CN104104611A - Method and device for achieving cluster load balancing dispatching - Google Patents

Method and device for achieving cluster load balancing dispatching Download PDF

Info

Publication number
CN104104611A
CN104104611A CN201410327814.6A CN201410327814A CN104104611A CN 104104611 A CN104104611 A CN 104104611A CN 201410327814 A CN201410327814 A CN 201410327814A CN 104104611 A CN104104611 A CN 104104611A
Authority
CN
China
Prior art keywords
server
hash
hash level
level
request
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.)
Granted
Application number
CN201410327814.6A
Other languages
Chinese (zh)
Other versions
CN104104611B (en
Inventor
辛永欣
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Inspur Beijing Electronic Information Industry Co Ltd
Original Assignee
Inspur Beijing Electronic Information Industry Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Inspur Beijing Electronic Information Industry Co Ltd filed Critical Inspur Beijing Electronic Information Industry Co Ltd
Priority to CN201410327814.6A priority Critical patent/CN104104611B/en
Publication of CN104104611A publication Critical patent/CN104104611A/en
Application granted granted Critical
Publication of CN104104611B publication Critical patent/CN104104611B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Landscapes

  • Computer And Data Communications (AREA)

Abstract

The invention discloses a method and device for achieving cluster load balancing dispatching. The method comprises the steps that when a cluster receives a request package, calculation is conducted, and a hash level is set as 0; a server corresponding to the hash level is determined according to the request package and the hash level; whether the obtained server meets a dispatching condition or not is judged, and when the server meets the dispatching condition, requests are distributed to the server; when the server does not meet the dispatching condition, the hash level is increased by 1, a server corresponding to the hash level is determined continually, and when the hash level is larger than the maximum value, the requests are rejected, wherein the value range of the hash level is from 0 to N-1, and N is the total number of the servers. According to the technical scheme, the probability of rejecting the requests is lowered, and the utilization rate of cluster resources is greatly improved.

Description

A kind of method and device of realizing cluster load balance scheduling
Technical field
The present invention relates to computer cluster field, espespecially a kind of method and device of realizing cluster load balance scheduling.
Background technology
In calculating, cluster often uses various load-balancing methods, the scalability of resolution system and transparent problem, by load balance scheduler, request is distributed to different servers efficiently and carries out, make the group system being formed by many platform independent computer become a virtual server; When client application and group system are mutual, just as the same with a high performance server interaction.
At present, common load equilibration scheduling method have wheel cry scheduling, weighted round robin scheduling, Least-Connection Scheduling, weighted least-connection scheduling, Locality-Based Least Connections Scheduling, tape copy based on the minimum link of locality, the scheduling of destination address hash, Source Hashing Scheduling etc., wherein the scheduling of destination address hash and Source Hashing Scheduling are more common static scheduling methods, in actual applications, use in firewall cluster in conjunction with these two kinds of dispatching methods, thereby ensure that whole system has unique gateway.
The general principle of Source Hashing Scheduling algorithm is: according to the source IP address of request, find out corresponding server as hashed key from the hash table of static allocation.This algorithm effective dispatch server fast, can ensure again that the request of identical address is to the identical server of scheduling simultaneously.But, the problem of this method is, if hash of source address institute to server overload, return to sky, refusal is processed follow-up request, so can make hash cannot be processed to the frequent requests of this server in the time that individual server load is larger, this has just reduced the utilance of cluster resource.
Summary of the invention
In order to solve the problems of the technologies described above, the invention provides a kind of method and device of realizing cluster load balance scheduling, can greatly reduce the probability of refusal request, greatly improve the utilance of cluster resource.
In order to reach foregoing invention object, the invention discloses a kind of method that realizes cluster load balance scheduling, comprising:
Calculate cluster and receive while asking bag, setting hash hash level is 0;
According to request bag and hash level, determine the server that hash level is corresponding;
Judge that whether the server obtaining meets schedulable condition, in the time that server meets schedulable condition, is distributed to request described server; In the time that server does not meet schedulable condition, hash level increases by 1, continue to determine server corresponding to hash level, and when hash level exceedes maximum, refusal request;
Wherein, the span of hash level is [0, N-1], total number that N is server.
Further, in the time that calculating cluster is received request bag, setting hash level is before 0, and the above-mentioned method of stating also comprises: set up and be used for storing the ServerNode table that calculates each server info of cluster and the one-to-one relationship of index value.
Further, request handbag is drawn together: the IP address of external user.
Further, determine the server that hash level is corresponding, comprising:
Obtain the index value of server by formula below:
(ip_addr*2654435761UL) % (N-hash level); Wherein, ip_addr is the integer corresponding to IP address, and % is modulo operation.
According to the index value of server, in ServerNode table, find server corresponding to hash level.
Further, schedulable condition is: server survival, weights are not 0 and the linking number of the server weights that are less than 2 times.
The invention also discloses a kind of device of realizing cluster load balance scheduling, comprising: module, mapping block, judge module and distribution module are set; Wherein,
Module is set, and in the time that external user sends request bag to calculating cluster, setting hash hash level is 0;
Mapping block, and arranges module and is connected, and for according to request bag and hash level, determines the server that hash level is corresponding;
Whether judge module, meet schedulable condition for the server that judges acquisition;
Distribution module, in the time that server meets schedulable condition, is distributed to request this server;
Described mapping block, also in the time that server does not meet schedulable condition, hash level increases by 1, continue to determine server corresponding to hash level, when hash level exceedes maximum, refusal request;
Wherein, the span of hash level is [0, N-1], total number that N is server.
Further, said apparatus also comprises memory module, and module is set is connected, for: in the time calculating cluster and receive request bag, setting hash level is before 0, sets up the ServerNode table that calculates each server info of cluster and the one-to-one relationship of index value for storing.
Further, request handbag is drawn together: the IP address of external user.
Further, mapping block specifically for:
Obtain the index value of described server by formula below:
(ip_addr*2654435761UL) % (N-hash level); Wherein, ip_addr is the integer corresponding to described IP address, and % is modulo operation.
According to the index value of described server, in ServerNode table, find server corresponding to described hash level.
Further, schedulable condition is: server survival, weights are not 0 and the linking number of the server weights that are less than 2 times.
Present techniques scheme comprises: in the time that external user sends request bag to calculating cluster, setting hash hash level is 0; According to request bag and hash level, determine the server that hash level is corresponding; Judge that whether the server obtaining meets schedulable condition, in the time that server meets schedulable condition, is distributed to request described server; In the time that server does not meet schedulable condition, hash level increases by 1, continue to determine server corresponding to hash level, and when hash level exceedes maximum, refusal request; Wherein, the span of hash level is [0, N-1], total number that N is server.The application's technical scheme has reduced the probability of refusal request, has greatly improved the utilance of cluster resource.
Brief description of the drawings
Accompanying drawing described herein is used to provide a further understanding of the present invention, forms the application's a part, and schematic description and description of the present invention is used for explaining the present invention, does not form inappropriate limitation of the present invention.In the accompanying drawings:
Fig. 1 is the flow chart that the present invention realizes the method for cluster load balance scheduling;
Fig. 2 is the flow chart that the present invention realizes the embodiment of the method for cluster load balance scheduling;
Fig. 3 is the structural representation that the present invention realizes the device of cluster load balance scheduling.
Embodiment
The present invention improves existing source address hash hash dispatching algorithm, introduces level parameter hash level (level), is used to specify hash level.When level is exactly the hash function that Source Hashing Scheduling uses for 0 time, if can not find the server that meets the demands when level=0, increase a level rank and continue to look for, so waterfall type increases level until find the server meeting the demands, the method only has just can announce can not find the server meeting the demands in the time that level reaches number of servers N, and level is while reaching N value, show to have traveled through Servers-all, Servers-all is all the server not meeting the demands.
Below in conjunction with drawings and the specific embodiments, the present invention is described in detail.
Fig. 1 is the flow chart that the present invention realizes the method for cluster load balance scheduling, as shown in Figure 1, comprising:
Step 101, calculates cluster and receives while asking bag, and setting hash hash level is 0.
Request bag in this step sends to calculating cluster from certain user of external network, and request comprises that source address IP is the IP address that external user uses.
Before this step, the inventive method also comprises: set up the ServerNode table for storing each server info of described calculating cluster and the one-to-one relationship of index value.
Wherein server info comprises: the IP of server, title, internal memory, hard disk.Belonging to common practise well known to those skilled in the art about how setting up ServerNode table, not repeating them here.
It should be noted that, ServerNode table has comprised Servers-all and information thereof, and each index value in ServerNode table is corresponding with a server in ServerNode table, finds index value, just can obtain corresponding with it server and the information of this server.
Above-mentioned request handbag is drawn together: the IP address of external user.
Step 102, according to request bag and hash level, determines the server that hash level is corresponding.
Specifically comprise:
First, obtain the index value of server corresponding to hash level by formula below:
(ip_addr*2654435761UL) % (N-hash level); Wherein, ip_addr is the integer corresponding to described IP address, and % is modulo operation.
Secondly,, according to the index value of this server, in ServerNode table, find server corresponding to this hash level.
It should be noted that, find the index value of server by above-mentioned formula, be similar to multiplication of prime numbers hash function in prior art, difference, is also core of the present invention place: first is not both, hash mask, be N-hash level, do not fix, but change according to hash level; Second is not both, and obtains server index value corresponding to hash level and be not to use with computing but with modulo operation, make computing more easy, and this has just saved operation time greatly.In addition, 2654435761UL is a value that the overall situation is fixing, be the prime number close to golden section between 2 to 2^32 (4294967296), 2654435761/4294967296=0.618033987, has ensured that different source addresses can be assigned in different servers as far as possible like this.
Can obtain by the following method corresponding ip_addr: the territory that less than is three adds 0 and supplies three.
Illustrate, in the time that IP address is 192.168.0.1, because the first two territory is three, so only need supply latter two territory, the 3rd territory supplements and becomes 000, the four territory for two 0 and supplement and become 001 for two 0, after above-mentioned IP address spaces, obtains ip_addr=192168000001.If totally 1000 station servers are N=1000, when IP address is 192.168.0.1, ip_addr=192168000001, in the time of hash level=0, N-hash level=1000 – 0=1000, (ip_addr*2654435761UL) % (N-hash level)=192168000001*2654435761%1000=761, be hash to the 761 station servers, in the time that hash level is 1, N-hash level=1000 – 1=999, (ip_addr*2654435761UL) % (N-hash level)=192168000001*2654435761%999=763, be hash to the 763 station servers.
Step 103, judges that whether the server obtaining meets schedulable condition, in the time that server meets schedulable condition, is distributed to request described server; In the time that server does not meet schedulable condition, hash level increases by 1, returns to step 102, when hash level exceedes maximum, and refusal request.
In this step, schedulable condition is: server survival, weights are not 0 and the linking number of the server weights that are less than 2 times.
If when the linking number of server is more than or equal to the weights of 2 times, show that server overloads.
In this step, it should be noted that, can use heartbeat detection technology of the prior art to judge that whether current server survives, and repeats no more herein.The weights of server and linking number are kept in scheduler, weights refer to be kept at the specified linking number of each server that scheduler can human-edited, can manually arrange, and the linking number of each server of renewal that can be real-time in the time of scheduling, linking number refers to the number of the external user being connected with server, can real time modifying.Method in employing prior art in Source Hashing Scheduling algorithm obtains and upgrades weights and the linking number of each server.
In said method, the span of hash level is [0, N-1], total number that N is server.
The inventive method only has just can announce can not find the server meeting the demands in the time that level reaches number of servers N, and level is while reaching N (total number of server) value, show to have traveled through Servers-all, be that Servers-all is all the server not meeting the demands, therefore the method is an optimum dispatching algorithm, this algorithm can be inherited the advantage of Source Hashing Scheduling algorithm simultaneously, there is scheduling locality, in the load situation in a basic balance of server, the request scheduling of identical ip addresses is arrived to same station server, locality of reference and the main memory hit rate of each station server are improved, thereby improve the disposal ability of whole group system.
Fig. 2 is the flow chart that the present invention realizes the embodiment of the method for cluster load balance scheduling, as shown in Figure 2, comprises the following steps:
Step 201, calculates cluster and receives while asking bag, and setting hash hash level is 0.
Step 202, according to request bag and hash level, determines the server that hash level is corresponding.
This step is identical with step 102, does not repeat them here.
Step 203, judges whether server meets schedulable condition, if meet schedulable condition, proceeds to step 204, if server does not meet schedulable condition, proceeds to step 205.
Schedulable condition in this step is identical with requiring in step 103.
Step 204, is distributed to server by request.So far distributed tasks completes, and finishes.
Step 205, hash level increases by 1.Proceed to step 206.
Step 206, judges whether hash level exceedes maximum, if exceed maximum, proceeds to step 207; If do not exceed maximum, return to step 202.
The span of above-mentioned hash level is [0, N-1], total number that N is server.
Step 207, refusal request.
Fig. 3 is the structural representation that the present invention realizes the device of cluster load balance scheduling, as shown in 3, comprising: module, mapping block, judge module and distribution module are set.Wherein,
Module is set, and while receiving request bag for calculating cluster, setting hash hash level is 0.
Module is set and receives the request bag sending to calculating cluster from certain user of external network, request comprises that source address IP is the IP address that external user uses, and arranges module IP address and hash level are sent to mapping block.
Mapping block, and arranges module and is connected, and for according to request bag and hash level, determines the server that hash level is corresponding.
Further, above-mentioned mapping block specifically for:
Obtain the index value of above-mentioned server by formula below:
(ip_addr*2654435761UL) % (N-hash level); Wherein, ip_addr is the integer corresponding to described IP address, and % is modulo operation.
According to the index value of described server, in ServerNode table, find server corresponding to above-mentioned hash level.
It should be noted that, find the index value of server by above-mentioned formula, be similar to multiplication of prime numbers hash function in prior art, difference, is also core of the present invention place: first is not both, hash mask, be N-hash level, do not fix, but change according to hash level; Second is not both, and obtains server index value corresponding to hash level and be not to use with computing but with modulo operation, make more gradual change of computing, and this has just saved operation time greatly.In addition, 2654435761UL is a value that the overall situation is fixing, be the prime number close to golden section between 2 to 2^32 (4294967296), 2654435761/4294967296=0.618033987, has ensured that different source addresses can be assigned in different servers as far as possible like this.
Can obtain by the following method corresponding ip_addr: the territory that less than is three adds 0 and supplies three.
Illustrate, in the time that IP address is 192.168.0.1, because the first two territory is three, so only need supply latter two territory, the 3rd territory supplements and becomes 000, the four territory for two 0 and supplement and become 001 for two 0, after above-mentioned IP address spaces, obtains ip_addr=192168000001.If totally 1000 station servers are N=1000, when IP address is 192.168.0.1, ip_addr=192168000001, in the time of hash level=0, N-hash level=1000 – 0=1000, (ip_addr*2654435761UL) % (N-hash level)=192168000001*2654435761%1000=761, be hash to the 761 station servers, in the time that hash level is 1, N-hash level=1000 – 1=999, (ip_addr*2654435761UL) % (N-hash level)=192168000001*2654435761%999=763, be hash to the 763 station servers.
Whether judge module, meet schedulable condition for the server that judges acquisition.
Above-mentioned schedulable condition comprises: the survival of this server, weights are not 0 and the linking number of the described server weights that are less than 2 times.
If when the linking number of server is more than or equal to the weights of 2 times, show that server overloads.
It should be noted that, can use heartbeat detection technology of the prior art to judge that whether current server survives, and repeats no more herein.The weights of server and linking number are kept in scheduler, weights refer to be kept at the specified linking number of each server that scheduler can human-edited, and the linking number of each server of renewal that can be real-time in the time of scheduling, linking number refers to the number of the external user being connected with server.Method in employing prior art in Source Hashing Scheduling algorithm obtains and upgrades weights and the linking number of each server.
Distribution module, in the time that server meets schedulable condition, is distributed to request this server.
Above-mentioned mapping block, also in the time that server does not meet schedulable condition, hash level increases by 1, continues to determine the server of hash layer correspondence, when hash level exceedes maximum, refusal request.
In said apparatus, the span of hash level is [0, N-1], total number that N is server.
Said apparatus also comprises: memory module, with module be set be connected, be used for: in the time that calculating cluster is received request bag, setting hash level is before 0, set up the ServerNode table that is used for storing each server info of described calculating cluster and the one-to-one relationship of index value.
Wherein server info comprises: the IP of server, title, internal memory, hard disk.Belonging to common practise well known to those skilled in the art about how setting up ServerNode table, not repeating them here.
It should be noted that, ServerNode table has comprised Servers-all and information thereof, and each index value in ServerNode table is corresponding with a server in ServerNode table, finds index value, just can obtain corresponding with it server and the information of this server.
One of ordinary skill in the art will appreciate that all or part of step in said method can carry out instruction related hardware by program and complete, described program can be stored in computer-readable recording medium, as read-only memory, disk or CD etc.Alternatively, all or part of step of above-described embodiment also can realize with one or more integrated circuits.Correspondingly, the each module/unit in above-described embodiment can adopt the form of hardware to realize, and also can adopt the form of software function module to realize.The application is not restricted to the combination of the hardware and software of any particular form.
The above, be only preferred embodiments of the present invention, is not intended to limit protection scope of the present invention.Within the spirit and principles in the present invention all, any amendment of making, be equal to replacement, improvement etc., within all should being included in protection scope of the present invention.

Claims (10)

1. a method that realizes cluster load balance scheduling, is characterized in that, comprising:
Calculate cluster and receive while asking bag, setting hash hash level is 0;
According to request bag and hash level, determine the server that hash level is corresponding;
Judge that whether the server obtaining meets schedulable condition, in the time that server meets schedulable condition, is distributed to request described server; In the time that server does not meet schedulable condition, hash level increases by 1, continue to determine server corresponding to hash level, and when hash level exceedes maximum, refusal request;
Wherein, the span of hash level is [0, N-1], total number that N is server.
2. method according to claim 1, it is characterized in that, in the time that described calculating cluster is received request bag, setting hash level is before 0, and described method also comprises: set up the ServerNode table for storing each server info of described calculating cluster and the one-to-one relationship of index value.
3. method according to claim 1 and 2, is characterized in that, described request handbag is drawn together: the IP address of external user.
4. method according to claim 3, is characterized in that, server corresponding to described definite hash level, comprising:
Obtain the index value of described server by formula below:
(ip_addr*2654435761UL) % (N-hash level); Wherein, ip_addr is the integer corresponding to described IP address, and % is modulo operation;
According to the index value of described server, in ServerNode table, find server corresponding to described hash level.
5. method according to claim 1 and 2, is characterized in that, described schedulable condition is: the survival of described server, weights are not 0 and the linking number of the described server weights that are less than 2 times.
6. a device of realizing cluster load balance scheduling, is characterized in that, comprising: module, mapping block, judge module and distribution module are set; Wherein
Module is set, and in the time that external user sends request bag to calculating cluster, setting hash hash level is 0;
Mapping block, and arranges module and is connected, and for according to request bag and hash level, determines the server that hash level is corresponding;
Whether judge module, meet schedulable condition for the server that judges acquisition;
Distribution module, in the time that server meets schedulable condition, is distributed to request described server;
Described mapping block, also in the time that server does not meet schedulable condition, hash level increases by 1, continue to determine server corresponding to hash level, when hash level exceedes maximum, refusal request;
Wherein, the span of hash level is [0, N-1], total number that N is server.
7. device according to claim 6, it is characterized in that, described device also comprises memory module, with module be set be connected, be used for: in the time that described calculating cluster is received request bag, setting hash level is before 0, sets up the ServerNode table for storing each server info of described calculating cluster and the one-to-one relationship of index value.
8. according to the device described in claim 6 or 7, it is characterized in that, described request handbag is drawn together: the IP address of external user.
9. device according to claim 8, is characterized in that, described mapping block specifically for:
Obtain the index value of described server by formula below:
(ip_addr*2654435761UL) % (N-hash level); Wherein, ip_addr is the integer corresponding to described IP address, and % is modulo operation;
According to the index value of described server, in ServerNode table, find server corresponding to described hash level.
10. according to the device described in claim 6 or 7, it is characterized in that, described schedulable condition is: the survival of described server, weights are not 0 and the linking number of the described server weights that are less than 2 times.
CN201410327814.6A 2014-07-10 2014-07-10 A kind of method and device for realizing cluster load balance scheduling Active CN104104611B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410327814.6A CN104104611B (en) 2014-07-10 2014-07-10 A kind of method and device for realizing cluster load balance scheduling

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410327814.6A CN104104611B (en) 2014-07-10 2014-07-10 A kind of method and device for realizing cluster load balance scheduling

Publications (2)

Publication Number Publication Date
CN104104611A true CN104104611A (en) 2014-10-15
CN104104611B CN104104611B (en) 2017-12-12

Family

ID=51672419

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410327814.6A Active CN104104611B (en) 2014-07-10 2014-07-10 A kind of method and device for realizing cluster load balance scheduling

Country Status (1)

Country Link
CN (1) CN104104611B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104468823A (en) * 2014-12-24 2015-03-25 浪潮(北京)电子信息产业有限公司 Cluster load balancing scheduling method and system
CN112015552A (en) * 2020-08-27 2020-12-01 平安科技(深圳)有限公司 Hash ring load balancing method and device, electronic equipment and storage medium
CN112947333A (en) * 2021-02-05 2021-06-11 天津市普迅电力信息技术有限公司 Socket long connection-based balanced load fragmentation method

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1595906A (en) * 2003-12-17 2005-03-16 浪潮电子信息产业股份有限公司 Single address traffic distributor of cluster network
US20060218176A1 (en) * 2005-03-24 2006-09-28 International Business Machines Corporation System, method, and service for organizing data for fast retrieval
CN102325093A (en) * 2011-11-01 2012-01-18 西安电子科技大学 Routing system constructing method in structuralized P2P (peer-to-peer) network
CN102571991A (en) * 2012-03-27 2012-07-11 电子科技大学 Multistage-mapping-based large-scale multi-copy distributed storage system and application method thereof
CN103139093A (en) * 2013-02-22 2013-06-05 桂林电子科技大学 High speed network data flow load balancing scheduling method based on field programmable gate array (FPGA)

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1595906A (en) * 2003-12-17 2005-03-16 浪潮电子信息产业股份有限公司 Single address traffic distributor of cluster network
US20060218176A1 (en) * 2005-03-24 2006-09-28 International Business Machines Corporation System, method, and service for organizing data for fast retrieval
CN102325093A (en) * 2011-11-01 2012-01-18 西安电子科技大学 Routing system constructing method in structuralized P2P (peer-to-peer) network
CN102571991A (en) * 2012-03-27 2012-07-11 电子科技大学 Multistage-mapping-based large-scale multi-copy distributed storage system and application method thereof
CN103139093A (en) * 2013-02-22 2013-06-05 桂林电子科技大学 High speed network data flow load balancing scheduling method based on field programmable gate array (FPGA)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104468823A (en) * 2014-12-24 2015-03-25 浪潮(北京)电子信息产业有限公司 Cluster load balancing scheduling method and system
CN104468823B (en) * 2014-12-24 2018-02-06 浪潮(北京)电子信息产业有限公司 A kind of cluster load balance dispatching method and system
CN112015552A (en) * 2020-08-27 2020-12-01 平安科技(深圳)有限公司 Hash ring load balancing method and device, electronic equipment and storage medium
CN112947333A (en) * 2021-02-05 2021-06-11 天津市普迅电力信息技术有限公司 Socket long connection-based balanced load fragmentation method

Also Published As

Publication number Publication date
CN104104611B (en) 2017-12-12

Similar Documents

Publication Publication Date Title
US10079879B2 (en) Weighted rendezvous hashing
Krishnamurthy et al. Pratyaastha: an efficient elastic distributed sdn control plane
CN103051564B (en) The method and apparatus of dynamic resource allocation
CN107317887B (en) A kind of load-balancing method, device and system
CN102035737A (en) Adaptive load balancing method and device based on cognitive network
WO2015096681A1 (en) Method and device for processing service access
CN104850450A (en) Load balancing method and system facing mixed cloud application
CN103414752B (en) A kind of cloud data center virtual machine distribution method of network aware
KR102022441B1 (en) Hardware Acceleration Methods and Related Devices
CN104796496A (en) Cloud storage based load balancing calculating method
Wang et al. Presto: Towards efficient online virtual network embedding in virtualized cloud data centers
CN111131486B (en) Load adjustment method and device of execution node, server and storage medium
CN105516325A (en) Cloud load balancing method for carrying out elastic expansion and traffic distribution expansion according to application load
US20140115168A1 (en) Allocation design method and apparatus
CN102932271A (en) Method and device for realizing load balancing
CN106230992B (en) A kind of load-balancing method and load balancing node
US20160269297A1 (en) Scaling the LTE Control Plane for Future Mobile Access
Kang et al. Application of adaptive load balancing algorithm based on minimum traffic in cloud computing architecture
Islam et al. Task scheduling for big data management in fog infrastructure
Sahoo et al. LVRM: On the design of efficient link based virtual resource management algorithm for cloud platforms
CN115134371A (en) Scheduling method, system, equipment and medium containing edge network computing resources
Komarasamy et al. A novel approach for Dynamic Load Balancing with effective Bin Packing and VM Reconfiguration in cloud
CN104104611A (en) Method and device for achieving cluster load balancing dispatching
Sasidhar et al. Load Balancing Techniques for Efficient Traffic Management in Cloud Environment.
Kapgate Improved round robin algorithm for data center selection in cloud computing

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant