CN1835607A - Short message two-staged netgate based on PC server and service environment - Google Patents
Short message two-staged netgate based on PC server and service environment Download PDFInfo
- Publication number
- CN1835607A CN1835607A CNA2006100464105A CN200610046410A CN1835607A CN 1835607 A CN1835607 A CN 1835607A CN A2006100464105 A CNA2006100464105 A CN A2006100464105A CN 200610046410 A CN200610046410 A CN 200610046410A CN 1835607 A CN1835607 A CN 1835607A
- Authority
- CN
- China
- Prior art keywords
- client
- message
- server
- msgswitch
- oid
- 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
Links
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The platform comprises low short message gateway, server and service generator, and has a logic structure with three layers. Wherein, the firs layer is a transmission layer, the second one is a platform layer, and the third one is a service layer. The service generator adopts IFE. The server is Msgswitch. The application environment of the middleware Msgswitch of the message exchanging server comprises: multi programs, especially multi language mixing programs; multi data sources or isomeric data bases; 'old' or existed application saved in client server; the system enhancement, addition and modification to be made; the communication, between complex applications, including communication in an enterprise and communication between enterprises.
Description
(1) technical field
The present invention relates to the network communications technology, particularly relate to a kind of efficient means of communication of short message two-staged netgate and service environment, the user only needs to be grasped simple interface, just can finish be applied between the application efficiently, communication easily.Realized that simultaneously short message two-staged netgate moves on PC level server, but can reach and a kind of software implementation method that on minicomputer, moves same performance.
(2) background technology
Mainly quoting the document that relates to background technology as proof has:
" PHS short messaging gateway technical specification first fascicle short messaging gateway and service provider (SP) interface specification (CNGP) V2.0 ";
" PHS short messaging gateway technical specification (the second fascicle short messaging gateway and charging center interface specification V2.0) ";
" PHS short messaging gateway technical specification first fascicle short messaging gateway and service provider (SP) interface specification (CNGP) V2.0 ";
Ministry of Information Industry's " the point-to-point short message internetworking protocol of YD/T 1291-2003 requires ";
The database that present provincial two-staged netgate uses is general large-scale relevant database, for the system that requires that satisfies high performance index all moves on minicomputer.And our product meets the requirements of performance index and only need move on the PC server owing to adopted the embedded database of independent development.
Existing systems is respectively used communication following shortcoming:
1, point-to-point application interface is finished the integrated complexity height of application, and cost is big, seriously limits application flexibility.
2, can not provide use with use between the data transaction transmitted, transmission according to priority, in real time or batch fashion carry out.
3, the complicated service processing logic bring between application, mutual complexity between application and the people, a kind of simplicity of supporting flow management, handling intelligent exploitation can not be provided.
4, integrated system structure of application can not adapt to service logic, institutional variation.
(3) summary of the invention
Purpose of the present invention is at the problems referred to above, provides a kind of based on PC server short message two-staged netgate and service environment, constitutes quick network service.
The technical scheme that adopts is:
Based on PC server short message two-staged netgate and service environment, platform comprises note secondary net, server and professional maker, it is characterized in that the platform logical construction is divided into three layers, ground floor is a transport layer, the second layer is that podium level, the 3rd layer are professional industry layer, professional maker adopts IFE, and server is Msgswitch, and message exchanges server middleware Msgwitch applied environment and comprises:
A plurality of programs are arranged, particularly mix the multilingual program;
The database of a plurality of data sources or isomery;
In client/server, exist " old " or existing application;
Be about to carry out enhancing, increase and the modification of system;
The communication of the existing enterprises of communication between complicated application has the communication between enterprise again.
The MsgSwitch message exchange server can compile in 32/64 multitask system and move and move based on the textual interface control desk.Remote debugging function and remote console are provided, can carry out remote monitoring by the TCP network with telnet or XRC long-range.Transmit message based on node address.Node address is 4 numerals that are similar to the IP address, opens with ". " between the numeral, and each digital scope is 1-255.Node also can be with 32 integer representations, be numbered one 32 integer, this integer can be converted by the function AMSAddreeeToInt cause node address in the client function library, and also can change an integer with AMSIntToAddress is node address.4 numerals in the node address all do not allow for numeral 0.Keep 0 in the MsgSwitch service routine.
MsgSwitch supports the message monitoring function, by the message monitoring client, uses Transmission Control Protocol to exchange messages, and port default is 10000.The Socket operational mode is the asynchronous mode preference pattern, can support a large amount of networks to connect, and can be supported by the majority operation system.Program adopts control desk order line control mode, and during operation, program will wait user input commands.The user can understand the ruuning situation and the control operation of system by order.
MsgSwitch is divided into two logical gates to the identification of client, is respectively Socket identification division and Client identification division, and each socket is connected to a Socket client, and a Socket client can support a plurality of Client nodes simultaneously.Socket client is the operating system Automatic Logos, and the Client node identifies by node serial number or node address.
The client-side interface storehouse can connect two MsgSwitch servers simultaneously, when client sends the message that needs forwarding, can respectively send a message by two servers, two servers all will be transmitted message independently of one another, receive client and can receive 2 identical message simultaneously, at this moment will delete the message that the back arrives, and only keep first message, and produce the OnMessage incident.Client can detect the server of inefficacy automatically again, and after the server that lost efficacy was resumed work, client can reconnect, register Client.MsgSwitch server and client side Socket supports the ActiveTest service; When starting the ActiveTest service, the MsgSwitch server end is given each Socket client with timed sending ActiveTest message, if client also starts ActiveTest.Client can automatically reply this message after receiving ActiveTest message; The MsgSwitch client supports inside story function of exchange (InternalSwitch) can realize the exchange and the information broadcast of message under the situation that does not connect the MsgSwitch server.
Above-mentioned Only DB embedded database is the comprehensive of a kind of object-oriented and network model data base method for designing.Object is to identify by unique object ID (OID) in inside, and this mode can be separated the ID of object and the attribute of self; Use B+ tree storage/access mode.A kind of such B+ tree index being arranged, be referred to as master index, in a set is that the form of the serial of major key has comprised all objects with OID.Also have additional B+ to set and safeguard index attributes.These attributes provide the mapping of a kind of dependency value to object OID.The combination of master index and secondary index makes that can carry out object by OID or index attributes value searches.Each object has a set of relationship with other object.
Only DB embedded database is adopted concurrent performance lock management; Only DB write operation is updated into database with all modifications.Use the get of B+ tree, set, and get-next service speed; Utilized local reference to improve buffer memory; Object storage is in the master index of OID as major key.Stored the serial object data in master index, the major key in the index is that OID and OID are orders, and makes each new object at first all be endowed an interim OID, and this interim OID will be replaced by a lasting OID when then submitting to.Only DB has introduced " group is submitted to " function, and so, application program will be finished identical functions and move with regard to not needing to do other.
The present invention is as the part of message-oriented middleware (MOM), and (Enterprise Application Integration, EAI) effect in becomes clear day by day the Message Agent middleware in enterprise's application integration.Under the distributed environment, service unit scatters everywhere, functions such as the Message Agent middleware mainly provides the sending of the necessary data of application integration, collects, translates, filters, mapping and route, shield wide gap and difference between different hardware platforms, database, message format, the communication protocol, provide be applied between the application efficiently, communication capacity easily.
(4) description of drawings
Fig. 1 is system logic structure figure.
Fig. 2 is that middle figure Netcom two-staged netgate is its location drawing in whole system of example.
(5) embodiment
The forwarding capability that the MsgSwitch message exchange server mainly gives information.Based on the adaptive communications Environmental Technology, can in nearly all 32/64 multitask system, compile and move and move based on the textual interface control desk.Remote debugging function and remote console are provided, can carry out remote monitoring by the TCP network with telnet or XRC (a same character mode Long-distance Control instrument) long-range based on adaptive technique.It transmits message based on node address.Node address is 4 numerals that are similar to the IP address, uses between the numeral ". " separately, each digital scope is 1-255.For example: 10.1.1.1 is exactly a legal node address, node also can be with 32 integer representations, be numbered one 32 integer, this integer can be converted by the function AMSAddreeeToInt cause node address in the client function library, and also can change an integer with AMSIntToAddress is node address.4 numerals in the node address all do not allow for numeral 0.For example: 10.1.0.1,10.0.0.1,0.1.1.1,10.1.1.0 are illegal addresses.For the MsgSwitch service routine, 0 the representative be broadcast address.Therefore kept.Example: send a message to the 10.1.1.0 address, then expression sends the node that identical message has been registered to each of 10.1.1.1-10.1.1.255.Send a message to address 10.1.0.0 and represent to send identical message to 10.1.1.1-10.1.255.255.Send a message to the 0.0.0.0 address and then represent to send a message to all nodes.
MsgSwitch supports the message monitoring function, and by the message monitoring client, the user can see all message through MsgSwitch at an easy rate.The use Transmission Control Protocol exchanges messages, and port default is 10000.The Socket operational mode is the asynchronous mode preference pattern, and number of connection and number of threads are irrelevant, can support a large amount of networks to connect, and can be supported by the majority operation system.Program adopts control desk order line control mode, and during operation, program will wait user input commands.The user can understand the ruuning situation and the control operation of system by order.
MsgSwitch is divided into two logical gates to the identification of client, is respectively Socket identification division and Client identification division, and each socket is connected to a Socket client, and a Socket client can support a plurality of Client nodes simultaneously.Socket client is the operating system Automatic Logos, and the Client node identifies by node serial number or node address.By definition to the special joint address, server support broadcasting or multicast services.
MsgSwitch message service itself does not need to provide redundancy scheme, and for the environment that needs high reliability, it is fault-tolerant to carry out hot-standby redundancy to MsgSwitch, and this function is realized by the client-side interface storehouse.In order to obtain the redundancy fault-tolerant ability, the client-side interface storehouse can connect two MsgSwitch servers simultaneously, when client sends the message that needs forwarding, can respectively send a message by two servers, two servers all will be transmitted message independently of one another, receive client and can receive 2 identical message simultaneously, at this moment will delete the message that the back arrives, and only keep first message, and produce the OnMessage incident.When a server breaks down, for example situation such as power down the time, message still can be passed through another server forwards, can not have any impact to system.Client can detect the server of inefficacy automatically again, and after the server that lost efficacy was resumed work, client can reconnect, register Client.This redundancy fault-tolerant mechanism compares with other modes, and advantage is simple, is easy to maintenance upgrade, and sufficiently high reliability and stability can be provided.Shortcoming is that message all is to send in pairs, and sending message number and being does not have 2 times of redundant system.Therefore operating efficiency is lower.This shortcoming can be come partial offset by the efficient that improves server.According to test, at common low-grade server, the client transmission rate can surpass for 10000 bag/seconds.
MsgSwitch server and client side Socket supports ActiveTest service (heartbeat service), by ActiveTest service, whether client and MsgSwitch server itself can be found to be connected because unpredictable former thereby send interruption at an easy rate.In discovery, have no progeny the connection of server meeting broken clients end.And whether client will resume operation by the automatic inspection server, after the normal operation of discovery server, can reconnect automatically.When starting the ActiveTest service, the MsgSwitch server end is given each Socket client with timed sending ActiveTest message, if client also starts ActiveTest.Client can automatically reply this message after receiving ActiveTest message, do not need programming or manual intervention.If client is not activated the ActiveTest service then needs artificial treatment ActiveTest message, and responds this message, could realize the disconnection reconnect function of client like this.Attention: the MsgSwitch server does not have (to interrupt all can causing sending failure because connect in any case, so do not need to check and whether returned ActiveTest) overtime measuring ability, that is to say, even client is not responded ActiveTest message, server can this connection of active break yet, unless client initiatively connects.Server is not supported the ActiveTest function of each Client node, can waste a large amount of bandwidth because do like this, makes that the cost of network communication is expensive more.So if desired function needs the user to realize.
After message sends to the MsgSwitch server, if destination address is not registered in server, then can return address error message (ERRORADDRESS), if under the situation of using dual computer fault-tolerant, may all not find this address, will return 2 error messages so because of 2 servers, if one has been found the address, another does not find, and will only return one so, and message can be by the client of reliable transmission to appointment in this case.
The MsgSwitch client supports inside story function of exchange (InternalSwitch) can realize the exchange and the information broadcast of message under the situation that does not connect the MsgSwitch server.Also can under the situation of Connection Service device, improve communication efficiency by the inner exchanging function.Inner exchanging realizes that by detecting local register list if find that the address in this locality, then excites the OnMessage incident automatically, message does not need can directly be forwarded to by server the client of appointment.
When the MsgSwitch client is not connected to server, if there is message to send, also can produce address error message, at this time should check the connection situation of client, if determine not connect, then should after connection, resend this message.
Second problem to be solved by this invention is very to the high performance index requirement of short message two-staged netgate, in the Only DB embedded database that independent development realizes how with complicated and the potential data pattern that can dynamically adjust is reflected the right problem of key/data that is set to Only DB.Our logical data pattern comprises a lot of different entity classes and the myriad relationships between them, and we just need certain to design the adaptability that guarantees this pattern.Wherein the most important is, our per second must be handled at least 300 affairs and can hold 15000000 the record/objects of average object size for about 1KB in a database.Typical affairs will comprise read and write access to database to create a record.Since our database size has surpassed the amount of ram that target hardware can provide, we must support the efficient read and write access to hard disk.
Be the comprehensive of a kind of object-oriented and network model data base method for designing very to our solution of embedded database Only DB.Our database is the set of object-name attribute and relation with data exposed.Relation provides a kind of mechanism to search the object that is associated, and index attributes then provides a kind of ability of searching object by property value.Object is to identify by unique object ID (OID) in inside, and this point is with the realization extraordinary image of OODBMS.This mode can be separated the ID of object and the attribute of self, and makes and be more prone to when property value upgrades and new logic more effectively.
In design, we have used B+ tree storage/access mode.A kind of such B+ tree index being arranged, be referred to as master index, in a set is that the form of the serial of major key has comprised all objects with OID.
Also have additional B+ to set and safeguard index attributes.These attributes provide the mapping of a kind of dependency value to object OID, and in essence, the secondary index of searching as object exists.
The combination of master index and secondary index makes that can carry out object by OID or index attributes value searches.
At last, each object has a set of relationship with other object.
Concurrent performance lock management
Owing in using a large amount of concurrent operations are arranged, we have to guarantee that any moment all has only one to upgrade affairs in Only DB.Simultaneously, we do not wish to make the client of server to only limit to use concurrent renewal affairs yet.Our solution is the concurrent renewal affairs that allow user class, and the renewal that batch processing is all is write into database to all renewals in single-threaded mode again up to submitting to.The user class affairs produce all more of being recorded in the internal memory by us in new capital revise in the collection.When the user class affairs were submitted to, we started new affairs and carry out Only DB write operation, and all modifications are updated into database.When realizing, we revise collection for each one of B+ tree index maintenance, and use to revise to collect and follow the tracks of each modification of user class affairs.Any all must collection by modification the visit of indexed data storehouse carry out, and changes and can both correctly be reflected by the transactional that B+ tree index accesses causes with assurance.Such as, if the user has deleted an object, do not submit all will failing in the affairs of working together so to searching all of this object.In fact, this is to have had record to show that object just deleted in executory affairs by this because revise collection.
Single-threaded visit may move that some is slow, but in our application, but can't see any big decreased performance.This must give the credit to the atomic operation relatively fast of Only DB.We only use the get of B+ tree, set, and and get-next operation, and do not have complicated SQL query.We can also satisfy even exceed the performance requirement of 300 affairs each second.
Our operable lock strategy also has other benefit.The lock granularity of our common service recorder/object one-level, sometimes, too coarse even this granularity also may seem.Because our control lock granularity, we just can do very exquisitely.The lock management of customization makes us improve performance, and we can be dealt with and the hot zone with better adaptability when model and access module variation.
Solve the disk read-write performance bottleneck
The main bottleneck that we will solve in our application program is a disk read-write, and the transfer of data between storage and the run time format (serialization).
Only DB has a high-speed cache in the internal memory, is used for writing down the most frequently used page.Yet in the face of at random access module and a large database, it is also thus much that this high-speed cache can be done.The result is exactly the factor that disk access time (mainly being seek time) becomes Performance Constraints.
Our design has utilized local reference to improve buffer memory.In essence, the object that uses together usually generally is near creating simultaneously.B+ tree for Only DB is realized, very big may being stored in the same B+ tree node that record that the order major key is arranged that order is created, thereby also just in identical disk block.Our object storage is in the master index of OID as major key.Like this, near two objects creating simultaneously, its OID numerically also is approaching.This just means that very big these two objects of possibility will be stored in identical disk page or leaf.If these targets are used simultaneously, this page is also as long as read once just passable from disk.
The another kind of approach of reading performance that improves is to improve buffer efficiency.Buffer efficiency depends on the tightness degree of data and the situation of utilizing that B+ sets the page.When keyset was determined, order increases just can make utilized situation best.Here it is when tree increases, the reason that the B+ tree page separates.
Index maximum in our realization is exactly a master index, because it has stored the serial object data.Because the major key in this index is OID and OID is order, and for reaching desirable utilance, the generation order that guarantees OID that we will do is consistent with the order that object is submitted to database.In order to reach this target, at first each new object all is endowed an interim OID, and this interim OID will be replaced by a lasting OID when then submitting to.
Writing disk also is the bottleneck that we will solve.When submitting to, one of transaction journal will be write (flush) disk to guarantee persistence.On the other hand, writing buffering also makes disk move the effective and efficient manner process of arranging the time with magnetic head.Challenge wherein is how to write a plurality of affairs of buffering and simultaneously keep the affairs persistence, and this persistent request data must be stored in the long-time memory before submit operation returns.
Only DB has a useful function, and it supports the asynchronous write affairs.When submitting affairs to, needn't write data in the daily record at once.Then you can write these bufferings respectively with follow-up function.These characteristics make us can handle a plurality of affairs in single disk.
Only DB has introduced " group is submitted to " function, and so, application program will be finished identical functions and move with regard to not needing to do other.
Last bottleneck problem that we handle relates to the serialization expense.Data serializing is a process that object presentation in the internal memory is transformed into binary form.In our example, we need object of serialization and all its relation and attributes, and these relations and attribute may be complicated types.
When addressing this problem, we have realized the serialization of customization, and the predefined object class ID of compactness is provided for all important classes.This just feasible storage is compact more, and compact storage causes better serialization performance again, and to the read-write of disk, better buffer efficiency has suffered because present more data can be stored in buffer memory faster.
Claims (1)
1, based on PC server short message two-staged netgate and service environment, platform comprises note secondary net, server and professional maker, it is characterized in that the platform logical construction is divided into three layers, ground floor is a transport layer, the second layer is that podium level, the 3rd layer are professional industry layer, professional maker adopts IFE, and server is Msgswitch, and message exchanges server middleware Msgwitch applied environment and comprises:
A plurality of programs are arranged, particularly mix the multilingual program;
The database of a plurality of data sources or isomery;
In client/server, exist " old " or existing application;
Be about to carry out enhancing, increase and the modification of system;
The communication of the existing enterprises of communication between complicated application has the communication between enterprise again;
The MsgSwitch message exchange server can compile in 32/64 multitask system and move and move based on the textual interface control desk; Remote debugging function and remote console are provided, can carry out remote monitoring by the TCP network with telnet or XRC long-range; Transmit message based on node address; Node address is 4 numerals that are similar to the IP address, and with ". " separately, each digital scope is 1-255 between the numeral; Node also can be with 32 integer representations, be numbered one 32 integer, this integer can be converted by the function AMSAddreeeToInt cause node address in the client function library, and also can change an integer with AMSIntToAddress is node address; 4 numerals in the node address all do not allow for numeral 0; Keep 0 in the MsgSwitch service routine;
MsgSwitch supports the message monitoring function, by the message monitoring client, uses Transmission Control Protocol to exchange messages, and port default is 10000; The Socket operational mode is the asynchronous mode preference pattern, can support a large amount of networks to connect, and can be supported by the majority operation system; Program adopts control desk order line control mode, and during operation, program will wait user input commands; The user can understand the ruuning situation and the control operation of system by order;
MsgSwitch is divided into two logical gates to the identification of client, is respectively Socket identification division and Client identification division, and each socket is connected to a Socket client, and a Socket client can support a plurality of Client nodes simultaneously; Socket client is the operating system Automatic Logos, and the Client node identifies by node serial number or node address;
The client-side interface storehouse can connect two MsgSwitch servers simultaneously, when client sends the message that needs forwarding, can respectively send a message by two servers, two servers all will be transmitted message independently of one another, receive client and can receive 2 identical message simultaneously, at this moment will delete the message that the back arrives, and only keep first message, and produce the OnMessage incident; Client can detect the server of inefficacy automatically again, and after the server that lost efficacy was resumed work, client can reconnect, register Client; MsgSwitch server and client side Socket supports the ActiveTest service; When starting the ActiveTest service, the MsgSwitch server end is given each Socket client with timed sending ActiveTest message, if client also starts ActiveTest; Client can automatically reply this message after receiving ActiveTest message; The MsgSwitch client supports inside story function of exchange (InternalSwitch) can realize the exchange and the information broadcast of message under the situation that does not connect the MsgSwitch server;
Above-mentioned Only DB embedded database is the comprehensive of a kind of object-oriented and network model data base method for designing; Object is to identify by unique object ID (OID) in inside, and this mode can be separated the ID of object and the attribute of self; Use B+ tree storage/access mode; A kind of such B+ tree index being arranged, be referred to as master index, in a set is that the form of the serial of major key has comprised all objects with OID; Also have additional B+ to set and safeguard index attributes; These attributes provide the mapping of a kind of dependency value to object OID; The combination of master index and secondary index makes that can carry out object by OID or index attributes value searches; Each object has a set of relationship with other object;
Only DB embedded database is adopted concurrent performance lock management; Only DB write operation is updated into database with all modifications; Use the get of B+ tree, set, and get-next service speed; Utilized local reference to improve buffer memory; Object storage is in the master index of OID as major key; Stored the serial object data in master index, the major key in the index is that OID and OID are orders, and makes each new object at first all be endowed an interim OID, and this interim OID will be replaced by a lasting OID when then submitting to; Only DB has introduced " group is submitted to " function, and so, application program will be finished identical functions and move with regard to not needing to do other.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CNB2006100464105A CN100486345C (en) | 2006-04-25 | 2006-04-25 | Business system based on PC server |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CNB2006100464105A CN100486345C (en) | 2006-04-25 | 2006-04-25 | Business system based on PC server |
Publications (2)
Publication Number | Publication Date |
---|---|
CN1835607A true CN1835607A (en) | 2006-09-20 |
CN100486345C CN100486345C (en) | 2009-05-06 |
Family
ID=37003177
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CNB2006100464105A Expired - Fee Related CN100486345C (en) | 2006-04-25 | 2006-04-25 | Business system based on PC server |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN100486345C (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102118452A (en) * | 2011-04-18 | 2011-07-06 | 北京神州数码思特奇信息技术股份有限公司 | Three-layer server architecture applied in B/S (browser/server) structure |
CN101119392B (en) * | 2007-08-21 | 2012-04-25 | 南京联创科技集团股份有限公司 | SOCKET united access method |
CN102945242A (en) * | 2006-11-01 | 2013-02-27 | 起元技术有限责任公司 | Managing storage method, system, and computer system |
CN104794119A (en) * | 2014-01-17 | 2015-07-22 | 阿里巴巴集团控股有限公司 | Middleware message storage and transmission method and system |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104788769A (en) * | 2014-05-29 | 2015-07-22 | 天长市远洋船舶设备有限公司 | Modified butadiene-acrylonitrile rubber material for ship |
-
2006
- 2006-04-25 CN CNB2006100464105A patent/CN100486345C/en not_active Expired - Fee Related
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102945242A (en) * | 2006-11-01 | 2013-02-27 | 起元技术有限责任公司 | Managing storage method, system, and computer system |
CN101119392B (en) * | 2007-08-21 | 2012-04-25 | 南京联创科技集团股份有限公司 | SOCKET united access method |
CN102118452A (en) * | 2011-04-18 | 2011-07-06 | 北京神州数码思特奇信息技术股份有限公司 | Three-layer server architecture applied in B/S (browser/server) structure |
CN104794119A (en) * | 2014-01-17 | 2015-07-22 | 阿里巴巴集团控股有限公司 | Middleware message storage and transmission method and system |
CN104794119B (en) * | 2014-01-17 | 2018-04-03 | 阿里巴巴集团控股有限公司 | Storage and transmission method and system for middleware message |
Also Published As
Publication number | Publication date |
---|---|
CN100486345C (en) | 2009-05-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20210203742A1 (en) | Providing access to managed content | |
US9122698B2 (en) | Peer-to-peer redundant file server system and methods | |
CN1320483C (en) | System and method for implementing journaling in a multi-node environment | |
CN101535965B (en) | Technique for improving scalability and portability of a storage management system | |
US7512673B2 (en) | Rule based aggregation of files and transactions in a switched file system | |
US6889249B2 (en) | Transaction aggregation in a switched file system | |
CN1132109C (en) | System and method for efficient caching in a distributed file system | |
US7788335B2 (en) | Aggregated opportunistic lock and aggregated implicit lock management for locking aggregated files in a switched file system | |
US7509322B2 (en) | Aggregated lock management for locking aggregated files in a switched file system | |
CA2512312C (en) | Metadata based file switch and switched file system | |
CN102411598B (en) | Method and system for realizing data consistency | |
CN102088484B (en) | Write lock method of distributed file system and a system thereof | |
CN1723434A (en) | Apparatus and method for a scalable network attach storage system | |
CN111984696B (en) | Novel database and method | |
CN101576915A (en) | Distributed B+ tree index system and building method | |
CN103399894A (en) | Distributed transaction processing method on basis of shared storage pool | |
CN102981857A (en) | Method for loading parallel compressed mass data of database cluster | |
CN100486345C (en) | Business system based on PC server | |
CN1324844C (en) | Method for providing data to members of network synchronously by gating system | |
US6847968B2 (en) | Method for facilitating access to remote files | |
CN101377788B (en) | Method and system of caching management in cluster file system | |
CN114253929A (en) | Network disk system architecture based on distributed file storage | |
CN100471112C (en) | Memory-network memory-magnetic disc high speed reliable storage system and its reading/writing method | |
CN1831810A (en) | Method and middleware for data base duplicating using middleware in group communication system | |
CN1203427C (en) | Load balance modulator possessing TCP connection fault tolerant function and its modulating method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20090506 Termination date: 20160425 |