CN102223395B - Method and device for balancing dynamic load of middleware in radio frequency identification (RFID) network - Google Patents
Method and device for balancing dynamic load of middleware in radio frequency identification (RFID) network Download PDFInfo
- Publication number
- CN102223395B CN102223395B CN201110120942.XA CN201110120942A CN102223395B CN 102223395 B CN102223395 B CN 102223395B CN 201110120942 A CN201110120942 A CN 201110120942A CN 102223395 B CN102223395 B CN 102223395B
- Authority
- CN
- China
- Prior art keywords
- middleware
- load
- reader
- rfid
- network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Abstract
The embodiment of the invention discloses a method and a device for balancing the dynamic load of middleware in a radio frequency identification (RFID) network. The method comprises the following steps of: selecting the middleware with the maximum load to move readers; and always moving one or more readers on the middleware with the maximum load to the middleware with the minimum load within a certain period during moving. Some current policies are more suitable for the condition that all middleware has same configuration (middleware isomorphism). The method and the device for balancing the dynamic load are designed based on the middleware with different configuration (the configuration which affects the processing performance of the middleware comprises a central processing unit (CPU) of a host server, internal memory, network bandwidth and other factors), and the balancing on the load of the middleware in the RFID network is effectively realized.
Description
1. technical field
The present invention relates to radio frequency identification (RFID) networking technology area, particularly a kind of a kind of method that realizes radio frequency identification network middleware dynamic load leveling.
2. background technology
The present invention relates to the dynamic load equilibrium technology of multiple middlewares in radio frequency identification (RFID) network.RFID middleware is the middleware of a message-oriented, has at present a lot of mechanisms to study its standards system both at home and abroad.Massachusetts Polytechnics (MIT) has proposed EPCGlobal standards system, and many researchers have also proposed various modification and implementations based on this standards system.Meanwhile, some companies conduct in-depth research RFID middleware product with increasing income to organize also.RFID middleware, as a kind of new software engineering, has greatly advanced the data interaction between RFID equipment and enterprise's upper layer application, for various application have brought facility.RFID middleware can be processed the data from label by distributed message mode, filters and it is distributed to corresponding application terminal by various communication protocol.Due to distributed nature, the data bulk of each middleware server processing is often widely different.Therefore load balancing is a key technical problem of RFID middleware steady operation.The RFID middleware of overload may cause network delay aggravation and performance of middle piece difference.Be badly in need of effective Load Balancing Solution.
The people such as Chae have proposed a kind of " An Approach to Adaptive Load Balancing for RFID Middlewares " in International Journal of Mathematical and Computer Sciences in 2006, select the middleware of load maximum to carry out reader migration, during migration always for the previous period on the middleware of internal burden maximum one or more readers migrate to the middleware of least-loaded.This strategy is relatively applicable to the situation (middleware isomorphism) of all middleware configuration consistencies.For middleware, configure the situation of isomorphism not (configuration that affects middleware processes performance comprises the CPU of middleware home server, internal memory, the factors such as the network bandwidth), this scheduling strategy is inapplicable.
Realizing in process of the present invention, inventor finds that in prior art, at least there are the following problems: above-mentioned simple allocation schedule method, all can not solve user specification demand and the inconsistent RFID middleware load imbalance problem causing of physical server specification configuration.
Therefore the present invention designs load-balancing algorithm and the device that a foundation is dynamically adjusted, and solves better user specification demand and the inconsistent each physical server load imbalance problem causing of physical server specification configuration.
3. summary of the invention
Embodiments of the invention provide a kind of method and device of load balancing of the RFID of realization network middleware, can realize well the load balancing of multiple RFID middlewares.
The present invention considers extensive RFID network application feature and the current existing problem of RFID middleware system, design is a kind of based on message queue, meet the distributed RFID middleware of international standard EPCGlobal ALE 1.1.1 and EPCglobal-ReaderManagementrm 1.0.1, and the improvement to implementation that standard is advised proposed, for standard has increased new function, also take into account the dynamic comprehensive load balancing of middleware system simultaneously.The intermodule communication of this system based on message queue (native system when realizing based on the message queue MSMQ of Microsoft, can certainly be other message queue subassembly product), by the multiple middleware servers of RFID middleware management service management, its distributed framework can support more massive application, take into account the integrated load equilibrium of two different levels, the stable and high-performance for system is provided safeguard.This middleware system is a kind of advanced person's multi-site read write line middleware system.
EdgeServer is the relation of one-to-many with being connected of read write line, and between read write line and EdgeServer, is man-to-man relation.Due in actual conditions, the label that each read write line is read might not be identical, even differ greatly, when there is multiple EdgeServer, the load that the EdgeServer that has connected different read write lines produces while processing read write line passback label is also uneven with regard to level, for guaranteeing the tag processes efficiency of each EdgeServer, should be to carrying out load balancing migration between these EdgeServer.
For solving RFID integrated load equalization problem, we have designed EdgeServer load-balancing algorithm, as described below:
First we define a RFID network middleware system, reader set, and EdgeServer set, the load of reader r, the load of a middleware:
As shown in figure-1, M={m
1, m
2..., m
n, M is middleware set, EdgeServer[n] use m
nrepresent; CR={r
1 k, r
2 k..., r
l k, be to be connected to EdgServer[k] reader set; A reader is only connected to an EdgeServer in a period of time.
WL
r[r]: the load of a reader represents by the number of labels of its processing;
WL
m[m
i]: a middleware m
iload by the number of labels summation of all reader processing of its management, represented.
WL
m u[m
i] and WL
m l[m
i] be made as respectively middleware m
iload bound.
The middleware integrated load balance policy of the present invention's design is as follows:
1) pass of the quantity of definition CPU and memory usage and processing label is two groups of vectors: < WL
m[m
i], CPU_m
i> and < WL
m[m
i], Mem_m
i>,
(formula-4)
3) the unbalanced degree of integrated load
4) the average unbalanced degree L_M of middleware home server equals the unbalanced degree of all middleware home server integrated loads L_M[m
i] sum is again divided by server number n,
(formula-6)
CPU_m
imiddleware m
ithe current utilance of place home server CPU, Mem_m
imiddleware m
ithe current utilance of place home server internal memory, AVG_c is the mean value of middleware place home server cpu busy percentage, AVG_m is the mean value of middleware place home server memory usage.SpeC_m
imiddleware place home server CPU specification, SpeM_m
imiddleware place home server memory standards,
CPU_m
i urepresent middleware m
ithe cpu busy percentage of label full load, Mem_m
i urepresent middleware m
ithe memory usage of label full load.
CPU_b represents the cpu busy percentage that middleware itself is intrinsic, and Mem_b represents the memory usage that middleware itself is intrinsic.
The factor that integrated load is considered can be expanded and comprise CPU, internal memory, network bandwidth utilization factor etc.
5) allocation strategy: during newly-increased reader, select the middleware of L_M minimum to distribute.
6) migration strategy: main consideration overload (middleware label load factor exceedes the set upper limit) situation, select the middleware of overload to move, need to quantize to consider how many readers of migration are to which middleware, need the number of times that reduces as far as possible migration to avoid system to occur concussion simultaneously.Always move the less reader of load on the middleware of the average unbalanced degree L_M of middleware maximum to the middleware of the average unbalanced degree L_M minimum of middleware (or inferior little), until moved middleware nonoverload for this reason.
The unbalanced degree appraisal procedure of integrated load (variance)
4. accompanying drawing explanation
Fig. 1 is RFID middleware load balancing apparatus figure;
Fig. 2 is Edge Server system assumption diagram;
Fig. 3 distributed message middleware message flow graph;
Fig. 4 provides the flow chart of dispatching distribution and migration reader method;
Fig. 5 is a schematic diagram of load balancing migration reader.
By (corresponding text all describes in detail) being described with accompanying drawing, it is easier to understand that feature of the present invention will become.
5. embodiment
The embodiment of the present invention provides a kind of device (as shown in figure-1) of the RFID of realization network middleware load balancing,
For making the advantage of technical solution of the present invention clearer, below in conjunction with drawings and Examples, the present invention is elaborated.
Embodiment mono-
The present embodiment provides a kind of method that realizes load balancing between many physical servers, and as shown in figure-4, described method comprises:
101, start, first carry out initialization operation preparation, comprise (CPU, the internal memory) configuration and their utilance and the utilance upper limit that record the service of different middlewares place.
102, judge whether it is to newly increase reader, or not do not judge whether to need migration, enter 103; That newly-increased reader need to enter 107 and distributes.
Which reader preparation and the judgement of 103, moving reader select where move and move to.
104, which reader judgement selects where move and move to: principle is to avoid the network concussion that excessively migration causes, as far as possible few migration; The reader of least-loaded on middleware of always selecting during migration to transship starts, and is moved on the middleware of unbalanced degree minimum (below definition).
105 otherwise refusal migration task.
106, from overload middleware, the reader of least-loaded starts migration, and is moved on the middleware of load imbalance degree minimum (this is consistent with assigning process).
107, distribute reader: prepare to distribute new reader to a current middleware.
108, calculate the unbalanced degree of current middleware, current middleware is carried out to comprehensive unbalanced degree sequence.
109, newly-increased reader is assigned on the minimum middleware of current comprehensive unbalanced degree.
It is concrete that reader is distributed or move on which middleware can be with reference to following process:
Table-1 initial situation
Table-2 is that the specification of four EdgeServer and the CPU of full load, memory usage are (when server resource utilance exceedes full value, the extreme low of server performance will be caused, therefore, before not exceeding full value, just should carry out load balancing migration to the read write line of overload EdgeServer).
We give tacit consent to EdgeServer main frame self system by the resource that takies 5%, and resources occupation rate is just considered to overload when reaching 70%.
The state of current each EdgeServer:
EdgeServerA has connected a read write line, and this read write line is being processed 1500 labels;
EdgeServerB has connected two read write lines, and these two read write lines are being processed 2000 labels just respectively;
EdgeServerC has connected three read write lines, and these three read write lines are being processed 5000 labels just respectively;
EdgeServerD has connected a read write line, and this read write line is being processed 6000 labels just respectively;
In the situation that tag processes rate is directly proportional to resource utilization, can calculate the resource utilization of current each EdgeServer:
The resource utilization of EdgeServer before table-2 load balance process
CPU | Internal memory | |
EdgeServer A | 0.29 | 0.32 |
EdgeServer B | 0.30 | 0.33 |
EdgeServer C | 0.65 | 0.73 |
EdgeServer D | 0.21 | 0.23 |
By table ,-2 known EdgeServerC have exceeded overload valve limit, therefore the read write line on it are carried out to load balancing migration.Adopt the integrated load equalization algorithm of the present invention's suggestion, from the EdgeServerC of overload, get next read write line, simulate being respectively connected of each EdgeServer of it and all the other, and calculate L_M:
After being connected with EdgeServerA, this server cpu busy percentage and memory usage exceed 100%, therefore can not be connected with EdgeServerA.
After being connected with EdgeServerB, the average unbalanced degree L_M of middleware home server is 0.033.
After being connected with EdgeServerD, the average unbalanced degree L_M of middleware home server is 0.006.
As can be seen here, migration read write line out should be connected with EdgeServerD.Move the load of rear four Server respectively as table-3:
Each Server load after table-3 migrations
CPU | Internal memory | |
EdgeServer A | 0.29 | 0.32 |
EdgeServer B | 0.30 | 0.33 |
EdgeServer C | 0.45 | 0.50 |
EdgeServer D | 0.34 | 0.38 |
With the unbalanced degree appraisal procedure of integrated load (variance) to distribute after each EdgeServer assess:
L_M=0.012.
If adopt the suggesting methods such as Chae only to move, take off a read write line being connected with EdgeServerC, directly migrate on the Server that label reads minimum number, the label of EdgeServerA reads minimum number, but will produce overload after migrating on it, therefore move to label and read quantity time few EdgeServerB upper (processing 2000 labels), the load of moving rear four Server is divided into as table-4:
Each Server load suggesting methods such as () Chae after table-4 migrations
CPU | Internal memory | |
EdgeServer A | 0.29 | 0.32 |
EdgeServer B | 0.60 | 0.67 |
EdgeServer C | 0.45 | 0.50 |
EdgeServer D | 0.21 | 0.23 |
With the unbalanced degree appraisal procedure of integrated load (variance) to distribute after each EdgeServer assess:
L_M_Devi=0.066. many data results all show, use while only considering that label reads the moving method of quantity, and the result of comprehensive unbalanced degree is greater than the load-balancing algorithm of the present invention's suggestion.Because of length restriction, at this, do not enumerate.
One of ordinary skill in the art will appreciate that all or part of flow process realizing in above-described embodiment method, to complete by computer program instructions and relevant hardware, described program can be stored in a computer read/write memory medium, this program, when carrying out, can comprise as the flow process of the embodiment of above-mentioned each side method.Wherein, described storage medium can be magnetic disc, CD, read-only store-memory body (Read-Only Memory, ROM) or random store-memory body (Random Access Memory, RAM) etc.
The above; be only the specific embodiment of the present invention, but protection scope of the present invention is not limited to this, any be familiar with those skilled in the art the present invention disclose technical scope in; the variation that can expect easily or replacement, all should be encompassed in protection model of the present invention with within.Therefore, protection scope of the present invention should be as the criterion with the protection range of claim.
Claims (5)
1. a method for radio frequency identification network (RFID) middleware dynamic load leveling, is characterized in that, comprising:
Detect the utilance of middleware server CPU and internal memory; Judge the unbalanced degree of middleware integrated load; Selecting the minimum middleware of the unbalanced degree of integrated load to increase reader newly distributes;
Wherein the unbalanced degree of middleware integrated load is defined as follows: by the deviation of the utilance of middleware server CPU and internal memory and the average utilization of Servers-all, determine, with L_M, represent L_M=
CPU_m
imiddleware m
ithe current utilance of place home server CPU, Mem_m
imiddleware m
ithe current utilance of place home server internal memory, AVG_c is the mean value of middleware place home server cpu busy percentage, AVG_m is the mean value of middleware place home server memory usage;
The factor that integrated load is considered comprises CPU, internal memory and network bandwidth utilization factor.
2. the method for radio frequency identification network RFID middleware dynamic load leveling according to claim 1, is characterized in that, during newly-increased reader, RFID network middleware system always selects the minimum middleware of comprehensive unbalanced degree to distribute.
3. according to the method described in claim 1, it is characterized in that, while judging comprehensive unbalanced the spending of RFID middleware, when having n RFID middleware home server, determine by the deviation of n middleware server CPU and the utilance of internal memory and the average utilization of Servers-all.
4. method according to claim 1, is characterized in that, the method for the load of tolerance middleware system: define first respectively a RFID net
Network middleware system, reader set, EdgeServer set, the load of reader r, the load of a middleware is as follows:
M={m
1, m
2..., m
n, M is middleware set, EdgeServer[n] use m
nrepresent; CR={r
1 k, r
2 k..., r
1 k, be to connect
To EdgServer[k] reader set; A reader is only connected to an EdgeServer in a period of time;
The load of a reader is expressed as WL by the number of labels of its processing
r[r];
A middleware m
iload by the number of labels summation of all reader processing of its management, represented, be WL
m[m
i];
Be made as respectively middleware m
ithe load upper limit, WL
m u[m
i] and lower limit WL
m l[m
i], make actual loading between between the two, exceed and be limited to overload, lower than under be limited to underloading.
5. the moving method of middleware: consider overload, it is the situation that middleware label load factor exceedes the set upper limit, select the middleware of overload to move, need to quantize to consider how many readers of migration are to which middleware, need the number of times that reduces as far as possible migration to avoid system to occur concussion simultaneously; The unbalanced degree of middleware integrated load according to claim 1, always move the reader of the least-loaded on the middleware of the unbalanced degree of middleware integrated load L_M maximum to the middleware of the unbalanced degree of middleware integrated load L_M minimum, until by migration middleware nonoverload.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201110120942.XA CN102223395B (en) | 2011-05-11 | 2011-05-11 | Method and device for balancing dynamic load of middleware in radio frequency identification (RFID) network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201110120942.XA CN102223395B (en) | 2011-05-11 | 2011-05-11 | Method and device for balancing dynamic load of middleware in radio frequency identification (RFID) network |
Publications (2)
Publication Number | Publication Date |
---|---|
CN102223395A CN102223395A (en) | 2011-10-19 |
CN102223395B true CN102223395B (en) | 2014-05-07 |
Family
ID=44779821
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201110120942.XA Active CN102223395B (en) | 2011-05-11 | 2011-05-11 | Method and device for balancing dynamic load of middleware in radio frequency identification (RFID) network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102223395B (en) |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103576827B (en) * | 2012-07-25 | 2016-04-27 | 田文洪 | The method of the online energy-saving distribution of a kind of cloud computation data center |
CN105530192B (en) * | 2014-09-29 | 2019-08-23 | 中兴通讯股份有限公司 | A kind of method, apparatus and primary processor of more coprocessor load balancing |
CN104750542B (en) * | 2015-04-22 | 2018-01-16 | 成都睿峰科技有限公司 | A kind of data migration method based on cloud platform |
CN110781700B (en) * | 2019-11-11 | 2021-05-14 | 湖南大学 | RFID multi-reader coordination method |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101309208A (en) * | 2008-06-21 | 2008-11-19 | 华中科技大学 | Job scheduling system suitable for grid environment and based on reliable expense |
CN101551846A (en) * | 2009-04-24 | 2009-10-07 | 华中科技大学 | Dynamic load equilibria method for radio frequency identifying middleware |
CN101639793A (en) * | 2009-08-19 | 2010-02-03 | 南京邮电大学 | Grid load predicting method based on support vector regression machine |
US7669029B1 (en) * | 2006-11-15 | 2010-02-23 | Network Appliance, Inc. | Load balancing a data storage system |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101770273B (en) * | 2008-12-29 | 2013-03-20 | 中兴通讯股份有限公司 | Method for realizing energy saving of system with a plurality of central processing units of server and device therefor |
CN101504620A (en) * | 2009-03-03 | 2009-08-12 | 华为技术有限公司 | Load balancing method, apparatus and system of virtual cluster system |
-
2011
- 2011-05-11 CN CN201110120942.XA patent/CN102223395B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7669029B1 (en) * | 2006-11-15 | 2010-02-23 | Network Appliance, Inc. | Load balancing a data storage system |
CN101309208A (en) * | 2008-06-21 | 2008-11-19 | 华中科技大学 | Job scheduling system suitable for grid environment and based on reliable expense |
CN101551846A (en) * | 2009-04-24 | 2009-10-07 | 华中科技大学 | Dynamic load equilibria method for radio frequency identifying middleware |
CN101639793A (en) * | 2009-08-19 | 2010-02-03 | 南京邮电大学 | Grid load predicting method based on support vector regression machine |
Also Published As
Publication number | Publication date |
---|---|
CN102223395A (en) | 2011-10-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107800756A (en) | A kind of load-balancing method and load equalizer | |
CN102207891B (en) | Method for achieving dynamic partitioning and load balancing of data-partitioning distributed environment | |
CN102223395B (en) | Method and device for balancing dynamic load of middleware in radio frequency identification (RFID) network | |
CN102317916A (en) | Distributed system | |
CN102970379A (en) | Method for realizing load balance among multiple servers | |
CN104834569A (en) | Cluster resource scheduling method and cluster resource scheduling system based on application types | |
CN105656973A (en) | Distributed method and system for scheduling tasks in node group | |
EP1145127A2 (en) | Apparatus for and method of non-linear constraint optimization in storage system configuration | |
CN109933431B (en) | Intelligent client load balancing method and system | |
CN101447939A (en) | Functional distribution method and load balancer | |
CN109491788B (en) | Method and device for realizing load balance of virtualization platform | |
CN102299959A (en) | Load balance realizing method of database cluster system and device | |
CN106557471A (en) | Method for scheduling task and device | |
CN111176840B (en) | Distribution optimization method and device for distributed tasks, storage medium and electronic device | |
CN104023042A (en) | Cloud platform resource scheduling method | |
US20220075665A1 (en) | Scheduling method for selecting optimal cluster within cluster of distributed collaboration type | |
CN106998340B (en) | Load balancing method and device for board resources | |
CN106095531A (en) | A kind of dispatching method of virtual machine loaded based on grade and physical machine in cloud platform | |
CN105791166B (en) | A kind of method and system of load balancing distribution | |
CN105007336A (en) | Load balancing method and load balancing system of server | |
Tao et al. | Load feedback-based resource scheduling and dynamic migration-based data locality for virtual hadoop clusters in openstack-based clouds | |
CN111147604A (en) | Load balancing method for edge calculation of Internet of vehicles | |
CN102217247B (en) | Method, apparatus and system for implementing multiple web application requests scheduling | |
CN112363827A (en) | Multi-resource index Kubernetes scheduling method based on delay factors | |
Karatza et al. | Epoch load sharing in a network of workstations |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C53 | Correction of patent for invention or patent application | ||
CB02 | Change of applicant information |
Address after: Chenghua District Chengdu city Sichuan province 610054 Taoxi Road No. 189 building 4 unit 1 Taoxi Yiyuan No. 10 Applicant after: Tian Wenhong Address before: Sichuan city of Chengdu province Chenghua District 610054 foot Water Bridge No. 2 North Street 38 No. 506 dianzaikeda Dongyuan real Applicant before: Tian Wenhong |
|
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |