CN104519106A - Task migration method and network controller - Google Patents

Abstract

The invention discloses a task migration method and a network controller. The task migration method includes that the network controller receives task execution requests of equipment and selects task service nodes according to energy consumption and internet data traffic, wherein task service node selection includes executing tasks on the local equipment, migrating to other equipment in a network to execute the tasks by cooperation or migrating to remote cloud equipment to execute the tasks. The task migration method and the network controller have the advantages that computation redundancy and data transmission redundancy in the network can be eliminated to lower power consumption of mobile equipment and internet data traffic demand, so that average energy consumption of the equipment in the network is minimized, and external internet data traffic is controlled.

Description

A kind of task immigration method and network controller

Technical field

The present invention relates to wireless communication technology field, particularly relate to a kind of task immigration method and network controller.

Background technology

Along with the progress of computer technology and wireless communication technology, the mobile device taking smart mobile phone as representative, from basic communication instrument, progressively develops into acquisition instrument and the processing platform of information.Mobile subscriber can by various ways accessing Internets such as 3G, WIFI, browsing page, reading Email, shopping, amusement etc. anywhere or anytime.But mobile device still exists a certain distance, especially battery life with permanent plant in CPU disposal ability, memory capacity etc., become the Main Bottleneck of restriction Mobile solution development.

A kind of effective ways solving mobile device resource restricted problem are by task migrating technology, and performing the task immigration of mobile device to resourceful permanent plant or server goes, is mobile device conserve energy, reduces expense, shortens and performs delay.Calendar year 2001, the people such as A.Balasubramanian propose the concept of " network is looked for food ", are intended to the calculation task on mobile device to move to long-distance support on idle fixing computer, reduce its energy consumption while strengthening mobile device computing capability with this.In recent years, along with the development of mobile cloud computing technology, there is the computation migration technological frame of a lot of energy optimization.MAUI and ThinkAir provides a kind of computation migration method of method rank, and does not need any extra support of operating system.But such computation migration technology needs the source code of programmer's access application, artificial division is carried out to program.Clonecloud and Cloudlet utilizes virtual machine technique, beyond the clouds or on powerful server, for mobile device sets up application execution environment.Such computation migration technology does not need programmer's application programs to do any change, but needs the support of operating system.SociableSense is a kind of computation migration system set up for social type application program.COMET system then achieves other computation migration of thread-level not needing source program to be done to any change, and its applicability is stronger.

These technical schemes are all pay close attention to how static or dynamic application programs to divide above, and the computation migration of how implementation method level or thread-level.How their target mainly saves the calculating energy consumption of single mobile device by computation migration.

We notice, although the coverage of 3G network is higher than WIFI network, directly realize still there are some problems based on the computation migration of cloud computing by 3G.The bandwidth of 3G network is relatively low, but transmission delay and transmission energy consumption are higher than WIFI network.In addition, the growth rate of 3G network capacity (network is eated dishes without rice or wine) cannot meet the mobile Internet traffic demand increased rapidly.

The people such as S.Ha have designed and developed time-based pricing system, and object is by the Dynamic Pricing in different time sections, better manage the 3G network traffic demand increased rapidly.K.Lee points out under study for action, and in urban environment, when not by any delay transmission policy, WIFI network has shared the network traffic data of 3G network 65%, and has saved the energy content of battery of 55% for mobile device.They find, the time of mobile subscriber average 70% is in (time of 63% is daytime) in WIFI overlay environment, and the average time once stopped was more than 2 hours, and after leaving WIFI overlay environment, the time of again getting back to WIFI overlay environment is 40 minutes.More data, mainly through predicting that WIFI's can transmit data by connectivity delay, are transmitted by WIFI network, are alleviated with this pressure alleviating 3G network further by Wiffler.The transfer of data streaming system of the energy optimization that the people such as M.-H.Chen open, the data allowing the close mobile subscriber in position to be downloaded by WIFI Connection Sharing, reduce the connection requirement of 3G network with this, reduce Internet Transmission redundancy.By above technology, we recognize, WIFI network has been regarded as the main technique methods alleviating 3G network pressure.

But, in large-scale wireless local area network (LAN) (such as enterprise network), simultaneously online mobile device quantity gets more and more, and a large amount of mobile device is by same access point or network controller access the Internet, cause serious network congestion, cause actual available bandwidth to reduce rapidly.For network manager, reduce the energy consumption of mobile device from the angle of the whole network, control external flows has been a more and more important problem.

Summary of the invention

The technical problem to be solved in the present invention is to provide a kind of task immigration method and network controller, is intended to the average energy consumption minimizing equipment in network, controls external Internet data traffic simultaneously.

For solving the problems of the technologies described above, a kind of task immigration method of the present invention, comprising:

The tasks carrying request of network controller receiving equipment, according to energy consumption and Internet data traffic, selects the service node of task;

Wherein, the service node of described selection task comprises:

Execute the task in equipment this locality;

Move on other equipment in network and cooperate to execute the task; Or,

Move to long-range cloud device to execute the task.

Further, described according to energy consumption and Internet data traffic, select the service node of task, comprising:

According to ${\mathrm{DPP}}_{\mathrm{bound}}=V\•\mathrm{\ϵ}\left(C_h^i\right)+{\mathrm{\Σ}}_{j=1}^n{Q}_j\left(t_h\right)(b_i\left(t_h\right)-d_j\left(t_h\right))$ Other equipment respectively in computing equipment this locality, network and long-range cloud device perform DPP corresponding to described task _bound;

Select minimum DPP _boundcorresponding equipment is as the service node of task;

Wherein, DPP _boundfor drift penalty factor, V is the coefficient of balance realizing the whole network energy consumption of mobile equipment and external Internet data traffic tradeoffs, expression task, i is the numbering of the equipment sending tasks carrying request, and h is mission number, and j is device numbering, and n is the sum of the equipment in network, for the energy consumption needed of executing the task, Q _j(t _h) for equipment time grain t _hbill length, b _i(t _h) be the length that after the equipment transportation task of transmission tasks carrying request, self bill increases, d _j(t _h) for equipment is the length that after other equipment are executed the task, self bill reduces, described bill length associates with Internet data traffic.

Further, according to $\mathrm{\ϵ}\left(C_h^i\right)={\mathrm{\Σ}}_{j=1}^{n+1}{\mathrm{\α}}_j\left(C_h^i\right)\·{\mathrm{\ϵ}}_O^{\mathrm{ij}}\left(C_h^i\right)$ Described in calculating

Wherein, j=1 ~ n represents the equipment in network, and j=n+1 represents long-range cloud device;

${\mathrm{\ϵ}}_O^{\mathrm{ij}}\left(C_h^i\right)=\left\{\begin{array}{c}{\mathrm{\ϵ}}_L^{\mathrm{ij}}\left(C_h^i\right)+{\mathrm{\ϵ}}_R^{\mathrm{ij}}\left(C_h^i\right),j=1,...,n;\\ {\mathrm{\ϵ}}_L^{\mathrm{ij}}\left(C_h^i\right),j=n+1;\end{array}\right.,$ for initiating the energy consumption of the equipment of tasks carrying request, the energy consumption of the equipment of executing the task.

Further, according to ${\mathrm{\ϵ}}_R^{\mathrm{ij}}\left(C_h^i\right)={\mathrm{\ρ}}_T^{\mathrm{ij}}\·\frac{D_h^{\mathrm{in}}}{{\mathrm{\φ}}_T^{\mathrm{ij}}}+{\mathrm{\ρ}}_T^{\mathrm{ij}}\·\frac{D_h^{\mathrm{out}}}{{\mathrm{\φ}}_T^{\mathrm{ij}}}+{\mathrm{\ϵ}}_C^j\left(C_h^i\right)$ Described in calculating wherein, represent from equipment i to the through-put power of equipment j, represent the transmission rate current time is from equipment i to the wireless link of equipment j, the input size of data of expression task, represent the transmission rate current time is from equipment j to the wireless link of equipment i, the output size of data of expression task;

According to ${\mathrm{\ϵ}}_R^{\mathrm{ij}}\left(C_h^i\right)={\mathrm{\ρ}}_T^{\mathrm{ji}}\•\frac{D_h^{\mathrm{in}}}{{\mathrm{\φ}}_T^{\mathrm{ij}}}+{\mathrm{\ρ}}_T^{\mathrm{ji}}\•\frac{D_h^{\mathrm{out}}}{{\mathrm{\φ}}_T^{\mathrm{ji}}}+{\mathrm{\ϵ}}_C^j\left(C_h^i\right)$ Described in calculating wherein, represent from equipment j to the through-put power of equipment i, represent the calculating energy consumption of executing the task.

Further, according to described in calculating wherein, for the processor power of equipment, f _jfor the processor clock frequency of equipment, X is complexity factor.

Further, also comprise after the service node of the task of selection:

If task performed by equipment i this locality, then keep the bill length of equipment i constant;

If task performed by long-range cloud device, then the bill length of equipment i is increased b _i(t _h);

If task performed by other equipment in network, then the bill length of equipment i is increased b _i(t _h), the bill length of the equipment of executing the task is reduced d _j(t _h).

Further, a kind of network controller, comprising: receiving element and service sensor selection problem unit, wherein:

Described receiving element, for the tasks carrying request of receiving equipment;

Described service node selected cell, for according to energy consumption and Internet data traffic, selects the service node of task;

Wherein, the service node of described selection task comprises:

Execute the task in equipment this locality;

Move on other equipment in network and cooperate to execute the task; Or,

Move to long-range cloud device to execute the task.

Further, described service node selected cell, according to energy consumption and Internet data traffic, is selected the service node of task, being comprised:

According to ${\mathrm{DPP}}_{\mathrm{bound}}=V\•\mathrm{\ϵ}\left(C_h^i\right)+{\mathrm{\Σ}}_{j=1}^n{Q}_j\left(t_h\right)(b_i\left(t_h\right)-d_j\left(t_h\right))$ Other equipment respectively in computing equipment this locality, network and long-range cloud device perform DPP corresponding to described task _bound;

Select minimum DPP _boundcorresponding equipment is as the service node of task;

Wherein, DPP _boundfor drift penalty factor, V is the coefficient of balance realizing the whole network energy consumption of mobile equipment and external Internet data traffic tradeoffs, expression task, i is the numbering of the equipment sending tasks carrying request, and h is mission number, and j is device numbering, and n is the sum of the equipment in network, for the energy consumption needed of executing the task, Q _j(t _h) for equipment time grain t _hbill length, b _i(t _h) be the length that after the equipment transportation task of transmission tasks carrying request, self bill increases, d _j(t _h) for equipment is the length that after other equipment are executed the task, self bill reduces, described bill length associates with Internet data traffic.

Further, described service node selected cell according to described in calculating

Wherein, j=1 ~ n represents the equipment in network, and j=n+1 represents long-range cloud device;

${\mathrm{\ϵ}}_O^{\mathrm{ij}}\left(C_h^i\right)=\left\{\begin{array}{c}{\mathrm{\ϵ}}_L^{\mathrm{ij}}\left(C_h^i\right)+{\mathrm{\ϵ}}_R^{\mathrm{ij}}\left(C_h^i\right),j=1,...,n;\\ {\mathrm{\ϵ}}_L^{\mathrm{ij}}\left(C_h^i\right),j=n+1;\end{array}\right.,$ for initiating the energy consumption of the equipment of tasks carrying request, the energy consumption of the equipment of executing the task.

Further, described service node selected cell according to described in calculating wherein, represent from equipment i to the through-put power of equipment j, represent the transmission rate current time is from equipment i to the wireless link of equipment j, the input size of data of expression task, represent the transmission rate current time is from equipment j to the wireless link of equipment i, the output size of data of expression task; Further,

According to ${\mathrm{\ϵ}}_R^{\mathrm{ij}}\left(C_h^i\right)={\mathrm{\ρ}}_T^{\mathrm{ji}}\•\frac{D_h^{\mathrm{in}}}{{\mathrm{\φ}}_T^{\mathrm{ij}}}+{\mathrm{\ρ}}_T^{\mathrm{ji}}\•\frac{D_h^{\mathrm{out}}}{{\mathrm{\φ}}_T^{\mathrm{ji}}}+{\mathrm{\ϵ}}_C^j\left(C_h^i\right)$ Described in calculating wherein, represent from equipment j to the through-put power of equipment i, represent the calculating energy consumption of executing the task.

Further, described service node selected cell according to described in calculating wherein, for the processor power of equipment, f _jfor the processor clock frequency of equipment, X is complexity factor.

Further, also comprise bill unit, wherein:

Described bill unit, if for task performed by equipment i this locality, then keep the bill length of equipment i constant; If task performed by long-range cloud device, then the bill length of equipment i is increased b _i(t _h); If task performed by other equipment in network, then the bill length of equipment i is increased b _i(t _h), the bill length of the equipment of executing the task is reduced d _j(t _h).

In sum, the present invention can eliminate computing redundancy in network and transfer of data redundancy, reduce energy consumption and the Internet data traffic demand of mobile device with this, reach the average energy consumption minimizing equipment in network, control the object of external Internet data traffic simultaneously.

Accompanying drawing explanation

Fig. 1 is the deployment schematic diagram of the execution environment of task immigration method of the present invention;

Fig. 2 is the logical schematic of the energy consumption calculation of task immigration method of the present invention;

Fig. 3 is the flow chart of task immigration method of the present invention;

Fig. 4 is the performance test figure of task immigration method of the present invention;

Fig. 5 is the Organization Chart of network controller of the present invention.

Embodiment

The task immigration method of the application is distributed the task scheduling between equipment (as mobile device), is intended to the average energy consumption minimizing equipment in network, controls external Internet data traffic simultaneously.The method of the application can be applicable in the networks such as mobile Internet.Herein by task immigration method called after E2COM, E2COM controller (network controller, hereinafter referred to as controller) be deployed on the access controller (AC) of wireless network, based on energy consumption and Internet data traffic, carry out scheduling decision to each task of mobile device, it specifically disposes operational mode as shown in Figure 1.According to decision-making, the task of mobile device can perform in this locality, move in network miscellaneous equipment on cooperation perform or move to long-range cloud device and perform.E2COM utilizes data fingerprint to mark task attribute, can by sharing the execution result of task between mobile subscriber.These tasks can be computational intesiveness tasks, also can transmit intensive task, such as, obtain positional information, optical character identification (OCR), picture processing etc.

In order to encourage the cooperation between mobile device, reduce network outbound data flow, the application utilizes the computation migration quantity of service that in bill record network, each mobile device obtains.If the bill length of a mobile device exceedes threshold value, computation migration of cannot reentrying service.On the other hand, mobile device by providing computation migration service for miscellaneous equipment, can repay the bill of oneself.The size of bill associates with the Internet data traffic of task.The object of controlling the data flows just can be reached by controlling bill length.

In order to solve task matching and the deployment issue of E2COM, the application devises a kind of online task scheduling algorithm (OTS).OTS under the prerequisite controlling WLAN (wireless local area network) outbound data flow, can minimize the average energy consumption of the whole network.Network manager hereinafter can be introduced by adjustment coefficient V(), realize the tradeoffs of the whole network energy consumption of mobile equipment and external Internet data traffic.

Below calculating energy consumption used in this application and transmission Calculation Method of Energy Consumption are described:

Suppose, in a time period T, have n mobile device in network, they have m task to need decision-making to perform, and are labeled as { C ₁, C ₂..., C _m.Each task presentation is: wherein h is mission number, h1,2 ..., m, i are the numbering of the equipment sending tasks carrying request, i1,2 ..., n; ID identifies task type; the input size of data of expression task; the output size of data of expression task.

Mobile device j executes the task calculating energy consumption be expressed as:

${\mathrm{\ϵ}}_C^j\left(C_h^i\right)={\mathrm{\ρ}}_C^j\•\frac{X\·D_h^{\mathrm{in}}}{f_j};$

Wherein, j is device numbering, and f _jbe respectively cpu power and the clock frequency of mobile device, complexity factor X is the parameter relevant to task attribute.

The WLAN disposing E2COM can be expressed as a directed graph G (M, L), and wherein M and L represents the set of node and directed edge respectively.Each node i ∈ M represents a mobile device in network, and every bar directed edge (i, j) ∈ L represents a wireless links of equipment in network i to equipment j, and with one group of weighted value be associated, wherein, represent from equipment i to the through-put power of equipment j, represent the transmission rate on current time wireless link (i, j).

The task of mobile device i move in mobile device j implementation, the energy consumption of i is designated as the energy consumption of j is designated as then have:

${\mathrm{\ϵ}}_L^{\mathrm{ij}}\left(C_h^i\right)={\mathrm{\ρ}}_T^{\mathrm{ij}}\•\frac{D_h^{\mathrm{in}}}{{\mathrm{\φ}}_T^{\mathrm{ij}}}+{\mathrm{\ρ}}_T^{\mathrm{ij}}\•\frac{D_h^{\mathrm{out}}}{{\mathrm{\φ}}_T^{\mathrm{ji}}};$

${\mathrm{\ϵ}}_R^{\mathrm{ij}}\left(C_h^i\right)={\mathrm{\ρ}}_T^{\mathrm{ji}}\•\frac{D_h^{\mathrm{in}}}{{\mathrm{\φ}}_T^{\mathrm{ij}}}+{\mathrm{\ρ}}_T^{\mathrm{ji}}\•\frac{D_h^{\mathrm{out}}}{{\mathrm{\φ}}_T^{\mathrm{ji}}}+{\mathrm{\ϵ}}_C^j\left(C_h^i\right).$

Due to task immigration perform to high in the clouds time, the energy consumption in high in the clouds is not counted in native system, thus calculation task perform energy consumption can be collectively expressed as:

${\mathrm{\ϵ}}_O^{\mathrm{ij}}\left(C_h^i\right)=\left\{\begin{array}{c}{\mathrm{\ϵ}}_L^{\mathrm{ij}}\left(C_h^i\right)+{\mathrm{\ϵ}}_R^{\mathrm{ij}}\left(C_h^i\right),j=1...,n;\\ {\mathrm{\ϵ}}_L^{\mathrm{ij}}\left(C_h^i\right),j=n+1.\end{array}\right.$

When i and j is equal, energy consumption when indication equipment this locality performs.

E2COM carries out decision-making with discrete-time manner, grain t when being defined as the time of implementation of each task _h, h1,2 ..., m, utilizes decision vector in grain when each carry out decision-making, concrete meaning is:

Wherein, j1,2 ..., n+1. expression task move to high in the clouds to perform.Execute the task energy consumption can formal unification be e2COM to the logic of various energy consumption calculation as shown in Figure 2.

In time period T, execute the task set { C ₁, C ₂..., C _mthe average energy consumption of the whole network be designated as:

$\stackrel{\‾}{\mathrm{\ϵ}}=\underset{m\→\∞}{\lim }\sup \frac{1}{m}{\mathrm{\Σ}}_{h=1}^{m}E\left\{\right|\mathrm{\ϵ}\left(C_h^i\right)\left|\right\}.$

The application utilizes the computation migration quantity of service that in bill record network, each mobile device obtains.If the bill length of a mobile device exceedes threshold value, the E2COM that cannot reentry provides computation migration service.On the other hand, mobile device by providing computation migration service for miscellaneous equipment, can repay the bill of oneself.The enjoyment computation migration service continued to enable each mobile device, must guarantee that the bill length of each mobile device can not increase without limitation.The size of bill associates with the data traffic of task.The object of controlling the data flows just can be reached by controlling bill length.

What migration execution task generation data traffic can be similar to is expressed as what perform due to this locality execution and mobile device cooperation does not produce external Internet data traffic, can normalization be expressed as so perform an external Internet data traffic of generation

Definition service pricing function is f ().If mobile device i time grain t _h, by calculation task move on high in the clouds or miscellaneous equipment and perform, then the bill of i increases if mobile device j time grain t _h, for terminal i performs calculation task then the bill of j reduces if execute the task in equipment this locality, then b _i(t _h) and d _j(t _h) be all zero.Service pricing function f () specifically can limit according to customer flow demand for control.

The task scheduling algorithm of the application is as follows:

Start time of grain th when each, in network, the current bill length of all mobile devices is defined as time grain t _hin, all mobile devices are because the bill obtaining service increase is expressed as (b ₁(t _h), b ₂(t _h) ..., b _n(t _h)), because the bill providing service to reduce for miscellaneous equipment is expressed as (d ₁(t _h), d ₂(t _h) ..., d _n(t _h)).

Defining each mobile device in the bill length of initial time is 0, i.e. Q _i(t ₀)=0, at t _hthe bill length of grain will be in lower a period of time:

Q _i(t _h＋1)＝max[Q _i(t _h)-d _j(t _h)+b _i(t _h),0]。

In order to limit the bill length of each mobile device, need the stability of the average length ensureing bill, namely

$\stackrel{\&OverBar;}{Q}=\underset{m\&RightArrow;\&infin;}{\lim }\sup \frac{1}{m}{\mathrm{\&Sigma;}}_{h=1}^m{\mathrm{\&Sigma;}}_{i=1}^{n}E\left\{\right|Q_i\left(t_h\right)\left|\right\}<\&infin;.$

Require and bill stability requirement based on energy optimization, the optimization problem of the application can formalization representation be:

$\begin{array}{cc}\min & \stackrel{\&OverBar;}{\mathrm{\&epsiv;}}\end{array}$

$\begin{array}{cc}S.T.& \stackrel{\&OverBar;}{Q}<\&infin;\end{array},$

${\mathrm{\&alpha;}}_j\left(C_h^i\right)\&Element;\left\{\mathrm{0,1}\right\},h=\mathrm{1,2},...,m$

j＝1,2,...,n+1

${\mathrm{\&Sigma;}}_{j=1}^{n+1}{\mathrm{\&alpha;}}_j\left(C_h^i\right)=1,h=\mathrm{1,2},...,m$

The optimal solution of this optimization problem is set to grain t when each _hstart time, E2COM is with minimized average energy consumption and to limit average bill length be foundation, for each task does migration scheduling decision-making.

Based on Lyapunov stability theory, this optimization problem can be solved by the upper bound minimizing drift penalty factor (drift-plus-penalty), namely minimizes following formula:

${\mathrm{DPP}}_{\mathrm{bound}}=V\&bull;\mathrm{\&epsiv;}\left(C_h^i\right)+{\mathrm{\&Sigma;}}_{j=1}^n{Q}_j\left(t_h\right)(b_i\left(t_h\right)-d_j\left(t_h\right));$

Wherein, V is the coefficient of balance realizing the whole network energy consumption of mobile equipment and external Internet data traffic tradeoffs, it reflects network manager to the attention degree of network to equipment energy consumption and data traffic.V larger energy consumption saving degree is higher, and traffic consumes is more, and vice versa.

Based on the method for solving of this optimization problem, the application devises online task immigration dispatching method (OTS), when namely when each, grain starts, according to above formula, is the service node of each task choosing the best; After tasks carrying, according to tasks carrying situation, revise the bill length of each mobile device.

Below in conjunction with accompanying drawing to a preferred embodiment of the present invention will be described in detail, should be appreciated that following illustrated preferred embodiment is only for instruction and explanation of the present invention, is not intended to limit the present invention.

The task immigration method that the application proposes is applicable to large enterprise's level network, at the access controller deploy E2COM controller of network.E2COM controller periodic collection also preserves network state information, and specifying information comprises:

(1) current time, the through-put power in network between wireless link:

When subscript is identical, the through-put power of indication equipment this locality, value is zero.

(2) current time, the transmission rate in network between mobile device:

When subscript is identical, the transmission rate of indication equipment this locality, value is infinitely great.

(3) current time, the bill length (Q of each equipment ₁(t _h), Q ₂(t _h) ..., Q _n(t _h));

When mobile subscriber in network initiates tasks carrying request, E2COM controller operation task dispatching algorithm, for the mobile subscriber in network carries out task scheduling and distribution.As shown in Figure 3, to run OTS algorithm, realize especially by following steps:

Step 301, mobile device i sends tasks carrying request to E2COM controller

Step 302, the network of relation state information input OTS that E2COM controller will be preserved, carry out scheduling decision, concrete steps are as follows:

(1) DPP is calculated for each equipment _boundvalue, and for j=1,2 ..., n+1 finds optimum DPP _boundvalue, wherein j=n+1 represents the corresponding index of cloud device;

Utilize following formula, calculate the execution energy consumption of each task.

${\mathrm{\&epsiv;}}_O^{\mathrm{ij}}\left(C_h^i\right)=\left\{\begin{array}{c}{\mathrm{\&epsiv;}}_L^{\mathrm{ij}}\left(C_h^i\right)+{\mathrm{\&epsiv;}}_R^{\mathrm{ij}}\left(C_h^i\right),j=1,...,n;\\ {\mathrm{\&epsiv;}}_L^{\mathrm{ij}}\left(C_h^i\right),j=n+1.\end{array}\right.$

$\mathrm{\&epsiv;}\left(C_h^i\right)={\mathrm{\&Sigma;}}_{j=1}^{n+1}{\mathrm{\&alpha;}}_j\left(C_h^i\right)\&CenterDot;{\mathrm{\&epsiv;}}_O^{\mathrm{ij}}\left(C_h^i\right).$

The bill length Q current according to each terminal _i(t _h), and the energy consumption calculated by following formulae discovery DPP _bound:

${\mathrm{DPP}}_{\mathrm{bound}}=V\&bull;\mathrm{\&epsiv;}\left(C_h^i\right)+{\mathrm{\&Sigma;}}_{j=1}^n{Q}_j\left(t_h\right)(b_i\left(t_h\right)-d_j\left(t_h\right)).$

(2) according to minimum DPP _boundvalue, is the service station j of task choosing optimum, and result notice request is performed equipment i;

Step 303, E2COM controller according to the result of decision, recalculates and upgrades the bill length of each mobile device, obtaining (Q ₁(t _h+1), Q ₂(t _h+1) ..., Q _n(t _h+1)), circular is:

(1) if task performed by i this locality, then the bill length of i is constant;

(2) if task performed by high in the clouds, then the bill length of i increases b _i(t _h);

(3) if task completed by j equipment, then the bill length of i increases b _i(t _h), the bill length of j reduces d _j(t _h).

Step 304, E2COM controller waits for that next task performs request and arrives, and enters new round scheduling decision.

In order to the practicality of the application is described, we are for following scene, and the deployment introducing this patent is in detail implemented.

Embodiment one

At tourist attraction, passenger utilizes smart mobile phone, takes pictures to sight name, and the AP that passage is deployed on sight spot uploads to high in the clouds, carries out OCR identification, and obtains the sight spot recommended information of the various ways such as word, Voice & Video.Visitor within the scope of same sight spot can share these recommended informations by E2COM.This has saved external data traffic from the whole network angle, has saved smart mobile phone energy consumption, has decreased transmission delay, improve Consumer's Experience from user perspective.

Embodiment two

In the stream of people such as subway station, coffee-house concentrated area, people's custom utilizes smart mobile phone to access social station, browses news, watches video online.In the WLAN (wireless local area network) in these regions, dispose E2COM, the people in the same area can share the identical information obtained from high in the clouds.This achieves the saving of the whole network angle outbound data flow equally, and the saving of user's smart mobile phone energy consumption.

In sum, the present invention proposes the cooperative computation migration mechanism (E2COM) in a kind of large-scale wireless local area network (LAN) clearly, is intended to minimize the average energy consumption of mobile device in network, controls the external Internet data traffic of WLAN (wireless local area network) simultaneously.E2COM controller portion is deployed on the access controller (AC) of wireless network, the energy requirement of task based access control and Internet data traffic demand, carries out scheduling decision to each task of mobile device.By sharing the execution result of task, the computing redundancy in network and transfer of data redundancy can be eliminated, reduce energy consumption and the Internet data traffic demand of mobile device with this between mobile subscriber.Network manager can pass through adjustment coefficient V, realizes the tradeoffs of the whole network energy consumption of mobile equipment and external Internet data traffic.The actual consumption of the application, data traffic optimization and counterbalance effect are shown in Fig. 4.

As shown in Figure 5, present invention also provides a kind of network controller, comprising: receiving element and service sensor selection problem unit, wherein:

Receiving element, for the tasks carrying request of receiving equipment;

Service node selected cell, for according to energy consumption and Internet data traffic, selects the service node of task;

Wherein, the service node of task is selected to comprise:

Execute the task in equipment this locality;

Move on other equipment in network and cooperate to execute the task; Or,

Move to long-range cloud device to execute the task.

Service node selected cell, according to energy consumption and Internet data traffic, is selected the service node of task, being comprised:

According to ${\mathrm{DPP}}_{\mathrm{bound}}=V\&bull;\mathrm{\&epsiv;}\left(C_h^i\right)+{\mathrm{\&Sigma;}}_{j=1}^n{Q}_j\left(t_h\right)(b_i\left(t_h\right)-d_j\left(t_h\right))$ Other equipment respectively in local, the network of computing equipment and long-range cloud device are executed the task corresponding DPP _bound;

Select minimum DPP _boundcorresponding equipment is as the service node of task;

Wherein, DPP _boundfor drift penalty factor, V is the coefficient of balance realizing the whole network energy consumption of mobile equipment and external Internet data traffic tradeoffs, expression task, i is the numbering of the equipment sending tasks carrying request, and h is mission number, and j is device numbering, and n is the sum of the equipment in network, for the energy consumption needed of executing the task, Q _j(t _h) for equipment time grain t _hbill length, b _i(t _h) be the length that after the equipment transportation task of transmission tasks carrying request, self bill increases, d _j(t _h) for equipment is the length that after other equipment are executed the task, self bill reduces, bill length associates with Internet data traffic.

Service node selected cell according to $\mathrm{\&epsiv;}\left(C_h^i\right)={\mathrm{\&Sigma;}}_{j=1}^{n+1}{\mathrm{\&alpha;}}_j\left(C_h^i\right)\&CenterDot;{\mathrm{\&epsiv;}}_O^{\mathrm{ij}}\left(C_h^i\right)$ Calculate

Wherein, j=1 ~ n represents the equipment in network, and j=n+1 represents long-range cloud device;

${\mathrm{\&epsiv;}}_O^{\mathrm{ij}}\left(C_h^i\right)=\left\{\begin{array}{c}{\mathrm{\&epsiv;}}_L^{\mathrm{ij}}\left(C_h^i\right)+{\mathrm{\&epsiv;}}_R^{\mathrm{ij}}\left(C_h^i\right),j=1,...,n;\\ {\mathrm{\&epsiv;}}_L^{\mathrm{ij}}\left(C_h^i\right),j=n+1;\end{array}\right.,$ for initiating the energy consumption of the equipment of tasks carrying request, the energy consumption of the equipment of executing the task.

Service node selected cell according to calculate wherein, represent from equipment i to the through-put power of equipment j, represent the transmission rate current time is from equipment i to the wireless link of equipment j, the input size of data of expression task, represent the transmission rate current time is from equipment j to the wireless link of equipment i, the output size of data of expression task; Further,

According to ${\mathrm{\&epsiv;}}_R^{\mathrm{ij}}\left(C_h^i\right)={\mathrm{\&rho;}}_T^{\mathrm{ji}}\&bull;\frac{D_h^{\mathrm{in}}}{{\mathrm{\&phi;}}_T^{\mathrm{ij}}}+{\mathrm{\&rho;}}_T^{\mathrm{ji}}\&bull;\frac{D_h^{\mathrm{out}}}{{\mathrm{\&phi;}}_T^{\mathrm{ji}}}+{\mathrm{\&epsiv;}}_C^j\left(C_h^i\right)$ Calculate wherein, represent from equipment j to the through-put power of equipment i, represent the calculating energy consumption of executing the task.

Service node selected cell according to calculate wherein, for the processor power of equipment, f _jfor the processor clock frequency of equipment, X is complexity factor.

Network controller also comprises bill unit, wherein:

Bill unit, if for task performed by equipment i this locality, then keep the bill length of equipment i constant; If task performed by long-range cloud device, then the bill length of equipment i is increased b _i(t _h); If task performed by other equipment in network, then the bill length of equipment i is increased b _i(t _h), the bill length of the equipment of executing the task is reduced d _j(t _h).

Those skilled in the art should be understood that, above-mentioned of the present invention each module or each step can realize with general calculation element, they can concentrate on single calculation element, or be distributed on network that multiple calculation element forms, alternatively, they can realize with the executable program code of calculation element, thus, they can be stored and be performed by calculation element in the storage device, and in some cases, step shown or described by can performing with the order be different from herein, or they are made into each integrated circuit modules respectively, or the multiple module in them or step are made into single integrated circuit module to realize.Like this, the present invention is not restricted to any specific hardware and software combination.

The foregoing is only the preferred embodiments of the present invention, be not limited to the present invention, for a person skilled in the art, the present invention can have various modifications and variations.Within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.

Although above to invention has been detailed description, the present invention is not limited thereto, those skilled in the art of the present technique can carry out various amendment according to principle of the present invention.Therefore, all amendments done according to the principle of the invention, all should be understood to fall into protection scope of the present invention.

Claims (12)

1. a task immigration method, comprising:

The tasks carrying request of network controller receiving equipment, according to energy consumption and Internet data traffic, selects the service node of task;

Wherein, the service node of described selection task comprises:

Execute the task in equipment this locality;

Move on other equipment in network and cooperate to execute the task; Or,

Move to long-range cloud device to execute the task.

2. the method for claim 1, is characterized in that, described according to energy consumption and Internet data traffic, selects the service node of task, comprising:

According to ${\mathrm{DPP}}_{\mathrm{bound}}=V\&bull;\mathrm{\&epsiv;}\left(C_h^i\right)+{\mathrm{\&Sigma;}}_{j=1}^n{Q}_j\left(t_h\right)(b_i\left(t_h\right)-d_j\left(t_h\right))$ Other equipment respectively in computing equipment this locality, network and long-range cloud device perform DPP corresponding to described task _bound;

Select minimum DPP _boundcorresponding equipment is as the service node of task;

Wherein, DPP _boundfor drift penalty factor, V is the coefficient of balance realizing the whole network energy consumption of mobile equipment and external Internet data traffic tradeoffs, expression task, i is the numbering of the equipment sending tasks carrying request, and h is mission number, and j is device numbering, and n is the sum of the equipment in network, for the energy consumption needed of executing the task, Q _j(t _h) for equipment time grain t _hbill length, b _i(t _h) be the length that after the equipment transportation task of transmission tasks carrying request, self bill increases, d _j(t _h) for equipment is the length that after other equipment are executed the task, self bill reduces, described bill length associates with Internet data traffic.

3. method as claimed in claim 2, is characterized in that:

According to $\mathrm{\&epsiv;}\left(C_h^i\right)={\mathrm{\&Sigma;}}_{j=1}^{n+1}{\mathrm{\&alpha;}}_j\left(C_h^i\right)\&CenterDot;{\mathrm{\&epsiv;}}_O^{\mathrm{ij}}\left(C_h^i\right)$ Described in calculating

Wherein, j=1 ~ n represents the equipment in network, and j=n+1 represents long-range cloud device;

${\mathrm{\&epsiv;}}_O^{\mathrm{ij}}\left(C_h^i\right)=\left\{\begin{array}{c}{\mathrm{\&epsiv;}}_L^{\mathrm{ij}}\left(C_h^i\right)+{\mathrm{\&epsiv;}}_R^{\mathrm{ij}}\left(C_h^i\right),j=1,...,n;\\ {\mathrm{\&epsiv;}}_L^{\mathrm{ij}}\left(C_h^i\right),j=n+1;\end{array}\right.,$ for initiating the energy consumption of the equipment of tasks carrying request, the energy consumption of the equipment of executing the task.

4. method as claimed in claim 3, is characterized in that:

According to described in calculating wherein, represent from equipment i to the through-put power of equipment j, represent the transmission rate current time is from equipment i to the wireless link of equipment j, the input size of data of expression task, represent the transmission rate current time is from equipment j to the wireless link of equipment i, the output size of data of expression task;

According to ${\mathrm{\&epsiv;}}_R^{\mathrm{ij}}\left(C_h^i\right)={\mathrm{\&rho;}}_T^{\mathrm{ji}}\&bull;\frac{D_h^{\mathrm{in}}}{{\mathrm{\&phi;}}_T^{\mathrm{ij}}}+{\mathrm{\&rho;}}_T^{\mathrm{ji}}\&bull;\frac{D_h^{\mathrm{out}}}{{\mathrm{\&phi;}}_T^{\mathrm{ji}}}+{\mathrm{\&epsiv;}}_C^j\left(C_h^i\right)$ Described in calculating wherein, represent from equipment j to the through-put power of equipment i, represent the calculating energy consumption of executing the task.

5. method as claimed in claim 4, is characterized in that:

According to described in calculating wherein, for the processor power of equipment, f _jfor the processor clock frequency of equipment, X is complexity factor.

6. the method as described in one of as any in claim 2 ~ 5, is characterized in that:

Also comprise after the service node of the task of selection:

If task performed by equipment i this locality, then keep the bill length of equipment i constant;

If task performed by long-range cloud device, then the bill length of equipment i is increased b _i(t _h);

If task performed by other equipment in network, then the bill length of equipment i is increased b _i(t _h), the bill length of the equipment of executing the task is reduced d _j(t _h).

7. a network controller, comprising: receiving element and service sensor selection problem unit, wherein:

Described receiving element, for the tasks carrying request of receiving equipment;

Described service node selected cell, for according to energy consumption and Internet data traffic, selects the service node of task;

Wherein, the service node of described selection task comprises:

Execute the task in equipment this locality;

Move on other equipment in network and cooperate to execute the task; Or,

Move to long-range cloud device to execute the task.

8. network controller as claimed in claim 7, is characterized in that:

Described service node selected cell, according to energy consumption and Internet data traffic, is selected the service node of task, being comprised:

According to ${\mathrm{DPP}}_{\mathrm{bound}}=V\&bull;\mathrm{\&epsiv;}\left(C_h^i\right)+{\mathrm{\&Sigma;}}_{j=1}^n{Q}_j\left(t_h\right)(b_i\left(t_h\right)-d_j\left(t_h\right))$ Other equipment respectively in computing equipment this locality, network and long-range cloud device perform DPP corresponding to described task _bound;

Select minimum DPP _boundcorresponding equipment is as the service node of task;

Wherein, DPP _boundfor drift penalty factor, V is the coefficient of balance realizing the whole network energy consumption of mobile equipment and external Internet data traffic tradeoffs, expression task, i is the numbering of the equipment sending tasks carrying request, and h is mission number, and j is device numbering, and n is the sum of the equipment in network, for the energy consumption needed of executing the task, Q _j(t _h) for equipment time grain t _hbill length, b _i(t _h) be the length that after the equipment transportation task of transmission tasks carrying request, self bill increases, d _j(t _h) for equipment is the length that after other equipment are executed the task, self bill reduces, described bill length associates with Internet data traffic.

9. network controller as claimed in claim 8, is characterized in that:

Described service node selected cell according to $\mathrm{\&epsiv;}\left(C_h^i\right)={\mathrm{\&Sigma;}}_{j=1}^{n+1}{\mathrm{\&alpha;}}_j\left(C_h^i\right)\&CenterDot;{\mathrm{\&epsiv;}}_O^{\mathrm{ij}}\left(C_h^i\right)$ Described in calculating

Wherein, j=1 ~ n represents the equipment in network, and j=n+1 represents long-range cloud device;

${\mathrm{\&epsiv;}}_O^{\mathrm{ij}}\left(C_h^i\right)=\left\{\begin{array}{c}{\mathrm{\&epsiv;}}_L^{\mathrm{ij}}\left(C_h^i\right)+{\mathrm{\&epsiv;}}_R^{\mathrm{ij}}\left(C_h^i\right),j=1,...,n;\\ {\mathrm{\&epsiv;}}_L^{\mathrm{ij}}\left(C_h^i\right),j=n+1;\end{array}\right.,$ for initiating the energy consumption of the equipment of tasks carrying request, the energy consumption of the equipment of executing the task.

10. network controller as claimed in claim 9, is characterized in that:

Described service node selected cell according to described in calculating wherein, represent from equipment i to the through-put power of equipment j, represent the transmission rate current time is from equipment i to the wireless link of equipment j, the input size of data of expression task, represent the transmission rate current time is from equipment j to the wireless link of equipment i, the output size of data of expression task; Further,

According to ${\mathrm{\&epsiv;}}_R^{\mathrm{ij}}\left(C_h^i\right)={\mathrm{\&rho;}}_T^{\mathrm{ji}}\&bull;\frac{D_h^{\mathrm{in}}}{{\mathrm{\&phi;}}_T^{\mathrm{ij}}}+{\mathrm{\&rho;}}_T^{\mathrm{ji}}\&bull;\frac{D_h^{\mathrm{out}}}{{\mathrm{\&phi;}}_T^{\mathrm{ji}}}+{\mathrm{\&epsiv;}}_C^j\left(C_h^i\right)$ Described in calculating wherein, represent from equipment j to the through-put power of equipment i, represent the calculating energy consumption of executing the task.

11. network controllers as claimed in claim 10, is characterized in that:

Described service node selected cell according to described in calculating wherein, for the processor power of equipment, f _jfor the processor clock frequency of equipment, X is complexity factor.

12. one of as any in claim 7 ~ 11 as described in network controller, it is characterized in that, also comprise bill unit, wherein:

Described bill unit, if for task performed by equipment i this locality, then keep the bill length of equipment i constant; If task performed by long-range cloud device, then the bill length of equipment i is increased b _i(t _h); If task performed by other equipment in network, then the bill length of equipment i is increased b _i(t _h), the bill length of the equipment of executing the task is reduced d _j(t _h).