CN106775949B - Virtual machine online migration optimization method capable of sensing composite application characteristics and network bandwidth - Google Patents
Virtual machine online migration optimization method capable of sensing composite application characteristics and network bandwidth Download PDFInfo
- Publication number
- CN106775949B CN106775949B CN201611231403.2A CN201611231403A CN106775949B CN 106775949 B CN106775949 B CN 106775949B CN 201611231403 A CN201611231403 A CN 201611231403A CN 106775949 B CN106775949 B CN 106775949B
- Authority
- CN
- China
- Prior art keywords
- virtual machine
- network
- network bandwidth
- application
- 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.)
- 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
-
- 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
- G06F2009/45595—Network integration; Enabling network access in virtual machine instances
Abstract
The invention discloses a virtual machine online migration optimization method for sensing composite application characteristics and network bandwidth, and belongs to the technical field of software. The method comprises the following steps: 1) sensing a virtual machine application characteristic environment and a network bandwidth environment, and collecting the number of internal memory dirty pages; 2) predicting the number of the dirty pages in the memory by using a gray prediction model; 3) calculating the dirty page rate of the iteration cycle of the virtual machine; 4) collecting the use condition of network bandwidth; 5) and judging whether the virtual machine is a network intensive virtual machine or not according to the network bandwidth required by the application in the virtual machine, and then reserving the network bandwidth. When the virtual machine migration of the network intensive application or the memory intensive application is faced, the method can reduce the extra overhead in the migration process, improve the transmission efficiency in the migration process and effectively reduce the migration time.
Description
Technical Field
The invention designs an optimization method for online migration of a virtual machine sensing application characteristics and a network environment, particularly designs a network bandwidth reservation adjustment algorithm sensing application characteristics, and belongs to the technical field of software.
Background
With the wide application of virtualization technology, dynamic management of virtual machines is becoming more and more important, and online migration is an important means for dynamic management of virtual machines. Online migration is a network intensive activity that requires the transfer of several GB or even tens of GB of virtual machine memory state from a source host to a destination host. Besides occupying network resources, the online migration consumes additional physical resources such as memory and CPU.
At present, the traditional online migration methods include pre-copy (pre-copy), post-copy (post-copy) and hybrid copy (hybrid copy) online migration methods. The online migration of pre-copy is the mainstream dynamic migration technology of the virtual machine at present, and the pre-copy is divided into 3 stages: first copy, iterative copy and stop copy. The specific migration process is shown in fig. 1. Firstly, copying all memories of the virtual machine to a target host machine, wherein the operation of the virtual machine is not interrupted in the process, and the stage is called as a 'first copying' stage; and then, entering an 'iterative copy stage', and iteratively copying the memory dirty page (the last modified memory page) generated in the last copy process to the destination host, wherein the process does not interrupt the operation of the virtual machine. After each iteration, whether the condition of entering the stop-and-coy stage (i.e. the stop copy stage) is met currently needs to be judged. If yes, entering a third stage, namely a 'stop copy' stage, and otherwise, continuing to carry out iterative copy. The virtual machines on the source host are suspended in the shutdown copy phase, and then the dirty pages in the rest virtual machines are synchronized to the destination host. Meanwhile, the system information of the virtual machine, including the CPU and the network state, is synchronized to the destination terminal. After the information is transmitted, the virtual machine on the destination host machine restores the system according to the transmitted system information of the virtual machine. Unlike pre-copy online migration, memory synchronization for post-copy online migration is after the virtual machine resumes running on the destination host. Hybridcopy online migration is a special case of post-copy online migration. The method integrates the characteristics of a pre-copy online migration method and a post-copy online migration method, and transfers the most frequently accessed memory page subset to the destination host before the system information of the virtual machine is transferred to the destination host. The remaining infrequently used memory pages are retrieved from the source host as needed for the virtual machine to run.
The mainstream virtualization platforms (KVM, Xen, VMware and the like) support pre-copy algorithm for online migration. The pre-copy online migration method has been widely used and developed in the commercial and academic fields. The method can effectively shorten the migration time and improve the migration performance, but in practical application, the migration performance is not ideal when a pre-copy online migration method is used for migrating the network-intensive virtual machine and the memory-intensive virtual machine under the influence of iterative convergence, different application characteristics of the virtual machine and resource limitation. For example, if a memory-intensive application is running in a virtual machine, the memory of the virtual machine is continuously and rapidly modified, and when the modification speed is greater than the transmission speed of the dirty memory page, the pre-copy iterative copy time is prolonged, the physical resource occupation time is increased, and the shutdown copy stage is difficult to enter, which may seriously affect other services, and even may cause migration failure. The influence of the virtual machine application characteristic environment on the dirty pages in the virtual machine is not considered in the method for predicting the dirty pages in the virtual machine. At present, a pre-copy online migration improvement method reduces data transmission of a dirty memory page by methods of deleting repeated memory page data, compressing the memory page, inhibiting generation of similar memory pages and the like, so as to reduce migration time and downtime. However, the above method often increases a large amount of extra CPU consumption, and does not consider the influence of the network bandwidth environment on the transmission of the dirty pages.
Disclosure of Invention
The invention aims to: the method for optimizing the online migration of the virtual machine by sensing the application characteristics and the network environment overcomes the problems that the influence of the application characteristics on the number of the dirty pages of the memory is not considered in the prior relevant achievement, the combined action of the number of the dirty pages of the memory and the network bandwidth is ignored, the migration performance problems such as long iteration period, increased downtime and the like in the migration scene of the network-intensive or memory-intensive virtual machine are solved, the competition of network resources can be effectively reduced, the network transmission efficiency is improved, the migration time is reduced, and the purpose of providing the migration performance of the virtual machine is achieved.
In the invention, in order to obtain the specific value of the internal memory dirty page, a grey prediction model is adopted to predict the internal memory dirty page. The reason for selecting the method is that the dirty page information in the memory copying process is difficult to acquire, the acquired amount is small, and the gray prediction model can perform better prediction under the condition of less data samples.
As shown in fig. 2, the method for optimizing online migration of a virtual machine in sensing application characteristics and a network environment of the present invention is as follows:
(1) sensing application characteristics and a network environment of the virtual machine, and collecting the number of internal dirty pages; the sensing of the application characteristics of the virtual machine refers to the dynamic acquisition of the use condition of application resources in the virtual machine, including the memory use rate, the CPU use rate and the network bandwidth, and the trend of the change of the application resource use rate can be mastered; the perception of the network environment refers to the dynamic acquisition of the utilization rate of the network bandwidth of the cloud data center; the memory dirty page refers to a memory page modified in the migration process of the virtual machine;
(2) according to the application characteristics and the network environment of the perception virtual machine in the step (1) and the acquired internal dirty page number, using a grey prediction model to predict the internal dirty page number;
(3) calculating the dirty page rate of the iteration cycle of the virtual machine according to the internal memory dirty page number obtained by prediction in the step (2), wherein the dirty page rate is the dirty page number generated in the iteration cycle/the iteration cycle time;
(4) obtaining the use condition of network bandwidth;
(5) and (4) judging whether the virtual machine is a network intensive virtual machine according to the network bandwidth use condition obtained in the step (4), and then reserving the network bandwidth. In the network bandwidth reservation process, bandwidth reservation is carried out according to the network bandwidth and the dirty page ratio required by the application of the virtual machine, so that the sufficient network bandwidth in each iteration period in the migration process is ensured, and network congestion is reduced. The method performs network bandwidth reservation for network-containing intensive application in combination with a network environment, can reduce competition of network bandwidth in a migration process, improves migration efficiency, reduces iteration time, and reduces migration time, thereby achieving the purpose of improving migration performance. Meanwhile, the migration of the non-network intensive virtual machines is preferentially ensured, network resources are reasonably allocated, and the utilization rate of the network is improved.
The process of predicting the number of dirty pages in the next period by using the gray prediction model in the step (2) is as follows:
(21) for the collection of step (1) data, namely X1: migrating the quantity of internal memory dirty pages generated in different iteration cycles; x2: migrating the memory utilization rate of the virtual machine in different iteration periods; x3: migrating the CPU utilization rate of the virtual machine in different iteration cycles; x4: migrating the network bandwidth of the virtual machine in different iteration periods; x5: the time of the last iteration; x6: and migrating the network use conditions of the cloud data center in different iteration periods. These data are converted into a matrix as a gray prediction sequence X(0)And for said sequence X(0)Accumulating to generate AGO sequence with X(1)；
(22) X obtained according to step (21)(1)Solving a sequence generated by the neighbor value, namely a mean sequence;
(23) in the gray prediction model, assume X of step (21)(1)A relationship exists with step (22), thereby establishing a gray differential equation;
(24) establishing a whitening differential equation according to step (22);
(25) according to the equation set in the step (23), establishing an equation set based on GM (1, N), and solving a parameter sequence of the GM (1, N) equation by using a least square method;
(26) substituting the parameters into the whitening differential equation in the step (24), solving to obtain a GM (1, N) discrete response function, and solving to obtain a predicted value sequence through an accumulation reduction formula;
(27) in order to reduce the error of the grey prediction model, the prediction precision is improved by correcting the predicted value through the residual error.
The network bandwidth reservation process in the step (5) is as follows:
(31) collecting network bandwidth information required by the application in the virtual machine, judging whether the virtual machine belongs to a network intensive application virtual machine, and jumping to the step (32) if the migration virtual machine is not the network intensive application virtual machine; if the virtual machine is a network-intensive virtual machine, jumping to the step (34);
(32) calculating the internal memory dirty page rate of the virtual machine according to the step (3), and judging whether the current idle bandwidth meets the transmission bandwidth of the number of the internal memory dirty pages next, namely whether the network bandwidth is greater than the internal memory dirty page rate; if yes, no adjustment is carried out; otherwise, adjusting the network bandwidth, namely jumping to the step (33);
(33) preferentially ensuring non-network intensive application migration and delaying the network intensive virtual machine migration at the same time;
(34) and reserving application bandwidth according to the historical network transmission flow data, and subtracting the application reserved bandwidth from the current idle network bandwidth to obtain the bandwidth allocated for the current virtual machine migration.
Compared with the prior art, the invention has the advantages that:
(1) in the prior art, methods such as repeating memory page data deletion and memory page compression, inhibiting generation of similar memory pages and the like bring larger additional resource overhead such as CPU resource overhead, and meanwhile, the efficiency of the methods also depends on the efficiency of a deletion algorithm, a compression algorithm and a decompression algorithm, and the influence of a network bandwidth environment on transmission of the memory dirty pages is not considered. Compared with the traditional pre-copy strategy, the method can predict the amount of dirty pages in the virtual machine under the environment of sensing different application characteristics, and meanwhile, in order to improve the network utilization rate, the method can perform network bandwidth reservation adjustment under the environment of sensing a network according to different application characteristics.
(2) Compared with the traditional pre-copy online migration method, in the virtual machine migration process containing the network intensive application or the memory intensive application, the network bandwidth reservation adjustment not only considers the network flow in the virtual machine migration process, but also reserves the network bandwidth required by the network intensive application in the virtual machine. The invention can effectively reduce competition of network resources, improve network transmission efficiency and reduce migration time.
Drawings
FIG. 1 is a schematic diagram of a pre-copy online migration process;
FIG. 2 is a process diagram of a virtual machine online migration optimization method for sensing composite application features and network environment according to the present invention;
FIG. 3 shows the GM (1, N) predicting the number of dirty pages in memory;
fig. 4 is a schematic diagram of a network reservation adjustment process.
Detailed Description
The present invention will be described in detail below with reference to specific embodiments and the accompanying drawings.
The virtual machine online migration optimization strategy for sensing the application characteristics and the network environment, which is provided by the invention, collects the application characteristic data of the virtual machine and the network environment characteristic data, and performs network bandwidth reservation adjustment according to the data, so that the purpose of optimizing the migration performance of the virtual machine is achieved, and the virtual machine online migration optimization strategy has an obvious effect particularly under the migration scene of the virtual machine containing network intensive application and memory intensive application.
After collecting samples of certain virtual machine application characteristics and network environment, the samples are input into a prediction model based on the number of dirty pages in GM (1, N) (i.e., gray prediction model).
Selecting a dirty page prediction variable in a virtual machine, the invention considers the following factors capable of reflecting the application characteristic environment of the virtual machine: x1: migrating the quantity of internal memory dirty pages generated in different iteration cycles; x2: migrating the memory utilization rate of the virtual machine in different iteration periods; x3: migrating the CPU utilization rate of the virtual machine in different iteration cycles; x4: migrating the network bandwidth of the virtual machine in different iteration periods; x5: the time of the last iteration; x6: and migrating the network use conditions of the cloud data center in different iteration periods. And (4) totally carrying out S times of iterations in the migration process, wherein the quantity of the dirty pages generated in each time is related to the application characteristics of the previous iteration period and the previous iteration time.
The building process of the prediction model of the internal memory dirty pages based on G (1, N) is shown in FIG. 3, and the specific steps are as follows:
the method comprises the following steps: giving a grey prediction sequence X(0)Let X(1)Is X(0)The AGO sequence of (1):
wherein, the elements of X1 represent the number of internal dirty pages generated in each cycle of S iteration cycles; x2 elements, which respectively represent the memory usage rate of each iteration virtual machine in S iteration cycles; x3 elements, namely respectively representing the CPU utilization rate of each iteration virtual machine in S iteration cycles; the number of the elements is X4, namely the elements respectively represent the network bandwidth of each iteration virtual machine in S iteration cycles; x5 elements, which respectively represent the last iteration cycle; the X6 elements respectively represent the network use condition of the cloud data center of each iteration in S iteration cycles;
step two: solving for
step three: establishing a gray differential equation:
wherein k is 1,2, …, S; a is a coefficient of development; bi(i-2, 3, …,6) is a drive coefficient.
Step four: establishing a whitening differential equation:
wherein k is 1,2, …, S; a is a coefficient of development; bi (i ═ 2,3, …,6) is the drive coefficient.
Step five: establishing a system based on GM (1, N) equation:
make the parameters listed as
The GM (1, N) ash differential equation is then:
wherein k is 1,2, …, S; a is a coefficient of development; bi (i ═ 2,3, …,6) is the drive coefficient.
Solving by least squares
Step six: solving for the GM (1, N) discrete response function:
substituting the solved parameters into a differential equation, the GM (1, N) discrete response function is:
wherein k is 1,2,3, …, 6; e is a mathematical constant. And because of
The cumulative subtraction reduction formula is:
step seven: residual correction
The gray prediction model sometimes predicts that the data with large fluctuation can influence the development rule of the data column, so that model correction needs to be carried out on GM (1, N). Order to
In order to be the original residual sequence,
if k is present0The following conditions are satisfied:
(1)
are the same in sign;
(2) when S-k0When the value is more than or equal to 4, the product can be called
For modellable residual tail segment, note
To pair
Establishing GM (1,1) model, and calculating its parameter list
Calculate out
Analog value of (d):
the modified form of the GM (1, N) discrete response function is:
after the number of internal memory dirty pages is predicted, the internal memory dirty page rate is calculated. According to the internal memory dirty page rate generated by different application characteristics, network bandwidth adjustment is carried out by combining a network environment, and a network bandwidth reservation adjustment algorithm based on application characteristic perception is provided.
The network reservation adjustment algorithm is shown in fig. 4. As can be seen from fig. 4, it is determined whether the virtual machine is a network-intensive application according to the network bandwidth condition required by the virtual machine application, that is, if the average load of the network bandwidth of the virtual machine is greater than the set network bandwidth-intensive threshold, the virtual machine is a network-intensive application virtual machine; and, conversely, is not a network bandwidth intensive application virtual machine. The method comprises the following specific steps:
the method comprises the following steps: and judging whether the virtual machine belongs to a network intensive application virtual machine or not by collecting network bandwidth information required by the application in the virtual machine. If the migration virtual machine is not the network intensive application virtual machine, jumping to the second step; if the virtual machine is a network intensive virtual machine, jumping to the fourth step;
step two: and judging whether the current idle bandwidth meets the transmission bandwidth of the next internal dirty page quantity (namely whether the network bandwidth is greater than the internal dirty page rate) according to the predicted internal dirty page quantity. If yes, no adjustment is carried out; otherwise, adjusting the network bandwidth, namely jumping to the third step;
step three: and the migration of non-network intensive applications is preferentially ensured, and the migration of network intensive virtual machines at the same time is delayed. Step four: and reserving application bandwidth according to the historical network transmission flow data. The bandwidth allocated for the virtual machine migration now is obtained by subtracting the application reserved bandwidth from the idle network bandwidth now.
The flow regulation and control are mainly realized by a flow control module, namely, a flow control (Trafficcontrol) module, which is mainly divided into 6 modules of an input interface (InputInterfaces), a flow limitation (IngressPolicing), an input Multiplexing (InputDe-Multiplexing), a Forwarding (Forwarding), a queue scheduling (QueueScheduling) and an output interface (OutputInterface). The input interface is responsible for receiving the data packet and transmitting the data packet to the flow restriction module. The flow limiting module is responsible for screening data packets, transmitting data packets which meet the regulations to the input multiplexing module and discarding data packets which do not meet the regulations. The input multiplexing module judges and analyzes the input data packet and determines the flow direction of the data packet. If the data packet is of the host, the data packet is handed to an upper layer (UpperLayer); otherwise, transmitting the data packet to a forwarding module for forwarding: the next hop for the packet is determined by looking at the routing table flow. The data packets are then transmitted to a queue scheduling module, which queues the data packets and transmits the data packets to the output interface in the queue order. The output interface transmits the data packet to the next hop in accordance with the transmission. Flow control is mainly handled and implemented when queues are scheduled and arranged.
In conclusion, when the virtual machine migration of the network intensive application or the memory intensive application is faced, the method can reduce the extra overhead in the migration process, improve the transmission efficiency in the migration process and effectively reduce the migration time.
Claims (1)
1. A virtual machine online migration optimization migration method for sensing composite application characteristics and network bandwidth is characterized by comprising the following steps:
(1) sensing application characteristics and a network environment of the virtual machine, and collecting the number of internal dirty pages; the sensing of the application characteristics of the virtual machine refers to the dynamic acquisition of the use condition of application resources in the virtual machine, including the memory use rate, the CPU use rate and the network bandwidth, and the trend of the change of the application resource use rate can be mastered; the perception of the network environment refers to the dynamic acquisition of the utilization rate of the network bandwidth of the cloud data center; the memory dirty page refers to a memory page modified in the migration process of the virtual machine;
(2) according to the application characteristics and the network environment of the perception virtual machine in the step (1) and the acquired internal dirty page number, using a grey prediction model to predict the internal dirty page number;
(3) calculating the dirty page rate of the iteration cycle of the virtual machine according to the internal memory dirty page number obtained by prediction in the step (2), wherein the dirty page rate is the dirty page number generated in the iteration cycle/the iteration cycle time, and reserving the network bandwidth for the step (5);
(4) obtaining the use condition of network bandwidth;
(5) judging whether the virtual machine is a network intensive virtual machine or not according to the network bandwidth use condition obtained in the step (4), and then reserving network bandwidth; in the network bandwidth reservation process, performing bandwidth reservation according to the network bandwidth and the dirty page rate required by the application of the virtual machine, so as to ensure that the network bandwidth is sufficient in each iteration period in the migration process and reduce network congestion;
the process of predicting the number of dirty pages in the next period by using the gray prediction model in the step (2) is as follows:
(21) for the collection of step (1) data, namely X1: migrating the quantity of internal memory dirty pages generated in different iteration cycles; x2: migrating the memory utilization rate of the virtual machine in different iteration periods; x3: migrating the CPU utilization rate of the virtual machine in different iteration cycles; x4: migrating the network bandwidth of the virtual machine in different iteration periods; x5: the time of the last iteration; x6: migrating the network use conditions of the cloud data centers in different iteration periods, converting the data into a matrix as a gray prediction sequence X(0)And for said sequence X(0)Accumulating to generate AGO sequence with X(1)；
(22) X obtained according to step (21)(1)Solving a sequence generated by the neighbor value, namely a mean sequence;
(23) in the gray prediction model, assume X of step (21)(1)Establishing a gray differential equation in relation to the step (22);
(24) establishing a whitening differential equation according to step (22);
(25) according to the equation set in the step (23), establishing an equation set based on GM (1, N), and solving a parameter sequence of the GM (1, N) equation by using a least square method;
(26) substituting the parameters into the whitening differential equation in the step (24), solving to obtain a GM (1, N) discrete response function, and solving to obtain a predicted value sequence through an accumulation reduction formula;
(27) in order to reduce the error of the grey prediction model, the prediction precision is improved by correcting the predicted value through the residual error;
the network bandwidth reservation process in the step (5) is as follows:
(31) collecting network bandwidth information required by the application in the virtual machine, judging whether the virtual machine belongs to a network intensive application virtual machine, and jumping to the step (32) if the migration virtual machine is not the network intensive application virtual machine; if the virtual machine is a network-intensive virtual machine, jumping to the step (34);
(32) calculating the internal memory dirty page rate of the virtual machine according to the step (3), and judging whether the current idle bandwidth meets the transmission bandwidth of the number of the internal memory dirty pages next, namely whether the network bandwidth is greater than the internal memory dirty page rate; if yes, no adjustment is carried out; otherwise, adjusting the network bandwidth, namely jumping to the step (33);
(33) preferentially ensuring non-network intensive application migration and delaying the network intensive virtual machine migration at the same time;
(34) and reserving application bandwidth according to the historical network transmission flow data, and subtracting the application reserved bandwidth from the current idle network bandwidth to obtain the bandwidth allocated for the current virtual machine migration.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611231403.2A CN106775949B (en) | 2016-12-28 | 2016-12-28 | Virtual machine online migration optimization method capable of sensing composite application characteristics and network bandwidth |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611231403.2A CN106775949B (en) | 2016-12-28 | 2016-12-28 | Virtual machine online migration optimization method capable of sensing composite application characteristics and network bandwidth |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106775949A CN106775949A (en) | 2017-05-31 |
| CN106775949B true CN106775949B (en) | 2020-08-18 |
Family
ID=58921091
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611231403.2A Active CN106775949B (en) | 2016-12-28 | 2016-12-28 | Virtual machine online migration optimization method capable of sensing composite application characteristics and network bandwidth |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106775949B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107479944B (en) * | 2017-07-20 | 2021-11-30 | 上海交通大学 | Virtual machine memory self-adaptive thermal migration scheduling method and system in hybrid cloud mode |
| CN108491248A (en) * | 2018-03-07 | 2018-09-04 | 山东大学 | A kind of triggering method and realization system of the dynamic migration of virtual machine based on prediction |
| CN109062669A (en) * | 2018-08-07 | 2018-12-21 | 郑州云海信息技术有限公司 | Virtual machine migration method and system under a kind of Random Load |
| CN109343931B (en) * | 2018-09-12 | 2021-02-12 | 西安交通大学 | Load balancing oriented application perception virtual machine scheduling method in IaaS environment |
| CN111399977A (en) * | 2020-02-29 | 2020-07-10 | 苏州浪潮智能科技有限公司 | Virtual machine migration method, device, equipment and medium |
| CN111381934B (en) * | 2020-03-11 | 2023-07-21 | 重庆邮电大学 | Method for improving quick migration reliability of virtual machine under disaster risk model |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102096461B (en) * | 2011-01-13 | 2013-06-19 | 浙江大学 | Energy-saving method of cloud data center based on virtual machine migration and load perception integration |
-
2016
- 2016-12-28 CN CN201611231403.2A patent/CN106775949B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102096461B (en) * | 2011-01-13 | 2013-06-19 | 浙江大学 | Energy-saving method of cloud data center based on virtual machine migration and load perception integration |
Non-Patent Citations (2)
| Title |
|---|
| 面向业务特征的自适应虚拟机迁移带宽分配算法;刘诗海等;《计算机学报2013年》;20130930;第36卷(第9期);第1816-1823页 * |
| 面向负载均衡的虚拟机动态迁移优化研究;李子堂;《中国优秀硕士学位论文全文数据库信息科技辑2016年》;20160115(第1期);第22-29、44页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106775949A (en) | 2017-05-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106775949B (en) | Virtual machine online migration optimization method capable of sensing composite application characteristics and network bandwidth | |
| CN111130904B (en) | Virtual network function migration optimization algorithm based on deep certainty strategy gradient | |
| CN111953758B (en) | Edge network computing unloading and task migration method and device | |
| CN107992353B (en) | Container dynamic migration method and system based on minimum migration volume | |
| CN110457131B (en) | Task scheduling method for supercomputing platform of power system based on Docker container | |
| CN111475252B (en) | Virtual network function deployment optimization method based on deep reinforcement learning | |
| CN103699433B (en) | One kind dynamically adjusts number of tasks purpose method and system in Hadoop platform | |
| CN104503838A (en) | Method for scheduling virtual CPU (Central Processing Unit) | |
| CN111538570B (en) | Energy-saving and QoS guarantee-oriented VNF deployment method and device | |
| CN106250240A (en) | A kind of optimizing and scheduling task method | |
| CN112087509A (en) | Task migration method in edge computing platform | |
| WO2021212933A1 (en) | Fpga-based concurrent graph data preprocessing method | |
| CN113904923A (en) | Service function chain joint optimization method based on software defined network | |
| CN108089914A (en) | A kind of cloud computing deploying virtual machine algorithm based on energy consumption | |
| CN112214301A (en) | Smart city-oriented dynamic calculation migration method and device based on user preference | |
| Zhou et al. | Cost-aware partitioning for efficient large graph processing in geo-distributed datacenters | |
| CN110597598B (en) | Control method for virtual machine migration in cloud environment | |
| CN110048966B (en) | Coflow scheduling method for minimizing system overhead based on deadline | |
| CN111901435B (en) | Load-aware cloud-edge collaborative service deployment method | |
| CN105393518B (en) | Distributed cache control method and device | |
| CN116302578B (en) | QoS (quality of service) constraint stream application delay ensuring method and system | |
| CN110850957B (en) | Scheduling method for reducing system power consumption through dormancy in edge computing scene | |
| CN116954905A (en) | Task scheduling and migration method for large Flink data | |
| CN114785692A (en) | Virtual power plant aggregation regulation and control communication network flow balancing method and device | |
| CN115361453A (en) | Load fair unloading and transferring method for edge service network |
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 |