CN101753405A - Cluster server memory management method and system - Google Patents
Cluster server memory management method and system Download PDFInfo
- Publication number
- CN101753405A CN101753405A CN200810227884A CN200810227884A CN101753405A CN 101753405 A CN101753405 A CN 101753405A CN 200810227884 A CN200810227884 A CN 200810227884A CN 200810227884 A CN200810227884 A CN 200810227884A CN 101753405 A CN101753405 A CN 101753405A
- Authority
- CN
- China
- Prior art keywords
- cluster
- server
- data
- memory
- sub
- 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.)
- Pending
Links
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a cluster server memory management method and system, the method comprising the following steps: each server of the cluster server is provided with a cluster manager, the cluster managers form the cluster management system of the cluster server through mutual communication; the cluster management system uniformly manages the memory of each server, therefore, the memory of the server is configured as the cluster memory of the cluster server. The invention discloses a cluster server memory management system, comprising the cluster managers which are respectively installed in each server of the cluster server, the cluster managers form the cluster management system through the mutual communication, and the cluster management system uniformly manages the memory of each server, therefore, the memory of the server is configured as the cluster memory of the cluster server.
Description
Technical field
The present invention relates to a kind of cluster server technology, particularly a kind of cluster server memory management method and system thereof.
Background technology
It is the calculating period at center that the computer nowadays technology has entered with the network, and a large amount of services and application all are to carry out round Web.This promotes Internet user's severe increase and Internet flow explosion type ground to increase.A lot of network services can't bear the heavy load because access times explosion type ground increases, and the server request of process user in time causes the user to wait for for a long time, greatly reduces service quality, and increasing bottleneck has appeared at server end.For the bottleneck of server end, people can be resolved by software or hardware mode.
Architecture is as shown in Figure 1 adopted in existing network service, is made up of server end and client.Client mainly by PC, mobile phone etc. can access network equipment form.And server end mainly is made up of the higher server of some performances.
The network service adopted request the working method of response.The user sends the request of service by user end to server; Server end is accepted user's request, and result of calculation is returned to client, finish once ask the response.So repeatedly, the user just can obtain the service that all need.Because the user of concurrent access server end asks meeting a lot of simultaneously, the separate unit server can not in time be handled all service requests, so server end generally can provide service by multiple servers.When a large amount of service requests arrive simultaneously, adopt the mode of load balancing, service request can be assigned to each server relatively uniformly.Like this, make server end to make response timely, guaranteed the stability of network service.
As shown in Figure 1, because server end has adopted multiple servers that identical service is provided, certainly existing a large amount of identical datas in the internal memory on each server, therefore, the internal memory of server end is not well utilized.
Because internal memory is the key factor that influences systematic function, low memory directly affects the speed of server response.For bigger commerce was used, the consumption of internal memory was great.Under this architecture, can only solve the demand of system by simple increase server memory to internal memory, this has increased the cost that service is provided undoubtedly.Simultaneously, from the angle of existing server software, the support of memory size is existed the upper limit, even internal memory can be increased to enough greatly sometimes, but can not get the support of software, problem still can not get solution.
Summary of the invention
The purpose of this invention is to provide a kind of cluster server memory management method that can make cluster server share the cluster internal memory.
Another object of the present invention provides a kind of cluster server memory management system that can make cluster server share the cluster internal memory.
According to an aspect of the present invention, the cluster server memory management method with a plurality of servers that is provided may further comprise the steps:
Load cluster manager dual system respectively in A, the every station server in cluster server, these cluster manager dual systems are by the cluster management system of the cluster server of intercommunication formation mutually;
The internal memory of B, the described every station server of described cluster management system unified management, thus make the internal memory of all described servers be configured to the cluster internal memory of cluster server.
Wherein, described cluster management system is configured to the memory space of described cluster internal memory the identical sub-memory space of a plurality of capacity and this a little memory space is distributed to each described server equably, and wherein each sub-memory space is assigned to different numberings.
Wherein, cluster management system is that every station server distributes data volume relatively uniformly, and add the identification information that is associated with the numbering of described sub-memory space for each data of distributing to every station server, set up the corresponding relation list item of described identification information and data simultaneously.
The internal memory of the described every station server of described unified management comprises that data are preserved operation, data search is operated and the data deletion action.
Preserve in the operation in data, cluster management system is calculated the sub-memory space of the data that will preserve according to identification information, data are saved in this sub-memory space then, and back up this data, and it is saved in the sub-memory space of another server.
In data search operation, cluster management system is calculated the numbering of the sub-memory space of preserving these data according to the corresponding identification information of searching with client of data, returns the data in the sub-memory space that has been kept at this numbering thus.
In the data deletion action, cluster management system foundation and the corresponding identification information of the data that will delete, calculate the numbering of the sub-memory space of preserving these data, delete the data of preserving in this sub-memory space thus, and further delete the Backup Data on another server.
For the optimization system performance, if be kept at the overtime not usefulness of data in the described sub-memory space, then cluster management system is considered as rubbish with it and reclaims, so that discharge sub-memory space.
Each cluster manager dual system of the present invention comprises the memory management assembly of a charge server memory management and the server member management assembly of a charge server member management at least.
Wherein, when described server member management assembly is received joining request of cluster server newcomer broadcasting, described server member management assembly distributes a new identifier (ID) for this newcomer, and described then memory management assembly distributes the sub-memory space of respective numbers to the newcomer with this identifier.
Wherein, when described server member management assembly received that withdrawing from of cluster server member broadcasting asked, the memory management assembly was regained the sub-memory space of having distributed to this member that will withdraw from, and it is transferred in other server.
In addition, each cluster manager dual system also comprises: be used for data packet dispatching assembly and packet distribution assembly that described memory management assembly and described server member management assembly message send, and the packet listening components and the packet receiving unit that are used for memory management assembly and described server member management assembly message sink.
Wherein, described data packet dispatching assembly is responsible for the standardization and the distribution of various types of messages and is unpacked; Described packet distribution assembly is responsible for all packets are sent to the designated destination; The packet listening components is responsible for monitoring the various types of messages packet; Described packet receiving unit is responsible for and will be received complete data packet queue, and is reduced into message, and the corresponding assembly that is distributed in described memory management assembly and the described server member management assembly is handled.
According to a further aspect in the invention, the cluster server memory management system that is provided comprises the cluster manager dual system in the every station server that is loaded into cluster server respectively, these cluster manager dual systems are by the cluster management system of intercommunication formation mutually, the internal memory of the described every station server of described cluster management system unified management, thus make the internal memory of all described servers be configured to the cluster internal memory of cluster server.
Wherein, described cluster management system is configured to the identical sub-memory space of a plurality of capacity to the memory space of the described cluster internal memory of described cluster server, and this a little memory space is distributed to each described server equably.
The internal storage data Managed Solution that provides efficient, transparent, be easy to expand for J2EE or other cluster service is devoted in cluster memory management of the present invention, makes memory shared and data management between service node just as simple at this machine.
For the cluster memory management technology, all buffer memory services all depend on the cluster management service, and the cluster management service is transparent to external call, and it guarantees that each node is to timely, the complete cognition of cluster topological structure in the cluster.On this basis, each buffer memory service could be managed data in buffer according to the data management policies of oneself.Each node is finished adding, leaves cluster, understands actions such as topologies change by the communication between the cluster management service, the buffer memory service is understood cluster by the cluster management service and is organized situation, receive various notices, thus finish data distribution, operation such as back up, obtain.The function that external call person uses the buffer memory service to provide is not known fully for data actual distribution situation and node distribution situation, just looks like that the operation local data is the same.
The present invention is described in detail below in conjunction with accompanying drawing.
Description of drawings
Fig. 1 is the allocation plan of the cluster server internal memory of existing enforcement network services;
Fig. 2 is the memory configurations figure of cluster server of the present invention;
Fig. 3 is an embodiment who realizes cluster server memory management of the present invention;
Fig. 4 is the schematic diagram of a cluster manager dual system in the cluster management system of the present invention shown in Figure 3;
Fig. 5 is that cluster management system of the present invention is used in the sub-memory space of preserving data and the schematic diagram of preserving data to the cluster server branch;
Fig. 6 shows the schematic diagram that concerns between sub-memory space of the present invention, data, the sign;
Fig. 7 is the schematic diagram of an instantiation of cluster manager dual system of the present invention.
Embodiment
Fig. 2 has shown the configuration of cluster server internal memory of the present invention, as shown in Figure 2, key of the present invention is that a cluster internal memory 2 of being shared by Servers-all in the cluster server 11,12,13 is set in cluster server, and this cluster internal memory is by cluster management system shown in Figure 33 management.
Fig. 3 has shown an embodiment of realization cluster of the present invention memory management, as can see from Figure 3, load cluster manager dual system 31,32,33 respectively in each server 11,12,13 of each cluster server, these cluster manager dual systems are by the cluster management system 3 of the present invention of intercommunication formation mutually.Cluster management system 3 can pass through the internal memory of every station server in cluster management interface 41,42, the 43 unified management cluster servers, thereby the internal memory 21,22,23 of the Servers-all of cluster server is configured to by the shared cluster internal memory 2 of cluster server.Therefore, cluster server memory management method of the present invention can may further comprise the steps:
A, load cluster manager dual system 31,32,33 respectively in the every station server 11,12,13 that becomes the cluster server member, these cluster manager dual systems constitute the cluster management system 3 of cluster server by intercommunication mutually;
B, by the internal memory 21,22,23 of the every station server in the described cluster management system unified management cluster server, thereby the internal memory 21,22,23 of the Servers-all 11,12,13 of cluster server is configured to the cluster internal memory 2 shared by cluster server.
Wherein, described cluster management system is configured to the memory space of the cluster internal memory 2 of described cluster server the identical sub-memory space of a plurality of capacity and this a little memory space is distributed to each described server equably, and wherein each sub-memory space is assigned to different numberings.
Fig. 5 has shown that cluster management system of the present invention is used in the sub-memory space of preserving data and the mode of preserving data to the cluster server branch.In Fig. 5, described sub-memory space is embodied in the bucket in the similar Hash table.
Wherein, cluster management system is that every station server distributes data volume relatively uniformly, and, and set up the corresponding relation list item of described identification information and data for each data of distributing to every station server are added the identification information that is associated with the numbering of described sub-memory space.
In Fig. 5, each server 11,12,13 internal memory is got data volume relatively uniformly.Suppose to have 4 data will distribute to this three station server 11,12,13, have so a station server for example server 11 minutes two data in full according to 1, data 2, other server 12,13 is respectively got data in full according to 1 and data 2.This shows that the data volume that each server is got differs one at most.
Cluster management system 2 is configured to a plurality of sub-memory spaces or bucket with the memory space of the cluster internal memory in the cluster server.Like this, can there be the bucket of some in the cluster internal memory, and is distributed to relatively uniformly in each server, and wherein each bucket is assigned with a numbering; Then, each data is saved in the bucket.Cluster management system adds an identification information for each data, and stores this identification information, sets up the corresponding relation list item of this identification information and data simultaneously.As an example, this identification information can be the key value of unique (specific), with these these data of key value sign.This key value should be associated with the numbering of bucket in addition, for example is associated by mapping table or hash table, and as special case, the numbering of bucket can be identical with the key value.Fig. 6 has shown the one-to-one relationship between key value, data and the bucket.
The present invention utilizes cache policy can guarantee that each server memory gets relatively data volume uniformly.For example, adopt more common Hash hashing algorithm, can data be saved in each bucket relatively uniformly according to key, then according to the algorithm that designs with bucket average be saved in every station server, realize load balancing with this.
The internal memory of the described every station server of described unified management comprises that data are preserved operation, data search is operated and the data deletion action.
Preserve in the operation in data, cluster management system is calculated the sub-memory space of the data that will preserve according to identification information, data are saved in this sub-memory space then, and back up this data, and it is saved in the sub-memory space of another server.Specifically, identification information such as pairing sub-memory space of key value or bucket that cluster management system 2 is calculated the data that will preserve according to for example hashing algorithm (hash table) are saved in data then in this sub-memory space or the bucket, and back up this data.
In data search operation, cluster management system is calculated the numbering of the sub-memory space of preserving these data according to the corresponding identification information of searching with client of data, returns the data in the sub-memory space that has been kept at this numbering thus.Specifically, cluster management system 2 is calculated the identification information of the data that client will search such as the numbering of pairing sub-memory space of key value or bucket according to for example hashing algorithm (hash table), returns the sub-memory space that has been kept at this numbering or the data in the bucket thus.
In the data deletion action, cluster management system foundation and the corresponding identification information of the data that will delete are calculated the numbering of the sub-memory space of preserving these data, delete data and the Backup Data preserved in this sub-memory space thus.
In the data deletion action, cluster management system foundation and the corresponding identification information of the data that will delete find or calculate the sub-memory space of preserving these data, delete the data and the backed up data of wherein being preserved then.
For the optimization system performance, if be kept at the overtime not usefulness of data in the described sub-memory space, then cluster management system is considered as rubbish with it and reclaims, so that discharge sub-memory space.
Each cluster manager dual system of the present invention comprises the memory management assembly of a charge server memory management at least, as Cache Service assembly; With the server member management assembly of a charge server member management, as Cluster Service assembly.
Wherein, when described server member management assembly is received joining request of cluster server newcomer broadcasting, described server member management assembly distributes a new identifier (ID) for this newcomer, and described then memory management assembly distributes the sub-memory space of respective numbers to the newcomer with this identifier.
Wherein, when described server member management assembly received that withdrawing from of cluster server member broadcasting asked, the memory management assembly was regained the sub-memory space of having distributed to this member that will withdraw from, and it is transferred in other server.
As shown in Figure 7, each cluster manager dual system also should comprise: be used for data packet dispatching assembly that described memory management assembly and described server member management assembly message sends (promptly, Packet Dispatcher assembly) and packet distribution assembly (promptly, Packet Publisher assembly), and the packet listening components that is used for memory management assembly and described server member management assembly message sink (promptly, the PacketListener assembly) and packet receiving unit (that is Packet Receiver assembly).
Wherein, described data packet dispatching assembly is responsible for the standardization and the distribution of various types of messages and is unpacked; Described packet distribution assembly is responsible for the designated destination that sends to all packets; The packet listening components is responsible for monitoring the various types of messages packet; Described packet receiving unit is responsible for and will be received complete data packet queue, and is reduced into message, and the corresponding assembly that is distributed in described memory management assembly and the described server member management assembly is handled.
For example, for the data operation request of outside, memory management assembly (being Cache Service assembly) can be found the position of this deposit data according to cache algorithm.Then, to the message of destination server transmission data manipulation, target machine can be known concrete content of operation and operating result is returned to source server according to the type of message and entrained content.
As shown in Figure 7, the transmission by message of CacheService assembly or Cluster Service assembly communicates with receiving, the transmission of message and the operation that receives roughly through three layers.At first, CacheService assembly or Cluster Service from the superiors, produce message, give down the PacketDispacter assembly of one deck, the PacketDispatcher assembly be responsible for message by appointment rule split into network packet, by the PacketPublisher assembly all bags are sent; Then, the PacketReceiver assembly of destination server can receive network packet, put it to the PacketReceiver assembly on upper strata, by the PacketReceiver assembly bag is reduced to message, at last, message is given the corresponding assembly in upper strata (one of CacheService assembly or Cluster Service assembly) and come processing messages.
The present invention also is provided with exception handling, and abnormality processing mainly refers to how to recover the data of preservation from backup machine when the server in the system takes place to delay machine unusually.
For the state of each server operation in the timely monitoring system,, a special PointScanner assembly is arranged in the system so that handle take place unusual.This assembly can be found the server of the machine of delaying timely, triggers the incident of a restore data then, and the mode by message sends to each server in the system, finishes the work of recovery by the Backup Data of preserving on other servers.
Specifically, server state is monitored by Pointer Scanner assembly, all be deployed with this assembly on every station server, do the ping operation mutually according to designated time intervals between the server, when reaching default threshold values and can't be responded by the server of ping, whole cluster sends broadcasting, and the machine data in server of delaying is recovered by backup server, and all buckets are balanced again according to assignment algorithm.
Cluster server memory management of the present invention system comprises the cluster manager dual system 31,32,33 that is loaded into respectively in the every station server that belongs to the cluster server member, these cluster manager dual systems are by the cluster management system 3 of intercommunication formation mutually, the internal memory 21,22,23 of the described every station server of cluster management system 3 unified managements makes the internal memory of all described servers be configured to the cluster internal memory 2 of cluster server.
Wherein, described cluster management system 3 is configured to identical sub-memory space of a plurality of capacity or bucket to the memory space of the described cluster internal memory of described cluster server, and this a little memory space is distributed to each described server equably.
In addition, the present invention can also provide following function:
Extended capability
If add the more service device in cluster, system can obtain bigger buffer memory capacity and handling property faster.For example distributed caching inquiry, the inquiry amount rises, but query time reduces.
There is not single failed point
System adopts the peer-to-peer model of equity fully, and guaranteeing does not have single failed point, and Servers-all can both be born whole roles of management buffer memory.
Failure is shifted and failure recovery fast
One of main design object of system realizes almost instantaneous failure transfer (load of shunting failed server) and failure recovery (reloading buffer memory after allowing to recover).
Efficient read and write access
Used that special applications is being used continuous updating storage data in order to optimize the agreement of read or write speed design, for example can obtain bigger performance boost during user session information.Combine data fault-tolerant and write buffer memory, the database throughput rises, and time-delay significantly reduces
Data consistent
Cluster wide in full accord. all members hold the same view for data cached guarantor.
Transactional synchronization and locking
System supports affairs, synchronous and locking, and these characteristics help to guarantee powerful data cached managerial ability.
The present invention also has following technique effect simultaneously:
1, smoothness: cluster member adds, withdraws from seamlessly transitting of (normally withdrawing from or unusually withdraw from), and is transparent to upper layer application, and upper-layer service for the change of cluster member completely without discovering.
2, high efficiency: the fast processing cluster member change, that load is shifted, avoid cleverly because the fully reciprocity efficiency of decision-making of bringing of node is low, simultaneously the computation capability of maximum each peer node of performance (democracy and the problem of concentrating). Good Concurrency Access ability for source synchronous access mutual exclusion and efficient, is got best meeting point (for synchronous meticulous control).
3, balance: the dynamic equilibrium of system load, no matter node adds or withdraws from, guarantee that group system obvious weakness can not occur, rely on superior algorithm, guarantee load variations that the node change brings the fastest reach poised state, make the cluster resource optimum use, like wooden barrel, each piece wood chip is all equally high.
Although above the present invention is had been described in detail, the invention is not restricted to this, those skilled in the art of the present technique can carry out various modifications according to principle of the present invention. Therefore, all modifications of doing according to the principle of the invention all should be understood to fall into protection scope of the present invention.
Claims (10)
1. cluster server memory management method, described cluster server comprises a plurality of servers, this method may further comprise the steps:
A) load cluster manager dual system respectively in every station server of cluster server, these cluster manager dual systems are by the cluster management system of the cluster server of intercommunication formation mutually;
B) internal memory of the described every station server of described cluster management system unified management, thus make the internal memory of all described servers be configured to the cluster internal memory of cluster server.
2. method according to claim 1, wherein said cluster management system is configured to the memory space of described cluster internal memory the identical sub-memory space of a plurality of capacity and this a little memory space is distributed to each described server equably, and wherein each sub-memory space is assigned to different numberings.
3. method according to claim 1, wherein cluster management system is that every station server distributes data volume relatively uniformly, and, and set up the corresponding relation list item of described identification information and data for each data of distributing to every station server are added the identification information that is associated with the numbering of described sub-memory space.
4. method according to claim 1, the internal memory of the described every station server of wherein said unified management comprise that data are preserved operation, data search is operated and the data deletion action.
5. method according to claim 4 is wherein preserved in the operation in data, and cluster management system is calculated the sub-memory space of the data that will preserve according to identification information, data are saved in this sub-memory space then.
6. method according to claim 4, wherein in the data search operation, cluster management system is calculated the numbering of the sub-memory space of preserving these data according to the corresponding identification information of searching with client of data, returns the data in the sub-memory space that has been kept at this numbering thus.
7. method according to claim 4, wherein in the data deletion action, cluster management system foundation and the corresponding identification information of the data that will delete are calculated the numbering of the sub-memory space of preserving these data, delete the data of preserving in this sub-memory space thus.
8. method according to claim 1, wherein each described cluster manager dual system comprises the memory management assembly of a charge server memory management at least; Server member management assembly with a charge server member management.
9. method according to claim 8, wherein:
When described server member management assembly is received joining request of cluster server newcomer broadcasting, described server member management assembly distributes a new identifier for this newcomer, and described then memory management assembly distributes the sub-memory space of respective numbers to the newcomer with this identifier;
When described server member management assembly received that withdrawing from of cluster server member broadcasting asked, the memory management assembly was regained the sub-memory space of having distributed to this member that will withdraw from, and it is transferred in other server.
10. cluster server memory management system, comprise the cluster manager dual system in the every station server that is loaded into respectively in the cluster server, these cluster manager dual systems are by the cluster management system of intercommunication formation mutually, the internal memory of the described every station server of described cluster management system unified management, thus make the internal memory of all described servers be configured to the cluster internal memory of cluster server.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200810227884A CN101753405A (en) | 2008-12-02 | 2008-12-02 | Cluster server memory management method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200810227884A CN101753405A (en) | 2008-12-02 | 2008-12-02 | Cluster server memory management method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101753405A true CN101753405A (en) | 2010-06-23 |
Family
ID=42479839
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200810227884A Pending CN101753405A (en) | 2008-12-02 | 2008-12-02 | Cluster server memory management method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101753405A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101989999A (en) * | 2010-11-12 | 2011-03-23 | 华中科技大学 | Hierarchical storage system in distributed environment |
| CN102523234A (en) * | 2011-12-29 | 2012-06-27 | 山东中创软件工程股份有限公司 | Implementation method and system for clustering of application servers |
| CN103226520A (en) * | 2013-04-02 | 2013-07-31 | 中国科学院信息工程研究所 | Self-adaptive cluster memory management method and server clustering system |
| CN103546522A (en) * | 2012-07-17 | 2014-01-29 | 联想(北京)有限公司 | Storage server determining method and distributed storage system |
| CN105094985A (en) * | 2015-07-15 | 2015-11-25 | 上海新储集成电路有限公司 | Low-power-consumption data center for sharing memory pool and working method thereof |
| CN105681366A (en) * | 2014-09-26 | 2016-06-15 | 广西盛源行电子信息有限公司 | Algorithm for accessing thousands of Beidou terminals to same server |
| CN106506693A (en) * | 2016-12-09 | 2017-03-15 | 北京奇虎科技有限公司 | The method and apparatus for deleting server from the server cluster that cloud service is provided |
| CN106649565A (en) * | 2016-11-14 | 2017-05-10 | 晶赞广告（上海）有限公司 | Online analysis processing method |
| CN110286854A (en) * | 2019-06-12 | 2019-09-27 | 北京达佳互联信息技术有限公司 | The method, apparatus of group members management and group's Message Processing, equipment and storage medium |
| CN110427280A (en) * | 2019-06-19 | 2019-11-08 | 中国平安财产保险股份有限公司 | A kind of buffer memory management method, system, computer equipment and storage medium |
| CN110519363A (en) * | 2019-08-23 | 2019-11-29 | 苏州浪潮智能科技有限公司 | A kind of memory device set group and its management method |
| CN117811987A (en) * | 2023-12-28 | 2024-04-02 | 杭州森格算力技术有限公司 | Method for automatic addressing and management of distributed memory resources |
-
2008
- 2008-12-02 CN CN200810227884A patent/CN101753405A/en active Pending
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101989999B (en) * | 2010-11-12 | 2012-09-26 | 华中科技大学 | Hierarchical storage system in distributed environment |
| CN101989999A (en) * | 2010-11-12 | 2011-03-23 | 华中科技大学 | Hierarchical storage system in distributed environment |
| CN102523234B (en) * | 2011-12-29 | 2015-12-02 | 山东中创软件工程股份有限公司 | A kind of application server cluster implementation method and system |
| CN102523234A (en) * | 2011-12-29 | 2012-06-27 | 山东中创软件工程股份有限公司 | Implementation method and system for clustering of application servers |
| CN103546522A (en) * | 2012-07-17 | 2014-01-29 | 联想(北京)有限公司 | Storage server determining method and distributed storage system |
| CN103226520B (en) * | 2013-04-02 | 2016-09-07 | 中国科学院信息工程研究所 | Self-adaptive cluster memory management method, server cluster system |
| CN103226520A (en) * | 2013-04-02 | 2013-07-31 | 中国科学院信息工程研究所 | Self-adaptive cluster memory management method and server clustering system |
| CN105681366A (en) * | 2014-09-26 | 2016-06-15 | 广西盛源行电子信息有限公司 | Algorithm for accessing thousands of Beidou terminals to same server |
| CN105094985A (en) * | 2015-07-15 | 2015-11-25 | 上海新储集成电路有限公司 | Low-power-consumption data center for sharing memory pool and working method thereof |
| CN106649565B (en) * | 2016-11-14 | 2019-11-26 | 晶赞广告（上海）有限公司 | A kind of on-line analytical processing method |
| CN106649565A (en) * | 2016-11-14 | 2017-05-10 | 晶赞广告（上海）有限公司 | Online analysis processing method |
| CN106506693A (en) * | 2016-12-09 | 2017-03-15 | 北京奇虎科技有限公司 | The method and apparatus for deleting server from the server cluster that cloud service is provided |
| CN106506693B (en) * | 2016-12-09 | 2019-06-18 | 北京奇虎科技有限公司 | The method and apparatus of server are deleted from the server cluster that cloud service is provided |
| CN110286854A (en) * | 2019-06-12 | 2019-09-27 | 北京达佳互联信息技术有限公司 | The method, apparatus of group members management and group's Message Processing, equipment and storage medium |
| CN110427280A (en) * | 2019-06-19 | 2019-11-08 | 中国平安财产保险股份有限公司 | A kind of buffer memory management method, system, computer equipment and storage medium |
| CN110427280B (en) * | 2019-06-19 | 2024-05-28 | 中国平安财产保险股份有限公司 | Cache management method, system, computer equipment and storage medium |
| CN110519363A (en) * | 2019-08-23 | 2019-11-29 | 苏州浪潮智能科技有限公司 | A kind of memory device set group and its management method |
| CN110519363B (en) * | 2019-08-23 | 2022-03-22 | 苏州浪潮智能科技有限公司 | Storage device cluster and management method thereof |
| CN117811987A (en) * | 2023-12-28 | 2024-04-02 | 杭州森格算力技术有限公司 | Method for automatic addressing and management of distributed memory resources |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101753405A (en) | Cluster server memory management method and system | |
| CN102244685B (en) | Distributed type dynamic cache expanding method and system for supporting load balancing | |
| CN101207550B (en) | Load balancing system and method for multi business to implement load balancing | |
| US10257255B2 (en) | Hierarchical organization for scale-out cluster | |
| CN105138615A (en) | Method and system for building big data distributed log | |
| CN105025053A (en) | Distributed file upload method based on cloud storage technology and system | |
| JP2008533564A (en) | Method and apparatus for data management | |
| CN101902498A (en) | Network technology based storage cloud backup method | |
| CN108833462A (en) | A kind of system and method found from registration service towards micro services | |
| CN101969468A (en) | Inquiry server cluster system and inquiry method | |
| CN105847391A (en) | Distributed cloud data central structure | |
| CN102970242A (en) | Method for achieving load balancing | |
| CN108900626A (en) | Date storage method, apparatus and system under a kind of cloud environment | |
| CN103401951B (en) | Based on the elastic cloud distribution method of peer-to-peer architecture | |
| Sadooghi et al. | Fabriq: Leveraging distributed hash tables towards distributed publish-subscribe message queues | |
| Zhao et al. | Research of P2P architecture based on cloud computing | |
| CN101521679A (en) | Self-organizing method based on composite structured peer-to-peer network | |
| Xu et al. | Integration of cloud computing and p2p: A future storage infrastructure | |
| Takemoto et al. | Distributed algorithms for balanced zone partitioning in content-addressable networks | |
| Xia et al. | Heterogeneity and load balance in structured P2P system | |
| Nakazato et al. | Data allocation method considering server performance and data access frequency with consistent hashing | |
| CN104537081A (en) | File management system and method | |
| Qu et al. | Efficient information retrieval by dispatching mobile agents in parallel | |
| CN103795766A (en) | P2SP (Peer to Server & Peer) multi-source download system | |
| WO2022206089A1 (en) | Data internet method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Open date: 20100623 |