of the next sampling period, and if the overload probability is greater than or equal to the relaxation parameter, trying to migrate a virtual machine from the normal host; s4: selecting a low-load host, and trying to migrate all the virtual machines running on the low-load host to other hosts in sequence; s5: steps S1-S4 are repeated. The cloud data center energy-saving management method can effectively reduce the energy consumption of the cloud data center while ensuring the service quality of the cloud data center, is energy-saving and environment-friendly, and is beneficial to the construction of green and energy-saving data centers.
Description
Technical Field
The invention relates to the technical field of information processing, in particular to a virtual machine energy-saving scheduling method based on anonymous random variables.
Background
The traditional virtual machine scheduling method usually focuses on constructing a cluster with high service quality, only focuses on an overloaded host, only can schedule the overloaded host after sampling data of a physical host of the cluster every time, but does not pay enough attention to energy conservation, and for a migration target host, a very heuristic simple energy-saving strategy (such as selecting a physical host with the lowest running load) is usually used, so that a large number of unnecessary physical hosts in the cluster cannot be effectively closed, and huge electric energy waste is caused.
Moreover, when the traditional scheduling method faces two scheduling targets of energy saving and service quality, the two scheduling targets are often considered, a quantitative method is not used, the description of the scheduling targets is very fuzzy, and the balance between skills and service quality cannot be effectively carried out, so that the traditional method is difficult to consider the service capacity and energy consumption of the cloud computing data center at the same time.
Disclosure of Invention
In order to solve the defects of the prior art, the invention provides an anonymous random variable-based virtual machine energy-saving scheduling method, which comprises the following steps:
s1: reading physical host data in the cluster, and dividing the physical host data into an overload host, a normal host, a low-load host and a dormant host according to the load value of each physical host;
s2: traversing all the overloaded hosts, selecting a virtual machine from each overloaded host, and trying to migrate the virtual machine to other hosts in the cluster;
s3: traversing all normal hosts, predicting the overload probability of the next sampling period, and if the overload probability is greater than or equal to the relaxation parameter, trying to migrate a virtual machine from the normal host;
s4: randomly selecting a low-load host, trying to sequentially migrate all the virtual machines running on the low-load host to other hosts in the cluster, and if all the virtual machines can be migrated, shutting down the low-load host to change the low-load host into a dormant host;
s5: after waiting for a predetermined time, steps S1-S4 are repeated.
In step S1, if the physical hosts of the cluster are all overloaded hosts, adding a sleeping host to the cluster;
in step S2-step S4, before the virtual machine is migrated, the overload probability of the target host after migration is determined, and if the overload probability after migration is greater than or equal to the relaxation parameter, the target host is searched again; if the overload probability of all the target hosts after the migration is greater than or equal to the relaxation parameter, stopping the migration or adding a dormant host to the cluster;
in the steps S2-S3, when migrating the virtual machine, the virtual machine migrates to the normal host first, if no normal host or the overload probabilities after all the normal hosts migrate are greater than or equal to the relaxation parameter, the virtual machine migrates to the low-load host, and if no low-load host or the overload probabilities after all the low-load hosts migrate are greater than or equal to the relaxation parameter, the virtual machine migrates to the dormant host; if no dormant host exists or the overload probability of the next sampling period of all the dormant hosts is greater than or equal to the relaxation parameter, adding the dormant host to the cluster;
in step S4, after selecting a low-load host, it is first determined whether the low-load host is in a low-load state after steps S2-S3, if yes, performing a migration operation of the virtual machine, and if not, searching for a next low-load host; if all the low-load hosts become the normal load state, selecting one low-load host which is in the normal load state to migrate.
In step S4, if the low-load hosts are still in the low-load state after steps S2-S3, one of the virtual machines is selected and sequentially migrated to the normal host first, if there is no normal host or the overload probabilities after all the normal hosts are migrated are greater than or equal to the relaxation parameter, the virtual machine is migrated to the other low-load hosts, and if there is no other low-load host or the overload probabilities after all the other low-load hosts are migrated are greater than or equal to the relaxation parameter, the virtual machine stops migrating;
if all the low-load hosts are in the normal load state through the steps S2-S3, one host in the normal load state is selected, the virtual machine of the host is sequentially migrated to other hosts in the normal load state, and if the overload probability after the migration of no other hosts in the normal load state or all the other hosts in the normal load state is greater than or equal to the relaxation parameter, the migration is stopped.
In step S1, the physical hosts are classified according to the load value of each physical host and the CPU overload threshold and the CPU underload threshold of the physical host, wherein,
the load value calculation method of the physical host comprises the following steps:
if the load value of the physical host is less than or equal to the underload threshold value, the physical host is a low-load host;
if the load value of the physical host is between the underload threshold and the overload threshold, the physical host is a normal host;
if the physical host is in the shutdown state, the physical host is a dormant host.
In step S3, the overload probability of the normal host in the next sampling period is calculated according to the historical load and the maximum entropy principle.
In the steps S2 to S4, the overload probability after the migration of the target host is calculated according to the historical load and the maximum entropy principle.
Wherein the relaxation parameters are introduced according to the following calculation method:
wherein the content of the first and second substances,
represents the set of overloaded hosts at the current duty cycle, the tth sampling point, T represents the total number of samplings, and H represents the set of all physical hosts in the cluster.
The cloud data center energy-saving management method has the advantages that the service quality of the cloud data center can be guaranteed, the energy consumption of the cloud data center can be effectively reduced, the energy is saved, the environment is protected, and the construction of the green energy-saving data center is facilitated.
Drawings
FIG. 1: the method of the invention is compared with the prior method for the overload times;
FIG. 2: the method of the present invention is energy consumption compared to prior art methods.
Detailed Description
In order to further understand the technical scheme and the beneficial effects of the present invention, the following detailed description of the technical scheme and the beneficial effects thereof is provided with the accompanying drawings.
In order to overcome the defects of the traditional scheduling method, the invention constructs an energy-saving scheduling model based on anonymous random variables, and introduces a relaxation factor, an overload threshold and an underload threshold so as to fully balance the cluster service capacity and the energy consumption of the data center in the energy-saving scheduling process.
The core concept of the invention is that under the condition of ensuring that each running physical host is in normal load, the number of the dormant hosts is increased as much as possible, thereby reducing the energy consumption of the data center to the maximum extent under the condition of ensuring the cluster service capability.
In order to realize the concept, the technical scheme of the invention mainly comprises two parts.
In one aspect, the invention constructs a load prediction model based on anonymous random variables. The model firstly assumes that the load of the virtual machine is an anonymous random variable subject to some unknown distribution in a period of time, and the random variable of the load of the physical host can be obtained by summing the loads of all the virtual machines on the same physical host. Therefore, the historical load of the virtual machine can be used as a random sample, and the probability that the load of the physical host is greater than the overload threshold value is calculated and output according to the historical load and the maximum entropy principle.
On the other hand, on the basis of the load prediction model, a novel energy-saving scheduling method is designed: firstly, reading physical host data of a data center, and classifying hosts according to load; then, attempting to migrate the virtual machine from the overloaded host; next, predicting the overload probability of the normal host, and carrying out virtual machine migration on the host with high overload probability; finally, an attempt is made to shut down the low-load host. In the process, the migration decision making needs to refer to the output result of the load prediction model, and the overload probability of the migration target machine is ensured to be less than the preset relaxation factor, namely the SLA quality parameter psla。
Firstly, construction of load prediction model
The load prediction model is constructed based on the following assumptions: at sampling time [ t-n +1, t +1]Virtual machine load
Is a random variable that determines the distribution,
the load of the current virtual machine in MIPS is shown, the subscript i, j indicates the jth virtual machine running on the ith physical host, and the superscript (t) indicates the current sampling time. Further assume that the virtual machine loads are independent of each other, when the physical host loads
By
The sum of the individual random variables gives:
wherein the content of the first and second substances,
representing the number of virtual machines running on the ith physical host at sample time t.
Without loss of generality, the SLA maintenance capacity of a physical host in a cloud computing data center is evaluated by the overload time of the physical host. The following describes the determination method of the overload and underload states of the physical host.
Let hostiThe observed value of the CPU load at time t is
When the following inequality is satisfied,
in the CPU overload state:
wherein, usetopIs the CPU overload threshold of the physical host, which corresponds to the virtual machine service capability specified in the SLA. Similarly, an underloaded state is assumed when the host load satisfies the following inequality:
usagebotbeing the physical host's underrun threshold, when the physical host is underrun, we should try to shut it down to save power, or create or introduce a new virtual machine on top.
In the present invention, the load value X represents a random variable and represents an abstract concept of the load value of the physical host, the load observed value use represents a specific load value X of the physical host at a certain time, and the load observed values use of the physical host at a plurality of different times form a normal distribution of the load value X.
Introduction of relaxation parameters
The introduction of the relaxation parameter c is based on the principle of balancing two scheduling objectives of service quality and energy consumption saving, and represents an upper limit of the probability of the occurrence of the overloaded host in the cluster, wherein the probability of the occurrence of the overloaded host at a certain time in the cluster is calculated by the following formula:
wherein the content of the first and second substances,
represents the set of overloaded hosts at the T-th sampling point, T represents the total number of samplings, and H represents the set of all physical hosts in the cluster.
The relaxation parameter C is a value introduced for human, and the significance is to control the value of Psla corresponding to each sampling point t in the cluster to be smaller than the relaxation parameter C as much as possible.
When the relaxation parameter C takes a smaller value, the algorithm preferentially ensures the SLA maintenance capability; and when the relaxation parameter C takes a larger value, the algorithm gives priority to the energy-saving effect.
Setting of energy-saving target
In order to realize that the value of Psla is smaller than the relaxation parameter C at each sampling point t, in the virtual machine migration process, it is required to ensure that the probability that each physical host is a load host at the next sampling point t +1 is also smaller than the relaxation parameter, that is, the energy-saving scheduling objective of the present invention can be described as the following constrained optimization problem:
wherein the content of the first and second substances,
is the set of sleeping hosts in the cluster.
The constraint target can be realized by combining technologies such as overload scheduling, load prediction (based on the load model), underload emptying and the like during each time of scheduling and sampling, and the specific realization method is as follows:
s1: reading physical host data in the cluster, and dividing the physical host data into an overload host, a normal host, a low-load host and a dormant host according to the load value of each physical host, wherein the specific judgment method is detailed in the construction part of the load model; and if the physical hosts of the cluster are all overloaded hosts, adding dormant hosts into the cluster.
S2: all overloaded hosts are traversed, a virtual machine is selected from each overloaded host, and the virtual machine is attempted to be migrated to other hosts in the cluster.
S3: and traversing all normal hosts, predicting the overload probability of the next sampling period of the normal hosts according to the historical load and the maximum entropy principle, and if the overload probability is greater than or equal to the relaxation parameter, attempting to migrate a virtual machine from the normal host.
S4: randomly selecting a low-load host, trying to sequentially migrate all the virtual machines running on the low-load host to other hosts in the cluster, and if all the virtual machines can be migrated, shutting down the low-load host to change the low-load host into a dormant host;
s5: after waiting for a predetermined time, steps S1-S4 are repeated.
In the present invention, in steps S2-S4, before migrating out of the virtual machine, the overload probability (calculated by the historical load and the maximum entropy principle) of the migrated target host is determined, and if the overload probability is greater than or equal to the relaxation parameter, the target host is searched again; in step S2-step S3, if the overload probabilities after all the target hosts are migrated are greater than or equal to the relaxation parameter, adding the sleeping hosts to the cluster; in step S4, if the overload probabilities after all the target hosts are migrated are greater than or equal to the relaxation parameter, the migration is stopped;
in the steps S2-S3, when migrating the virtual machine, the virtual machine migrates to the normal host first, and if there is no normal host or the overload probabilities after all the normal hosts migrate are greater than or equal to the relaxation parameters, the virtual machine migrates to the low-load host; if no low-load host computer exists or the overload probability after all the low-load host computers are migrated is larger than or equal to the relaxation parameter, migrating to the dormant host computer; and if no dormant host exists or the overload probability of the next sampling period of all the dormant hosts is greater than or equal to the relaxation parameter, adding the dormant hosts to the cluster.
After the steps S2-S3, some load hosts or virtual machines of normal hosts may migrate to the low-load host, so that the low-load host is in a normal state (since the overload probability after the migration of the target host is determined in advance before the migration of the above steps S2-S3, the low-load host will not be in an overload state after the migration of the target host through the steps S2-S3), therefore, in the step S4, after one low-load host is selected, it is required to determine whether the low-load host is in a low-load state, if so, the migration operation of the virtual machine is performed, and if not, the next low-load host is searched; if all the low-load hosts become the normal load state, selecting one low-load host which is in the normal load state for migration; in the migration process, some low-load hosts face the situation that the migration needs to be stopped after migrating a part of virtual machines, some low-load hosts can completely migrate the virtual machines, and if the low-load hosts can completely migrate the virtual machines, one physical host can be released, so that the energy consumption is saved.
Specifically, in the step S4, if the low-load hosts are still in the low-load state after the steps S2 to S3, one of the virtual machines is selected to be migrated to the normal host in sequence (the normal host includes the normal host determined in the step S1, and also includes the low-load host that is determined in the step S1 and is in the normal state after the steps S2 to S3), and if there is no normal host or the overload probabilities after all the normal hosts are migrated are greater than or equal to the relaxation parameter, the virtual machine is migrated to the other low-load hosts; if no other low-load host or the overload probability of all other low-load hosts after the migration is greater than or equal to the relaxation parameter, stopping the migration;
if all the low-load hosts are in the normal load state after the steps S2-S3, one host in the normal load state is selected, the virtual machine of the host is sequentially migrated to other hosts in the normal load state (including the normal host determined in the step S1, and also including the low-load host in the normal load state determined in the step S1, but after the steps S2-S3, the low-load host in the normal load state is assumed), and if no other host in the normal load state or all the other hosts in the normal load state have the overload probability after the migration greater than or equal to the relaxation parameter, the migration is stopped.
In conclusion, the energy-saving scheduling method for the virtual machine provided by the invention takes the anonymous random variable as a basis, and can fully balance the service quality and save the energy consumption by constructing the load model and combining the introduction of the relaxation parameter and the judgment of the overload probability.
Table 1 is a violation rate and energy consumption table of the scheduling method of the present invention and the existing scheduling method, and fig. 1 and fig. 2 are a comparison graph of overload times and energy consumption of the method of the present invention and the existing method, respectively, specifically, 480 virtual machines are operated on 40 physical hosts for 10 days; wherein, the ARV is the scheduling method of the invention, the NKPSA is the existing energy-saving scheduling method on the cloud management platform of the operating system (virtualized version) of the Kangsu server, and the MAO is the common overload scheduling method without the energy-saving strategy.
Experimental comparisons on specific data sets show that: compared with the existing energy-saving scheduling method on the cloud management platform of the operating system of the winning-bid kylin server, the scheduling method is more superior, and is embodied in that not only can the energy consumption of the cluster be reduced, but also the service quality of the cluster can be improved.
Table 1: violation rate and energy consumption table of scheduling method and existing scheduling method
The invention has the following beneficial effects:
1. a load model is constructed based on anonymous random variables, classification of physical hosts is achieved according to the load model, and a foundation is provided for cluster optimization in the later period.
2. By introducing relaxation parameters a balance is achieved between quality of service and saving energy.
3. By combining the load model with various scheduling strategies, a scheduling method capable of fully balancing the service quality and saving the energy consumption is provided.
Although the present invention has been described with reference to the preferred embodiments, it should be understood that the scope of the present invention is not limited thereto, and those skilled in the art will appreciate that various changes and modifications can be made without departing from the spirit and scope of the present invention.
Claims (6)
1. A virtual machine energy-saving scheduling method based on anonymous random variables is characterized by comprising the following steps:
s1: reading physical host data in the cluster, and dividing the physical host data into an overload host, a normal host, a low-load host and a dormant host according to the load value of each physical host;
s2: traversing all the overloaded hosts, selecting a virtual machine from each overloaded host, and trying to migrate the virtual machine to other hosts in the cluster;
s3: traversing all normal hosts, predicting the overload probability of the next sampling period, and if the overload probability is greater than or equal to the relaxation parameter, trying to migrate a virtual machine from the normal host;
s4: randomly selecting a low-load host, trying to sequentially migrate all the virtual machines running on the low-load host to other hosts in the cluster, and if all the virtual machines can be migrated, shutting down the low-load host to change the low-load host into a dormant host;
s5: after waiting for a predetermined time, repeating steps S1-S4;
in step S1, if the physical hosts of the cluster are all overloaded hosts, adding a sleeping host to the cluster;
in step S2-step S4, before the virtual machine is migrated, the overload probability of the target host after migration is determined, and if the overload probability after migration is greater than or equal to the relaxation parameter, the target host is searched again; if the overload probability of all the target hosts after the migration is greater than or equal to the relaxation parameter, stopping the migration or adding a dormant host to the cluster;
in the steps S2-S3, when migrating the virtual machine, the virtual machine migrates to the normal host first, if no normal host or the overload probabilities after all the normal hosts migrate are greater than or equal to the relaxation parameter, the virtual machine migrates to the low-load host, and if no low-load host or the overload probabilities after all the low-load hosts migrate are greater than or equal to the relaxation parameter, the virtual machine migrates to the dormant host; if no dormant host exists or the overload probability of the next sampling period of all the dormant hosts is greater than or equal to the relaxation parameter, adding the dormant host to the cluster;
in step S4, after selecting a low-load host, it is first determined whether the low-load host is in a low-load state after steps S2-S3, if yes, performing a migration operation of the virtual machine, and if not, searching for a next low-load host; if all the low-load hosts become the normal load state, selecting one low-load host which is in the normal load state to migrate.
2. The anonymous random variable-based virtual machine energy-saving scheduling method of claim 1, wherein:
in step S4, if the low-load hosts are still in the low-load state after steps S2-S3, one of the low-load hosts is selected, the virtual machine is sequentially migrated to the normal host first, if no normal host or the overload probabilities after all the normal hosts are migrated are greater than or equal to the relaxation parameter, the virtual machine is migrated to the other low-load hosts, and if the overload probabilities after no other low-load host or all the other low-load hosts are migrated are greater than or equal to the relaxation parameter, the virtual machine is stopped migrating;
if all the low-load hosts are in the normal load state through the steps S2-S3, one host in the normal load state is selected, the virtual machine of the host is sequentially migrated to other hosts in the normal load state, and if the overload probability after the migration of no other hosts in the normal load state or all the other hosts in the normal load state is greater than or equal to the relaxation parameter, the migration is stopped.
3. The anonymous random variable-based virtual machine energy-saving scheduling method of claim 1, wherein:
in step S1, the physical hosts are classified according to their load values and their CPU overload and underload thresholds, wherein,
the load value calculation method of the physical host comprises the following steps:
if the load value of the physical host is less than or equal to the underload threshold value, the physical host is a low-load host;
if the load value of the physical host is between the underload threshold and the overload threshold, the physical host is a normal host;
if the physical host is in the shutdown state, the physical host is a dormant host.
4. The anonymous random variable-based virtual machine energy-saving scheduling method of claim 1, wherein: in step S3, the overload probability of the normal host in the next sampling period is calculated according to the historical load and the maximum entropy principle.
5. The anonymous random variable-based virtual machine energy-saving scheduling method of claim 1, wherein: in the steps S2-S4, the overload probability after the migration of the target host is calculated according to the historical load and the maximum entropy principle.
6. The anonymous random variable-based virtual machine energy-saving scheduling method of claim 1, wherein: the relaxation parameters are introduced according to the following calculation method:
wherein the content of the first and second substances,
represents the set of overloaded hosts at the current duty cycle, the tth sampling point, T represents the total number of samplings, and H represents the set of all physical hosts in the cluster.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710371693.9A CN107423109B (en) | 2017-05-24 | 2017-05-24 | Anonymous random variable-based virtual machine energy-saving scheduling method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710371693.9A CN107423109B (en) | 2017-05-24 | 2017-05-24 | Anonymous random variable-based virtual machine energy-saving scheduling method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107423109A CN107423109A (en) | 2017-12-01 |
| CN107423109B true CN107423109B (en) | 2020-05-01 |
Family
ID=60428580
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710371693.9A Active CN107423109B (en) | 2017-05-24 | 2017-05-24 | Anonymous random variable-based virtual machine energy-saving scheduling method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107423109B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109669760B (en) * | 2018-11-01 | 2020-03-31 | 江苏南大苏富特科技股份有限公司 | Virtualized dynamic resource management system |
| CN109656678B (en) * | 2018-11-01 | 2020-05-08 | 江苏南大苏富特科技股份有限公司 | Dynamic resource management method based on virtualization |
| CN112148572B (en) * | 2020-08-25 | 2023-09-19 | 杭州叙简科技股份有限公司 | Virtual machine cluster test method, system, electronic device and storage medium |
| CN113342462B (en) * | 2021-06-02 | 2022-03-15 | 燕山大学 | Cloud computing optimization method, system and medium integrating limitation periodic quasi-dormancy |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102457583A (en) * | 2010-10-19 | 2012-05-16 | 中兴通讯股份有限公司 | Realization method of mobility of virtual machine and system thereof |
| CN102696000A (en) * | 2010-01-13 | 2012-09-26 | 美国日本电气实验室公司 | Methods and apparatus for coordinated energy management in virtualized data centers |
| CN105159751A (en) * | 2015-09-17 | 2015-12-16 | 河海大学常州校区 | Energy-efficient virtual machine migration method in cloud data center |
| CN105930202A (en) * | 2016-04-29 | 2016-09-07 | 合肥工业大学 | Migration policy for virtual machine with three thresholds |
| CN106201700A (en) * | 2016-07-19 | 2016-12-07 | 北京工业大学 | The dispatching method that a kind of virtual machine migrates online |
| 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 |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8489744B2 (en) * | 2009-06-29 | 2013-07-16 | Red Hat Israel, Ltd. | Selecting a host from a host cluster for live migration of a virtual machine |
-
2017
- 2017-05-24 CN CN201710371693.9A patent/CN107423109B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102696000A (en) * | 2010-01-13 | 2012-09-26 | 美国日本电气实验室公司 | Methods and apparatus for coordinated energy management in virtualized data centers |
| CN102457583A (en) * | 2010-10-19 | 2012-05-16 | 中兴通讯股份有限公司 | Realization method of mobility of virtual machine and system thereof |
| CN105159751A (en) * | 2015-09-17 | 2015-12-16 | 河海大学常州校区 | Energy-efficient virtual machine migration method in cloud data center |
| CN105930202A (en) * | 2016-04-29 | 2016-09-07 | 合肥工业大学 | Migration policy for virtual machine with three thresholds |
| CN106201700A (en) * | 2016-07-19 | 2016-12-07 | 北京工业大学 | The dispatching method that a kind of virtual machine migrates online |
| 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 |
Non-Patent Citations (1)
| Title |
|---|
| 基于负载特征聚类的节能资源调度算法;夏庆新,兰雨晴等;《北京航空航天大学学报》;20141219;第41卷(第4期);第680-685页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107423109A (en) | 2017-12-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107423109B (en) | Anonymous random variable-based virtual machine energy-saving scheduling method | |
| Askarizade Haghighi et al. | An energy-efficient dynamic resource management approach based on clustering and meta-heuristic algorithms in cloud computing IaaS platforms: Energy efficient dynamic cloud resource management | |
| Zhou et al. | Virtual machine placement algorithm for both energy-awareness and SLA violation reduction in cloud data centers | |
| CN105159751B (en) | The virtual machine migration method of energy efficient in a kind of cloud data center | |
| Xiao et al. | Multi-objective VM consolidation based on thresholds and ant colony system in cloud computing | |
| CN108196935B (en) | Cloud computing-oriented virtual machine energy-saving migration method | |
| CN109901932B (en) | Server integration method based on virtual machine | |
| CN103916438B (en) | Cloud testing environment scheduling method and system based on load forecast | |
| CN102624865B (en) | Cluster load prediction method and distributed cluster management system | |
| Xu et al. | A novel artificial bee colony approach of live virtual machine migration policy using Bayes theorem | |
| Xu et al. | Comparison between dynamic programming and genetic algorithm for hydro unit economic load dispatch | |
| Wang et al. | A task scheduling strategy in edge-cloud collaborative scenario based on deadline | |
| CN106201700A (en) | The dispatching method that a kind of virtual machine migrates online | |
| Dhingra et al. | Green cloud: heuristic based BFO technique to optimize resource allocation | |
| CN106775944A (en) | The method integrated based on cultural multi-ant colony algorithm virtual machine under cloud platform | |
| CN113535409A (en) | Server-free computing resource distribution system oriented to energy consumption optimization | |
| CN107506233A (en) | A kind of schedule virtual resources method, apparatus and server | |
| CN111176784B (en) | Virtual machine integration method based on extreme learning machine and ant colony system | |
| CN109976879B (en) | Cloud computing virtual machine placement method based on resource usage curve complementation | |
| Fu et al. | Network traffic based virtual machine migration in cloud computing environment | |
| CN108388471B (en) | Management method based on double-threshold constraint virtual machine migration | |
| Chen et al. | An energy‐efficient method of resource allocation based on request prediction in multiple cloud data centers | |
| CN105930202A (en) | Migration policy for virtual machine with three thresholds | |
| CN110888713A (en) | Trusted virtual machine migration algorithm for heterogeneous cloud data center | |
| CN110850957A (en) | Scheduling method for reducing system power consumption through dormancy in edge computing scene |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |