CN107919986B - VM migration optimization method among MEC nodes in ultra-dense network - Google Patents
VM migration optimization method among MEC nodes in ultra-dense network Download PDFInfo
- Publication number
- CN107919986B CN107919986B CN201711098566.2A CN201711098566A CN107919986B CN 107919986 B CN107919986 B CN 107919986B CN 201711098566 A CN201711098566 A CN 201711098566A CN 107919986 B CN107919986 B CN 107919986B
- Authority
- CN
- China
- Prior art keywords
- node
- energy consumption
- user
- representing
- 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
- 238000013508 migration Methods 0.000 title claims abstract description 88
- 230000005012 migration Effects 0.000 title claims abstract description 87
- 238000000034 method Methods 0.000 title claims abstract description 40
- 238000005457 optimization Methods 0.000 title claims description 13
- 238000005265 energy consumption Methods 0.000 claims abstract description 74
- 230000002776 aggregation Effects 0.000 claims abstract description 49
- 238000004220 aggregation Methods 0.000 claims abstract description 49
- 230000005540 biological transmission Effects 0.000 claims abstract description 37
- 230000003993 interaction Effects 0.000 claims abstract description 4
- 238000004364 calculation method Methods 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 4
- 238000012546 transfer Methods 0.000 claims description 4
- 239000000835 fiber Substances 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 2
- 230000008859 change Effects 0.000 claims 1
- 230000000779 depleting effect Effects 0.000 claims 1
- 238000004891 communication Methods 0.000 abstract description 5
- 230000008901 benefit Effects 0.000 description 6
- 230000008569 process Effects 0.000 description 6
- 239000000243 solution Substances 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 230000001413 cellular effect Effects 0.000 description 3
- 238000009826 distribution Methods 0.000 description 3
- 238000004088 simulation Methods 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005562 fading Methods 0.000 description 2
- 238000010521 absorption reaction Methods 0.000 description 1
- 230000003321 amplification Effects 0.000 description 1
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 239000011557 critical solution Substances 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 229910052760 oxygen Inorganic materials 0.000 description 1
- 239000001301 oxygen Substances 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 208000000649 small cell carcinoma Diseases 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/08—Configuration management of networks or network elements
- H04L41/0803—Configuration setting
- H04L41/0813—Configuration setting characterised by the conditions triggering a change of settings
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention belongs to the technical field of wireless communication, and particularly relates to a method for optimizing VM migration between mobile edge computing nodes in an ultra-dense network. The applied ultra-dense network comprises a gateway node, an aggregation node and an edge node; when the VM of the MEC node needs to be migrated, firstly, the initialization characteristic is calculated, and according to the predicted migration time of the user, the energy consumption generated by interaction between the user and the VM when the user moves in the same gateway node range is calculated, wherein the energy consumption comprises the energy consumption W for transmitting data to a destination nodemigEnergy consumption W for data transmission to source nodepreAnd energy consumption W for data transmission in connection with VM in coverage areas of three nodes when user positions changeafter(ii) a Secondly, establishing an optimal profit model; and finally, solving the optimal revenue model and selecting the optimal VM migration strategy. The method can realize the flexible migration function of the VM in a special scene, effectively reduce the energy consumption of the system and improve the migration efficiency; and the resource collocation is reasonably carried out, and the service requirement of the user is met.
Description
Technical Field
The invention belongs to the technical field of wireless communication, and particularly relates to a VM (virtual machine) migration optimization method among Mobile Edge Computing (MEC) nodes in an ultra-dense network (UDN).
Background
In recent years, the demand of wireless services is rapidly increasing, and according to the prediction in the industry, the demand of wireless services will be thousands of times in the near future, the existing access network service equipment is difficult to meet the increasing high-speed demand, and the dense deployment of heterogeneous cells (smallcell) becomes the most critical solution of the problem. At the same time, in order to accommodate the flexibility and reduce costs of dense cells, wireless transmission schemes have become an important choice for cells to connect to the core network through gateway nodes. In addition, wireless backhaul in different bands, such as non-line-of-sight propagation (below 6GHz) and straight millimeter wave propagation from point to point, can be satisfactory for wireless transmission in certain situations. Wireless backhaul technology has become one of the backhaul solutions for cellular deployments.
The novel application and cloud service have strict requirements on the transmission delay and the calculation delay of resources, and further have strict requirements on the placement and the transmission conditions of the cloud and the edge calculation server. Deploying edge computing servers at cellular nodes, closer to the users, may provide sufficient transmission and computing resources. However, deployment of the server in the cellular node inevitably reduces the coverage of the server, and the mobility of the user is more frequent and complicated, and VM migration is more frequent, especially between heterogeneous cells. Heterogeneous backhaul links are more complex, and the VM migration technology needs to consider these novel factors to reduce energy consumption and time delay of the system, optimize throughput, and ensure overall service efficiency of the system.
The selection of the migration position and the migration mode are the main research works of the current VM migration technology. In the case of multiple users, resource occupation of users is a balancing problem, and when a VM migrates, not only the workload status of the relevant server, i.e., the working status of other users on the server, but also the transmission status and the mobility of the users are considered. Since the VM migration needs to consider the conditions of VM components and memory update, there are memory transfer and calculation costs in the migration process, and the migration benefit becomes a main decision condition for VM migration. To improve the benefits of migration, mobility prediction and cost prediction become new influencing factors. In addition, the migration time is reduced by changing the migration mode, for example, the VM interruption time is reduced by a mode of transmitting less data amount by a compression algorithm and the like and by a memory pre-migration mode, so that the migration cost is reduced, and therefore, not only the execution delay can be reduced, but also the throughput of the system can be increased, and the pressure of the system load can be reduced.
A heterogeneous MEC server framework is established in document [1] (l.tong, y.li and w.gao, "a hierarchical edge layer architecture for mobile computing," IEEE INFOCOM 2016-The 35th international conference on Computer Communications, San Francisco, CA,2016, pp.1-9.), a layered framework can provide more flexible services, ensure reasonable utilization of resources and service unloading efficiency of users, and solve The problem of deployment of edge servers under different time-varying task volumes, but do not specifically discuss backhaul links and VM migration problems in such a case.
A Wireless backhaul node deployed in a dense network is discussed in document [2] (m.n.islam, a.sampath, a.maharshi, o.koymen and n.b.mandayam, "Wireless backhaul node deployment for small cell networks," 201448 th annual conference on Information Sciences and Systems (CISS), Princeton, NJ,2014, pp.1-6.) that uses different bands for inter-node communication, enabling flexible deployment of cells, solving the deployment problem of vertical nodes in cities, but does not incorporate other application scenarios such as deployment of MEC computing resources, and VM migration related parts.
In an ultra-dense network scene, access nodes are densely deployed, an edge computing server is closer to a user, the service range of the edge computing server is also smaller, along with the movement of the user, a virtual machine on the edge server needs to be migrated to a proper node to ensure the service efficiency of the user and the energy consumption in the network, and the existing VM migration does not consider the deployment characteristics of a heterogeneous network; in addition, the deployment of many nodes requires high flexibility, wireless backhaul becomes a backhaul technology choice in the scenario, and the power and transmission rate of transmission of different bands are also different, which also need to be considered in VM migration.
Disclosure of Invention
Aiming at the problem that the characteristics of a heterogeneous network are not considered in VM migration in the current ultra-dense network scene, the invention provides a VM migration optimization method among MEC nodes in the ultra-dense network in order to reduce energy consumption in the ultra-dense network, improve VM migration efficiency and introduce a heterogeneous scene of a cell.
The invention provides a VM migration optimization method among MEC nodes in an ultra-dense network, wherein the applied ultra-dense network comprises a gateway node, a aggregation node and an edge node, the coverage range of the three types of nodes is from large to small, each heterogeneous access point is provided with an MEC server, the MEC node is a node with the MEC server, and when the VM of the MEC node needs to be migrated, the following steps are executed:
first, an initialization feature is computed, comprising: according to predicted migration time of user
Calculating energy consumption generated by interaction between a user and a VM when the user moves in the range of the same gateway node, wherein the energy consumption comprises energy consumption W of data transmission to a destination nodemigEnergy consumption W of data transfer to source nodepreAnd energy consumption W for data transmission in connection with VM in coverage areas of three nodes when user positions changeafter；
Energy consumption WafterThe method comprises three stages: in the first stage, a user is positioned in the coverage area of the edge node A where the VM is positioned, and the energy consumption of data transmission between the user and the VM is
In the second stage, the user is under the coverage of the edge node B, the edge nodes A and B are under the same aggregation node, and the energy consumption of the user and the VM for data transmission in the second stage is
In the third stage, the user moves to an edge node C, the edge nodes A and C are under the same gateway node, and the energy consumption of data transmission between the user and the VM is
Secondly, establishing an optimal profit model:
wherein,
representing the number of VMs, ψ representing the set of nodes in the network; g (l) when the value is 1, the VM is migrated to the node l, and g (l) when the value is 0, the VM is not migrated to the node l; dkRepresenting the resources that VM k needs to occupy; mlRepresenting the total amount of resources of node l.
And finally, solving the optimal revenue model and selecting the optimal VM migration strategy.
Compared with the prior art, the invention has the advantages and positive effects that:
(1) the method can realize the flexible migration function of the VM in a special scene, combines the technologies of wireless backhaul, user mobility prediction and the like, and can effectively reduce the energy consumption of the system and improve the migration efficiency according to the simulation result;
(2) the method of the invention considers the coverage areas of different nodes in a super-dense network scene, comprehensively considers a plurality of factors such as resource deployment, the mobility characteristics and requirements of users, channel conditions and the like according to the size of the coverage area of the nodes, selects proper nodes for the VM needing to be migrated, reasonably collocates resources, ensures the calculation requirements and meets the service requirements of the users.
Drawings
FIG. 1 is a system model diagram of an ultra-dense network;
FIG. 2 is a schematic main flow diagram of a VM migration optimization method provided by the present invention;
FIG. 3 is a graph of energy consumption variation for different scenarios with a linear increase in the number of users;
FIG. 4 is a graph of energy consumption revenue variation for different types of nodes as the user increases when the method of the present invention is applied;
FIG. 5 is a graph of energy consumption revenue variation for an average user for different types of situations when the method of the present invention is applied;
FIG. 6 is a graph of energy consumption and revenue variation of different types of nodes with the average memory footprint of the VM when the method of the present invention is applied.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples.
The VM migration optimization problem among MEC nodes in an ultra-dense network is a multi-server and multi-user problem, in the VM migration process, due to the fact that a VM has multiple position choices and multi-user changes are complex and random, a Nash equilibrium problem is formed under the condition of complete optimization, the complexity is greatly increased, calculation is difficult in a very short time, an energy consumption optimization strategy based on VM dynamic migration is a simplified algorithm, and the method takes the benefit maximization as an optimization target. The target equation has a certain resource occupation amount and also generates corresponding value, the selection problem can be converted into a knapsack problem, and dynamic programming is utilized to solve the problem. Due to the heterogeneity of the MEC servers, the improved distributed dynamic planning can ensure the maximization of the service benefits, can also ensure the overload problem of a single MEC, and ensures the service quality of users.
As shown in fig. 1, a super-dense network model, Gateway Nodes (GN) are located at the top of a building, and their special functions are to access a core network through optical fiber communication. Aggregation Nodes (ANs) are located on top of a building in which the cells are deployed, and may be connected to gateway Nodes on the one hand, and enter the core network through the gateway Nodes. Edge Nodes (ENs) are located inside buildings, and the Nodes directly provide network services for users. The three nodes are responsible for different tasks and different coverage areas of the nodes, and determine the size and the deployment of the servers of the nodes. In an ultra-dense network applied by the method, each access point is provided with an MEC server, and the MEC node refers to a node provided with the MEC server, for example, as shown in figure 1, the MEC node comprises a heterogeneous gateway node, an aggregation node and an edge node, and the signal coverage ranges of the gateway node, the aggregation node and the edge node are reduced in sequence.
According to the method for optimizing VM migration among MEC nodes in the ultra-dense network, disclosed by the invention, when a single user is used, a proper node is selected for VM migration according to the movement characteristics of the user, so that the working efficiency of a VM can be improved, the interruption condition caused by VM migration can be reduced, and the total energy consumption can be reduced. For the multi-user situation, the resource occupation and other situations need to be considered, the position of VM migration is adjusted, and the overall benefit is ensured.
In the migration process, in order to ensure the QoS of the user, the user equipment still needs to maintain a normal operating state with the VM at the initial position, and the longer the VM migration time is, the more the energy consumption is, and the relatively lower the migration frequency is. The method of the invention firstly reasonably predicts the moving time of the user based on the moving characteristics of the user, then calculates the energy consumption of migration, distributes according to the calculation result, and then optimally selects each node to obtain the best result.
Power p between node i and node ji,jCan be solved by a shannon formula as follows:
where r is the transmission rate, N0Is the channel noise, w represents the channel bandwidth, gi,jRepresenting the link gain.
wherein M isk,0Representing the initial memory size of the VM with the label k, wherein the labels of the VM are numbered according to a natural number sequence; r is the data transfer rate of VM migration, here set to a fixed value; dkIs the dirty page rate, representing the memory generation rate, where R > D must be satisfiedkOtherwise, the migration process cannot be converged, the VM pre-migration is deteriorated, and server resources are extremely occupied; mthIf the dirty page threshold is reached, the switching stage is entered.
An implementation flow of the VM migration optimization method according to the present invention is described below with reference to fig. 2.
The method comprises the following steps: first, whether the VM needs to be migrated is judged. If the amount of calculation data of the user is not large enough or the resources of the destination server are not sufficient, the migration operation is not required. If migration is required, the movement characteristics of the user are called.
Step two: and initializing feature calculation. Heterogeneous nodes result in multiple positions where a VM may be placedWhen a user moves to a location, which is within the coverage of three different nodes, the VM can theoretically migrate to any one of the three nodes, but the best profit needs to be selected. The invention aims at all different types of nodes of each task (namely VM) covering users, the energy consumption rate of the different nodes is calculated, the migration energy consumption is pre-estimated according to the movement characteristics of the users, then the relative profit is calculated, and the VM is distributed to the optimal MEC server node according to the profit. In the migration process of the VM, the connection of the links needs to be ensured, and a certain transmission rate needs to be ensured, which consumes the time and the migration time mentioned above
There is a direct relationship.
WmigThe energy consumed by transferring the configuration file contained in the VM and the necessary data cache to the destination node in the VM migration process is shown.
f (x) is a normalization function, and f (x) is 0 when x is 0, and f (x) is 1 when x is not equal to 0.
φkThe position of the virtual machine k is represented by (i, j, z), i represents the gateway node serial number of the node where the VM is located, j is the aggregation node serial number of the node where the VM is located, z represents the edge node serial number of the node where the VM is located, if the VM is not located on the edge node, z is 0, if the VM is not located on the aggregation node, j is 0, i is the gateway node code number and is not 0, the VM is not necessarily located on the edge node or the aggregation node server, but is necessarily located under the coverage of one gateway node, wherein the VM is not necessarily located on the edge node or the aggregation node server, but is always
For the code numbers of different types of nodes, superscripts t +1 and t represent two different time periods, wherein t represents the last time period and generally represents the source node position of the VM, and t +1 represents the current position of the user.
Respectively represent the edge nodes where the VM is located at the previous moment and the current moment,
respectively represent the aggregation nodes where the VM is located at the previous moment and the current moment,
respectively representing the gateway nodes where the VMs are located at the previous time and the current time. To be provided with
For example, if
Is 1, indicating that VM k has the code number of
Run on the edge node.
The migration of the VM relates to four kinds of inter-node migration, including the migration between the edge node and the aggregation node, and the migration between the aggregation node and the gateway node. Let en denote edge node, an denote aggregation node, D denote destination, and R denote source.
Representing the power consumed by task k when it is transmitted from the aggregation node to the edge node,
representing the power consumed by task k when it is transmitted from the gateway node to the aggregation node,
representing the power consumed by task k when it is transmitted from the aggregation node to the gateway node,
representing the power consumed by the task k edge node in transmitting to the aggregation node.
Representing the power consumed by the task k fiber transmission.
Calculating energy W consumed by data transmission in VM (virtual machine) migration processpreRepresenting that the calculated data still needs to be transmitted to the source server within the migration time, wherein the transmission energy consumption is as follows:
after the VM migration is completed, the position of the user is changed, the position of the user can be divided into three types of areas, therefore, the consumption can also be divided into three stages, and the consumption W of data transmission is calculatedafter：
1) In the first stage, the time is
This stage is the energy consumption time of the user at the initial edge node, the energy consumption of the user:
wherein T represents time, and T is settier_3、Ttier_2、Ttier_1Respectively representing the predicted residence time of the user under the same edge node, the same aggregation node and the same gateway node with the VM.
Representing the time required for VM migration. As shown in equation (5), if the VM migrates to the edge node, the VM migrates to the edge node
A value of 0 indicates that the VM can serve the user at the edge node, and there is no energy consumption between nodes; if the data is migrated to the aggregation node or the gateway node, data transmission is required between the nodes at this time, which results in energy consumption.
2) In the second stage, the time is Ttier_2-Ttier_3The period is the energy consumption time of the user in the same aggregation node but different edge nodes, the user leaves the primary edge node but still is in the coverage area of the same aggregation node, and the energy consumption of the period is as follows:
where en' denotes another edge node located under the same aggregation node an as the edge node en.
Representing the power consumed by the task k when it is transmitted from the aggregation node to the other edge node en'.
3) At the third stage, the time is Ttier_1-Ttier_2And represents the energy consumption time of the user in the same gateway node but different aggregation nodes. This stage is under the coverage of the gateway node, so it is not necessary to consider the case of VM migration again. The energy consumption is as follows:
wherein en 'represents that the user position is under the edge node of another aggregation node range within the same gateway node range, and an' represents another aggregation node within the same gateway node range.
And through the initialized characteristic calculation in the step, judging the energy consumption generated by interaction with the VM when the user moves in the same gateway node range according to the predicted time of the user.
Step three: and calculating the distribution optimal profit. After the initialization calculation is completed, tasks are pre-distributed to corresponding stages, and the total profit is calculated by the nodes according to the resource amount of the nodes and the occupied space and the value of the users as follows:
wherein,
representing the number of users or VMs, which are in one-to-one correspondence. Psi denotes the set of all nodes. g (l) indicates whether VM is migrated to node l, MlRepresenting the total amount of resources of node l. DkIndicating the resources that user k or VM k needs to occupy. The constraint condition in equation (8) is to ensure that the resources occupied by the VMs on a node do not exceed the maximum resource value of the node.
The invention uses the energy consumption W of VM without migrationstaticAs a reference line,
can be changed by changing WpreThe conditions are found as follows:
the baseline energy consumption of VM k is
Obtained according to the formula (9)
Then will be
The difference value of (a) is used as a profit target, and a target expression for calculating the optimal profit can be converted into:
because server distribution exists, the overall optimal solution is required to be solved, the calculation complexity and the time consumption are very high, and therefore the invention aims to solve the suboptimal solution, namely each server recursively solves the local optimal solution through dynamic programming, and the iterative core formula is as follows:
W(n,m)＝max{W(n-1,m),W(n-1,m-Dn)+wk} (10)
wherein n is iteration number, M is resource occupation amount of the MEC server on the node, and the maximum value of the resource occupation amount of the server is set as M. w is akIs the expected revenue for a single VM k,
in this way, the maximum gain can be determined.
Step four: and (5) calculating the cyclic filling. And in the third step, the VM without resource storage selects a higher-layer server node, and performs primary solution and secondary iterative solution on the original basis.
The invention provides a VM (virtual machine) migration scheme based on heterogeneous nodes, which introduces a network system model, describes a VM migration scene and specific amplification, and finally gives performance simulation. The method can effectively reduce the energy consumption of the system and simultaneously improve the service quality of the user.
Example (b):
consider the following scenario: two gateway nodes, four aggregation nodes and twelve edge nodes. As shown in table 1, this embodiment selects a 5.8GHz bandwidth as over-the-horizon transmission and 60GHz as line-of-sight. The maximum number of users is set to 10000, the requirement of the transmission rate meets normal distribution N (300Mbps,100Mbps), the resource occupation amount is randomly distributed, and the range is [5,20 ]](ii) a Dirty page rate of [10, 20%]k/s, the transmission rate of the migration data is set as 200Mbps, and the Gaussian noise is 104The time of the user under the three nodes is randomly generated, and the generation rule is Ttier_1∈(Ttier_2,3Ttier_2)，Ttier_2∈(Ttier_1,3Ttier_1)，Ttier_3∈(0,7200)s。
TABLE 15.8 GHz and 60GHz Wireless channel characteristic Table
| Feature(s) | 5.8GHz | 60GHz |
| Rain-induced fading (dB) | 0 | 10 |
| Oxygen absorption (dB) | 0 | 15 |
| Channel gain (dB) | 17 | 38 |
| Maximum transmission power (dBm) | 19 | 25 |
| Fading margin (dB) | 15 | 25 |
| Channel bandwidth (MHz) | 40 | 160 |
| Number of |
6 | 6 |
For the above scenario, the method of the present invention is used for performing a simulation experiment on the VM with a user migration scheme and a non-migration scheme. Fig. 3 shows a graph of the energy consumption variation for three different scenarios with a linear increase in the number of users, wherein the curve of the method of the invention employs a selection migration strategy. The graph is mainly embodied in the situation of different users, the total energy consumption of the graph is gradually increased, wherein the state of the scheme is not migrated and is almost in a linear state, and the selective migration has better convergence than the non-migration, but has better performance when the user quantity is increased to a certain degree. In general, under normal conditions, the method has a good effect of saving energy consumption.
Fig. 4 is a graph of the total energy yield of the MEC servers of each class of energy consumption with the number of users in the optimization strategy. Compared with a non-migration state, the most profit is the edge node, because the edge node is the node closest to the user, if the user's feature belongs to a certain place for a long time, and if the VM migrates to the edge node, the energy consumption rate is the lowest at this time, the profit of the VM migrated to the edge node is the largest, which becomes the main part of the total profit, so the deployment of the edge node becomes the key.
Fig. 5 shows the average energy consumption per user, and as the number of users increases, the profit on different server nodes decreases, because the increase of users and the conflict of resources cause that some VMs make secondary server selection, and the profit inevitably decreases. Both of these two points indicate that more appropriate edge nodes need to be deployed and more resources need to be provided to satisfy the flexible migration of VMs in order to better achieve the effect of reducing energy consumption.
In fig. 6, a graph of energy consumption and revenue variation with the average memory footprint of the VM is shown. As the amount of resources of the VM increases, the energy consumption gains of the different types of servers are reduced. On different types of servers, the energy consumption of the gateway node is the largest, which shows that the mobility of a user is higher, and the larger the coverage of the VM migration position is, the higher the profit is, so that when server resources are deployed, the larger the coverage of the VM migration position is, such as the gateway node, the more resources can be deployed, but the side length of the transmission distance of the gateway node may sacrifice a part of transmission delay, thereby reducing the task computation rate.
Claims (7)
1. A VM migration optimization method among MEC nodes in a super-dense network is characterized in that the applied super-dense network comprises gateway nodes, aggregation nodes and edge nodes, each node is provided with an MEC server, wherein the node provided with the MEC server is called an MEC node, the MEC represents mobile edge calculation, and the VM represents a virtual machine;
when the VM of the MEC node needs to be migrated, the method executes the following steps:
first, an initialization feature is computed, comprising: according to predicted migration time of user
Calculating energy consumption generated by interaction between a user and a VM when the user moves in the range of the same gateway node, wherein the energy consumption comprises energy consumption W of data transmission to a destination nodemigEnergy consumption W of data transfer to source nodepreAnd energy consumption W for connecting with VM in coverage areas of three nodes when user position changesafter；
Energy consumption WafterThe method comprises three stages: in the first stage, a user is positioned in the coverage area of the edge node A where the VM is positioned, and the energy consumption of data transmission between the user and the VM is W1 after(ii) a In the second stage, the user is under the coverage of the edge node B, the edge nodes A and B are under the same aggregation node, and the energy consumption of the user and the VM for data transmission in the second stage is
In the third stage, the user moves to an edge node C, the edge nodes A and C are under the same gateway node, and the energy consumption of data transmission between the user and the VM is
Secondly, an optimal profit model is established as follows:
g(x)∈(0,1)
wherein,
representing the number of VMs, ψ representing the set of nodes in the network; g (l) represents whether the VM is migrated to the node l, the migration is represented by the value of 1, and the non-migration is represented by the value of 0; dkRepresenting the resources that VM k needs to occupy; mlRepresents the total amount of resources of the node l;
and finally, solving the optimal revenue model and selecting the optimal VM migration strategy.
2. The method of claim 1, wherein the migration time of the user is determined by a user profile
Calculated according to the following formula:
wherein M isk,0Represents the initial memory size of the VM with the label k; r is the data transmission rate of VM migration; dkIs the dirty page rate; mthIs the dirty page threshold.
3. Method according to claim 1, characterized in that the energy consumption W for the data transmission to the destination nodemigCalculated according to the following formula:
wherein, f (x) is a normalization function, and when x is 0, f (x) is 0, and when x is not equal to 0, f (x) is 1; f (x) 0 means that VM is not running on node x, and f (x) 1 means that VM is running on node x;
let phikThe position of the virtual machine k is represented by (i, j, z), i represents a gateway node serial number of a node where the VM is located, j represents an aggregation node serial number of the node where the VM is located, and z represents an edge node serial number of the node where the VM is located;
representing the edge node where the VM k is located at the current moment;
representing the aggregation node where the VM k is located at the current moment;
respectively representing an edge node and an aggregation node where the VM k is located at the previous moment; wherein
For the codes of different types of nodes, superscripts t +1 and t represent two different time periods, wherein t represents the last time period and represents the position of a source node of a VM, and t +1 represents the current position of a user;
4. Method according to claim 3, characterized in that said energy consumption W ispreObtained according to the following formula:
wherein,
representing the power consumed by task k when it is transmitted from the aggregation node to the edge node,
representing the power consumed by task k when it is transmitted from the gateway node to the aggregation node,
representing the power consumed by task k when it is transmitted from the edge node to the aggregation node,
representing the power consumed by task k when it is transmitted from the aggregation node to the gateway node,
representing the power consumed by the fiber transmission of task k.
5. The method of claim 3, wherein said depleting WafterDividing into three stages according to the position change of the user, and setting Ttier_3、Ttier_2、Ttier_1Respectively indicates that the user is at the same edge node with the VM,Predicting residence time under the same aggregation node and the same gateway node;
(1) in the first stage, the time is
Energy consumption W1 afterComprises the following steps:
wherein,
representing the power consumed by task k when it is transmitted from the gateway node to the aggregation node an,
representing the power consumption of the task k when it is transmitted from the aggregation node an to the edge node en;
(2) in the second stage, the time is Ttier_2-Ttier_3Energy consumption
Comprises the following steps:
wherein,
representing the energy consumption power when the task k is transmitted from the aggregation node an to another edge node en ', and the node en' and the en are under the aggregation node an;
(3) at the third stage, the time is Ttier_1-Ttier_2Energy consumption
Comprises the following steps:
wherein,
representing the energy consumption power when the task k is transmitted from the aggregation node an to another edge node en ', wherein the node en' and the en are under different aggregation nodes under the same gateway node;
representing the power consumed by the task k when it is transmitted from the gateway node to the aggregation node an ', both nodes an' and an being under one gateway node.
6. The method according to claim 1 or 3, wherein when the optimal profit model is solved, each VM uses the energy consumption without migration as a reference energy consumption;
reference energy consumption
Ttier_1Representing the predicted residence time of the user under the same gateway node with the VM;
for user k, the reference energy consumption is set as
Total energy consumption resulting from interacting with VMs
Will be provided with
Taking the difference as a profit target, and converting a target expression for calculating the optimal profit into:
7. the method of claim 6, wherein the optimal profit model is solved by dynamically and recursively solving for each node a local optimal solution;
let the n-th iteration of a node have an energy consumption W (n, m) of max { W (n-1, m), W (n-1, m-D)n)+wk}；
Where n is the number of iterations, m is the resource occupancy of the MEC server on the node, DnIs the memory size of the VM, wkIs the expected revenue for VM k and,
and when no resource is stored in the node for VM k, selecting a higher-layer server node for VM k, and then carrying out recursive iterative solution on the node.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711098566.2A CN107919986B (en) | 2017-11-09 | 2017-11-09 | VM migration optimization method among MEC nodes in ultra-dense network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711098566.2A CN107919986B (en) | 2017-11-09 | 2017-11-09 | VM migration optimization method among MEC nodes in ultra-dense network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107919986A CN107919986A (en) | 2018-04-17 |
| CN107919986B true CN107919986B (en) | 2020-09-08 |
Family
ID=61896038
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711098566.2A Active CN107919986B (en) | 2017-11-09 | 2017-11-09 | VM migration optimization method among MEC nodes in ultra-dense network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107919986B (en) |
Families Citing this family (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108984307B (en) * | 2018-07-19 | 2020-11-10 | 中国联合网络通信集团有限公司 | Computing task migration method and computing task migration device |
| CN109067583A (en) * | 2018-08-08 | 2018-12-21 | 深圳先进技术研究院 | A kind of resource prediction method and system based on edge calculations |
| CN109376374B (en) * | 2018-09-01 | 2023-04-07 | 哈尔滨工程大学 | Multi-user computing migration method based on multi-radio frequency communication |
| CN109151864B (en) * | 2018-09-18 | 2022-01-18 | 贵州电网有限责任公司 | Migration decision and resource optimal allocation method for mobile edge computing ultra-dense network |
| CN109375999A (en) * | 2018-10-23 | 2019-02-22 | 北京工业大学 | A kind of MEC Random Task moving method based on Bayesian network |
| CN109656681B (en) * | 2018-12-03 | 2022-12-02 | 华中科技大学 | Energy scheduling method in cloud fusion environment |
| CN109710374A (en) * | 2018-12-05 | 2019-05-03 | 重庆邮电大学 | The VM migration strategy of task unloading expense is minimized under mobile edge calculations environment |
| CN109802934A (en) * | 2018-12-13 | 2019-05-24 | 中国电子科技网络信息安全有限公司 | A kind of MEC system based on container cloud platform |
| DE102018009903A1 (en) * | 2018-12-20 | 2020-06-25 | Volkswagen Aktiengesellschaft | Device for a vehicle for outsourcing computing power |
| CN110753117B (en) * | 2019-10-24 | 2022-03-04 | 南京信息工程大学 | Computing migration method giving consideration to privacy protection under wireless metropolitan area network environment |
| CN110933609A (en) * | 2019-11-26 | 2020-03-27 | 航天科工网络信息发展有限公司 | Service migration method and device based on dynamic environment perception |
| CN113098917B (en) * | 2019-12-23 | 2024-06-04 | 华为云计算技术有限公司 | Method for migrating functional nodes and related equipment |
| CN111132253B (en) * | 2019-12-31 | 2021-03-30 | 北京邮电大学 | Joint mobility management method for communication switching and service migration |
| CN111290835B (en) * | 2020-01-22 | 2023-03-10 | 重庆锦禹云能源科技有限公司 | Method and device for migrating virtual machine of cloud platform and server |
| CN113301092B (en) * | 2020-07-31 | 2022-04-12 | 阿里巴巴集团控股有限公司 | Network reconnection method, device, system and storage medium |
| CN112822055B (en) * | 2021-01-21 | 2023-12-22 | 国网河北省电力有限公司信息通信分公司 | Edge computing node deployment method based on DQN |
| CN113115256B (en) * | 2021-04-14 | 2022-04-19 | 重庆邮电大学 | Online VMEC service network selection migration method |
| CN113206694B (en) * | 2021-04-22 | 2023-04-07 | 南京航空航天大学 | Computing efficiency optimization method in millimeter wave mobile edge computing system based on hybrid beam forming |
| CN113656952B (en) * | 2021-08-04 | 2024-04-05 | 之江实验室 | Modeling tool for cloud edge computing environment |
| CN114096006B (en) * | 2021-08-18 | 2024-07-05 | 东南大学 | Resource allocation and data compression combined optimization method in mobile edge computing system |
| CN115765826B (en) * | 2022-09-19 | 2024-05-31 | 西安电子科技大学 | Unmanned aerial vehicle network topology reconstruction method for on-demand service |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106454948A (en) * | 2016-10-11 | 2017-02-22 | 重庆邮电大学 | Node re-distribution method in wireless network virtualization |
| CN106775933A (en) * | 2016-11-29 | 2017-05-31 | 深圳大学 | A kind of virtual machine on server cluster dynamically places optimization method and system |
-
2017
- 2017-11-09 CN CN201711098566.2A patent/CN107919986B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106454948A (en) * | 2016-10-11 | 2017-02-22 | 重庆邮电大学 | Node re-distribution method in wireless network virtualization |
| CN106775933A (en) * | 2016-11-29 | 2017-05-31 | 深圳大学 | A kind of virtual machine on server cluster dynamically places optimization method and system |
Non-Patent Citations (2)
| Title |
|---|
| Bandwidth Management VMs Live Migration in Wireless Fog Computing for 5G Networks;Danilo Amendola 等;《2016 5th IEEE International Conference on Cloud Networking (Cloudnet)》;20161208;第21-26页 * |
| 多天线超密集网络统计性能分析与优化;陈正;《中国博士学位论文全文数据库 信息科技辑（月刊 ）》;20170915;第I136-30页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107919986A (en) | 2018-04-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107919986B (en) | VM migration optimization method among MEC nodes in ultra-dense network | |
| CN109413724B (en) | MEC-based task unloading and resource allocation scheme | |
| Liang et al. | Multi-cell mobile edge computing: Joint service migration and resource allocation | |
| CN108809695B (en) | Distributed uplink unloading strategy facing mobile edge calculation | |
| Liang et al. | A cluster-based energy-efficient resource management scheme for ultra-dense networks | |
| Ari et al. | Resource allocation scheme for 5G C-RAN: a Swarm Intelligence based approach | |
| CN108964817B (en) | Heterogeneous network joint computing unloading and resource allocation method | |
| Lin et al. | Three-tier capacity and traffic allocation for core, edges, and devices for mobile edge computing | |
| CN107613556B (en) | Full-duplex D2D interference management method based on power control | |
| Li et al. | Mobility-aware offloading and resource allocation in NOMA-MEC systems via DC | |
| CN112512065B (en) | Method for unloading and migrating under mobile awareness in small cell network supporting MEC | |
| El Haber et al. | Macro-cell assisted task offloading in MEC-based heterogeneous networks with wireless backhaul | |
| Li et al. | Reliability and robust resource allocation for Cache-enabled HetNets: QoS-aware mobile edge computing | |
| CN107911856A (en) | A kind of separation multi-access method based on matching game in super-intensive heterogeneous network | |
| Sharma et al. | Energy saving in heterogeneous cellular network via transfer reinforcement learning based policy | |
| CN112235387B (en) | Multi-node cooperative computing unloading method based on energy consumption minimization | |
| Zeng et al. | Energy minimization for delay constrained mobile edge computing with orthogonal and non-orthogonal multiple access | |
| Pang et al. | Joint wireless source management and task offloading in ultra-dense network | |
| Nguyen et al. | Computation offloading and resource allocation for backhaul limited cooperative MEC systems | |
| JP5558551B2 (en) | Wireless communication resource management for distributed cellular systems | |
| El Haber et al. | Computational cost and energy efficient task offloading in hierarchical edge-clouds | |
| Zhou et al. | Multi-server federated edge learning for low power consumption wireless resource allocation based on user QoE | |
| CN115278779A (en) | Rendering perception-based dynamic placement method for VR service module in MEC network | |
| Li et al. | Mobile edge computing for task offloading in small-cell networks via belief propagation | |
| Xia et al. | Dynamic task offloading and resource allocation for heterogeneous MEC-enable IoT |
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 |