CN101753558A - Distributed MRCP server load balancing system and balancing method thereof - Google Patents
Distributed MRCP server load balancing system and balancing method thereof Download PDFInfo
- Publication number
- CN101753558A CN101753558A CN 200910185900 CN200910185900A CN101753558A CN 101753558 A CN101753558 A CN 101753558A CN 200910185900 CN200910185900 CN 200910185900 CN 200910185900 A CN200910185900 A CN 200910185900A CN 101753558 A CN101753558 A CN 101753558A
- Authority
- CN
- China
- Prior art keywords
- mrcp
- resource management
- management server
- server
- heartbeat message
- 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
Abstract
The invention relates to a distributed MRCP server load balancing system, which comprises at least one MRCP resource management server. The MRCP resource management server communicates with one or more MRCP servers through an SIP, a load balancing module is arranged in the MRCP server, and the MRCP resource management server provides MRCP load balancing service for a client through the SIP. The invention also discloses a balancing method for the distributed MRCP server load balancing system. By introducing the MRCP resource management server, the related position and the capacity information of the MRCP servers do not need to be additionally distributed after adding or reducing the number of the MRCP servers during system deployment. Because the load is distributed to a plurality of MRCP servers, the load of each MRCP server is relatively light, the throughput of the system is remarkably increased, and the expansion performance is good.
Description
Technical field
The present invention relates to a kind of distributed MRCP server SiteServer LBS and equalization methods thereof.
Background technology
Media resource control protocol MRCP (Media Resource Control Protocol) is the procotol interface that is used to provide voice related media resource of Internet engineering duty group IETF exploitation.IETF has successively developed two versions, uses second edition in this patent, i.e. MRCPv2.MRCP provides request for client on the network and has controlled the mechanism of voice related media resource, and media resource comprises automatic speech recognition ASR engine, phonetic synthesis tts engine, speaker verification and identification engine etc.The MRCP agreement is operated in the application layer of ISO/OSI network reference model.It is independent of bottom transmission mode and physical network device.The MRCP agreement is not a kind of independently agreement, it uses Session Initiation Protocol (RFC3261) to finish conversation establishing, change and destruction function, use SDP (RFC2327) agreement to describe the type of institute's request resource, relevant parameters such as transmission channel, use RTP/RTCP (RFC3550) agreement to come transmission of audio, the MRCP message that use itself defines is finished the using and controlling of institute's request resource, so MRCP can regard a protocol suite as.
At present most of in the world voice technology manufacturer can both provide the MRCP server product of oneself, and domestic also have some manufacturers that similar products are provided.But the MRCP agreement has only defined the standardised networks interface that uses speech engine, does not define the implementation of this interface, does not also define the characteristic of server product, as performance and autgmentability etc.In actual applications, the MRCP server is mostly as telecom operators' level network platform, its requirement be stablize, efficiently, can expand and uninterruptedly service.Specifically, it is exactly the network platform wants to enlarge by the number of continuous increase server whole system under the situation of non-stop run capacity, also to satisfy the requirement of two aspects simultaneously, dilation process should be transparent for the user on the one hand, i.e. the particular capacity of the imperceptible system of user; On the other hand, system's each server after dilatation should reasonably be shared user's request, reaches load balancing.
Existing MRCP agreement and system have the following disadvantages:
First, though the MRCP agreement has defined the mode of user's initiation session and has been to use Session Initiation Protocol, come the resource of identification request with SIP-URI, but how regulation is not mapped to the session request on the concrete server address, general realization all is that a SIP-URI and a concrete server are interrelated, in case the user has specified SIP-URI just to equal to have specified a concrete server like this, this mechanism one to one can't be expanded, and also forms excessive pressure simultaneously easily and cause server to lose efficacy on a station server; The second, existing MRCP system, though some has management node (server), this management node forms the bottleneck of system easily, in case the management node collapse, whole system function is had a strong impact on; The 3rd, the MRCP agreement has only defined the characteristic of common most of voice application as standard, definitely support for the not relevant mark of the distinctive characteristic of some manufacturers oneself, for example, specifying the age of synthetic speech during TTS uses is children, grow up, or old, specify concrete speaker, it all is the content that standard agreement does not relate to that the code page of specify text (codepage) waits these.
Summary of the invention
Primary and foremost purpose of the present invention is to provide a kind of autgmentability strong, distributed MRCP server SiteServer LBS that can equally loaded.
For achieving the above object, the present invention has adopted following technical scheme: a kind of distributed MRCP server SiteServer LBS, comprise at least one MRCP resource management server, the MRCP resource management server carries out communication by Session Initiation Protocol and one or more MRCP server, a plurality of MRCP servers are formed a server zone in the local area network (LAN), be provided with load balancing module in the MRCP server, client and MRCP resource management server carry out information interaction, and the MRCP resource management server provides MRCP load balancing service by Session Initiation Protocol to client.
Another object of the present invention is to provide a kind of equalization methods of distributed MRCP server SiteServer LBS, this method comprises the step of following order:
(1) client sends solicited message to the MRCP resource management server by Session Initiation Protocol, after the MRCP resource management server receives solicited message, calculates the best MRCP server that related service can be provided according to this solicited message;
(2) the MRCP resource management server is sent to client with the SIP-URI address of best server;
(3) after client is received message, send solicited message to best MRCP server again, best MRCP server is understood Resources allocation after receiving solicited message, until the client ending request.
As shown from the above technical solution, the present invention makes system when disposing by introducing MRCP resource management server, need not in interpolation or the extra capacity information of issuing relevant position and MRCP server after reducing the MRCP number of servers.Because load has been assigned on a plurality of MRCP servers, each MRCP server load is lighter relatively, has significantly increased the throughput of system, and scalability is good.
Description of drawings
Fig. 1 is separate unit MRCP resource management server and heartbeat of Duo Tai MRCP server and service information interaction schematic diagram;
Fig. 2 is the message interaction process schematic diagram of separate unit MRCP resource management server and MRCP server;
Fig. 3 is that many MRCP resource management servers are main and auxiliary, the algorithm flow chart of backup;
Fig. 4 is the network topology structure schematic diagram that many MRCP resource management servers provide load balancing service;
Fig. 5 is the message interaction process schematic diagram of many MRCP resource management servers and MRCP server.
Embodiment
A kind of distributed MRCP server SiteServer LBS, comprise at least one MRCP resource management server 10, MRCP resource management server 10 carries out communication by Session Initiation Protocol and one or more MRCP server 20, a plurality of MRCP servers 20 are formed a server zone in the local area network (LAN), be provided with load balancing module in the MRCP server 20, client and MRCP resource management server carry out information interaction, MRCP resource management server 10 provides MRCP load balancing service by Session Initiation Protocol to client, as shown in Figure 1, 2.
In conjunction with Fig. 1,2, in order to realize that a SIP-URI is to the mapping between a plurality of MRCP servers 20, realize load-balancing function simultaneously, must introduce at least one MRCP resource management server 10, MRCP resource management server 10 has the Session Initiation Protocol stack, use unique SIP-URI to identify oneself, for example sip:mrm@example.com.MRCP resource management server 10 periodically sends oneself heartbeat message-Heart Beat by broadcasting or multicast mode to local area network (LAN), generally speaking, the MRCP resource management server sends one time heartbeat message 10 per 5 seconds.Heartbeat message has comprised the information such as title, IP address, port and type of MRCP resource management server 10.MRCP server 20 utilizes the heartbeat message in the load balancing module reception local area network (LAN), periodically report the information on services-Service Info of oneself simultaneously to MRCP resource management server 10, generally speaking, MRCP server 20 sends one time information on services every 5 seconds, described information on services comprises static information and multidate information two parts, wherein static information comprises nodal information, as service name, IP address, port etc.; Static load information is as CPU model, CPU number, memory size etc.; Engine information, as TTS, engine relevant informations such as ASR.Multidate information comprises dynamic load information, and as current memory usage, cpu busy percentage is licensed number, network utilization etc.; Dynamic statistics information, as network traffics, the service statistics of various engines etc.Described heartbeat message and information on services are the XML form, and this form is with good expansibility, and under message content was on the increase situation, various information need not be encapsulated in the network packet, can separately transmit various types of information.
In conjunction with Fig. 4,5, described MRCP resource management server 10 comprises a main MRCP resource management server 11, an auxilliary MRCP resource management server 12 and an one or more backup MRCP resource management server 13, in arbitrary network, it is main having only a MRCP resource management server 10, also have only a MRCP resource management server 10 to assist, other is backup MRCP resource management server 13.Three kinds of MRCP resource management servers 10 have unified SIP-URI address, and main MRCP resource management server 11 carries out communication with one or more MRCP servers 20.Also can realize the function of MRCP resource management server 10 by a sip proxy server (Proxy), sip proxy server can be to have also can making of state stateless, client requests directly mails to the optimum MRCP server of having selected 20 by this sip proxy server, and the sip message between client and the MRCP server 20 is all acted on behalf of by resource management server alternately like this.
In conjunction with Fig. 4,5, client sends solicited message to main MRCP resource management server 11 by Session Initiation Protocol, described main MRCP resource management server 11 sends heartbeat message to each MRCP server 20, each MRCP server 20 feeds back information on services respectively to main MRCP resource management server 11, after described main MRCP resource management server 11 is received the information on services that comes from each MRCP server 20, calculate the optimization FACTOR P RI of each MRCP server 20 respectively according to formula, computing formula is as follows: PRI=50 * LIC%+30 * CPU%+20 * MEM%, it is to authorize occupancy LIC%, the weighted value of CPU usage CPU% and memory usage MEM%, PRI numerical value is more little to illustrate that then the resource occupation of MRCP server 20 is few more, that is to say that this MRCP server 20 is outstanding more.Main MRCP resource management server 11 is sent to client with the SIP-URI address of best server; After client is received message, send solicited messages to best MRCP server 20 again, best MRCP server 20 is understood Resources allocation after receiving solicited message, until the client ending request,
In conjunction with Fig. 4,5,20 whiles of MRCP server are to the service message of main MRCP resource management server 11 and auxilliary MRCP resource management server 12 reports oneself, the benefit of Chu Liing is like this, can be at once by auxilliary MRCP resource management server 12 adapter work after main MRCP resource management server 11 lost efficacy.In order to allow main and auxiliary and to back up MRCP resource management server 11,12,13 and can realize wrong the recovery automatically, the present invention is used for leader's election algorithm of local area network management, produce main and auxiliary MRCP resource management server 11,12, this algorithm is further described below in conjunction with Fig. 3.
As shown in Figure 3, after described MRCP resource management server 10 starts, judge the heartbeat message that whether has had main MRCP resource management server 11 on the network, if do not exist, then MRCP resource management server 10 sends heartbeat message, and states it oneself is main MRCP resource management server 11; Otherwise, continue to judge the heartbeat message that whether has auxilliary MRCP resource management server 12 on the network, if do not exist, then MRCP resource management server 10 sends heartbeat message, and state it oneself is auxilliary MRCP resource management server 12, otherwise MRCP resource management server 10 announces it oneself is backup MRCP resource management server 13.
As shown in Figure 3, after described MRCP resource management server 10 statements oneself are main MRCP resource management server 11, continue to judge whether to receive the main heartbeat message of other MRCP resource management servers 10, if do not receive, then wait for the main heartbeat message of only surplus next MRCP resource management server 10 in the network; If receive, whether the IP address of then judging other MRCP resource management servers 10 send main heartbeat message is less than the IP address of this MRCP resource management server 10, if judged result is for denying, then wait for the main heartbeat message of only surplus next MRCP resource management server 10 in the network, otherwise, stop to send main heartbeat message, wait for the main heartbeat message of only surplus next MRCP resource management server 10 in the network.
As shown in Figure 3, after described MRCP resource management server 10 statements oneself are auxilliary MRCP resource management server 12, continue to judge whether to receive the auxilliary heartbeat message of other MRCP resource management servers 10, if do not receive, then wait for the auxilliary heartbeat message of only surplus next MRCP resource management server 10 in the network; If receive, whether the IP address of then judging other MRCP resource management servers 10 send auxilliary heartbeat message is less than the IP address of this MRCP resource management server 10, if judged result is for denying, then wait for the auxilliary heartbeat message of only surplus next MRCP resource management server 10 in the network, otherwise, stop to send auxilliary heartbeat message, wait for the auxilliary heartbeat message of only surplus next MRCP resource management server 10 in the network.
As shown in Figure 3, when in waiting until network, only remaining the main heartbeat message of next MRCP resource management server 10, MRCP resource management server 10 judges whether the IP address of this main heartbeat message is identical with the IP address of oneself, if identical, then determines it oneself is main MRCP resource management server 11; Otherwise, wait for the auxilliary heartbeat message of only surplus next MRCP resource management server 10 in the network, when in waiting until network, only remaining the auxilliary heartbeat message of next MRCP resource management server 10, MRCP resource management server 10 judges whether the IP address of this auxilliary heartbeat message is identical with the IP address of oneself, if identical, then determine it oneself is auxilliary MRCP resource management server 12; Otherwise, determine it oneself is backup MRCP resource management server 13.
From the above, because main MRCP resource management server 11 is the heartbeat message of multicast oneself on network constantly, therefore in a single day main MRCP resource management server 11 had lost efficacy, auxilliary MRCP resource management server 12 can perceive this variation at once, it will be automatically upgraded to main MRCP resource management server 11 itself, the heartbeat of broadcasting oneself on network simultaneously.At this time, owing to there has not been auxilliary MRCP resource management server 12, therefore, above algorithm can restart again, chooses a new auxilliary MRCP resource management server 12.By such process, realized the multi-computer back-up mechanism of MRCP resource management server 10, strengthened the fault-tolerance of system.
In addition, because most client-side program are not intercepted the ability that local area network multicast is understood the primary resource server location, the present invention is represented main MRCP Resource Server with standardization agreement: at first, all MRCP resource management servers 10 possess identical SIP-URI, promptly all MRCP resource management server 10 external signs have uniqueness, client need not manually to specify its position as long as this SIP-URI of request just can have access to main MRCP resource management server 11; Secondly, can use the mobile mechanism of Session Initiation Protocol in order to make the user, must dispose a sip agent and registrar on the network, all of user are all transmitted by this agency for the request of primary resource management server; At last, main MRCP resource management server 11 is regularly registered the position of oneself to this sip proxy server.After main MRCP resource management server 11 lost efficacy, can register own new positional information then at first by original log-on message on the SIP REGISTER message erasing sip proxy server by the new main MRCP resource management server 11 that auxilliary MRCP resource management server 12 upgradings form.
Pass through said method, all MRCP resource management servers 10 externally all provide a unified mirror image, the process of MRCP of client-requested service is as shown in Figure 6: client requests is at first sent to sip agent and registrar, determine the position of concrete main MRCP resource management server 11 then by this server, then request is forwarded to this main MRCP resource management server 11.Because sip agent and registrar have very high stability and efficient as the key equipment in the next generation network infrastructure, therefore need not to provide back mechanism to this kind equipment again.
The present invention is on the basis of above-mentioned load balancing, also expanded the SDP agreement in order to improve MRCP for the accuracy of resource request, make it to support further specifying for request resource, concrete method is: filter attributes of the increase after SDP is for the description field m of every kind of resource, attribute is represented with a field in SDP, after having increased this attribute, after MRCP resource management server 10 receives the INVITE request that client sends, inspection has the attribute of filter keyword, according to the setting of parameter that indicates among the filter and value, search qualified optimum MRCP server 20.If can find the MRCP server 20 that satisfies condition then indicate the position of this MRCP server 20 with SIP 302 message, if can not find the MRCP server 20 that satisfies condition fully then return immediate MRCP server 20 positions, if can not find the MRCP server 20 that satisfies condition fully then return SIP 4XX failed message.
Claims (10)
1. distributed MRCP server SiteServer LBS, it is characterized in that: comprise at least one MRCP resource management server, the MRCP resource management server carries out communication by Session Initiation Protocol and one or more MRCP server, a plurality of MRCP servers in the local area network (LAN) are formed a server zone, be provided with load balancing module in the MRCP server, client and MRCP resource management server carry out information interaction, and the MRCP resource management server provides MRCP load balancing service by Session Initiation Protocol to client.
2. distributed MRCP server SiteServer LBS according to claim 1, it is characterized in that: described MRCP resource management server comprises a main MRCP resource management server, an auxilliary resource management server and one or more backup MRCP resource management server, above-mentioned MRCP resource management server has unified SIP-URI address, and main MRCP resource management server and one or more MRCP server carry out communication.
3. the equalization methods of distributed MRCP server SiteServer LBS according to claim 1, this method comprises the step of following order:
(1) client sends solicited message to the MRCP resource management server by Session Initiation Protocol, after the MRCP resource management server receives solicited message, calculates the best MRCP server that related service can be provided according to this solicited message;
(2) the MRCP resource management server is sent to client with the SIP-URI address of best server;
(3) after client is received message, send solicited message to best MRCP server again, best MRCP server is understood Resources allocation after receiving solicited message, until the client ending request.
4. the equalization methods of distributed MRCP server SiteServer LBS according to claim 3, it is characterized in that: described MRCP resource management server comprises a main MRCP resource management server, an auxilliary MRCP resource management server and one or more backup MRCP resource management server, above-mentioned MRCP resource management server has unified SIP-URI address, and main MRCP resource management server and one or more MRCP server carry out communication.
5. the equalization methods of distributed MRCP server SiteServer LBS according to claim 4, it is characterized in that: described main MRCP resource management server sends heartbeat message to each MRCP server, each MRCP server feeds back information on services respectively to main, auxilliary MRCP resource management server, described heartbeat message and information on services are the XML form, described heartbeat message comprises the title of main MRCP resource management server, the IP address, port and type, described information on services comprises static information and multidate information, wherein static information comprises nodal information, static load information and Engine information, multidate information comprise dynamic load information and dynamic statistics information.
6. according to claim 3 or 4 or 5 described distributed MRCP server SiteServer LBSs, it is characterized in that: after described main MRCP resource management server is received the information on services that comes from each MRCP server, calculate the optimization FACTOR P RI of each MRCP server respectively according to formula, computing formula is as follows: PRI=50 * LIC%+30 * CPU%+20 * MEM%, the minimum MRCP server of PRI value is best MRCP server.
7. the equalization methods of distributed MRCP server SiteServer LBS according to claim 4, it is characterized in that: after described MRCP resource management server starts, judge the heartbeat message that whether has had main MRCP resource management server on the network, if do not exist, then the MRCP resource management server sends heartbeat message, and states it oneself is main MRCP resource management server; If there is the heartbeat message of main MRCP resource management server, continue to judge the heartbeat message that whether has auxilliary MRCP resource management server on the network, if do not exist, then the MRCP resource management server sends heartbeat message, and state it oneself is auxilliary MRCP resource management server, otherwise the MRCP resource management server announces it oneself is backup MRCP resource management server.
8. the equalization methods of distributed MRCP server SiteServer LBS according to claim 7, it is characterized in that: after described MRCP resource management server statement oneself is main MRCP resource management server, continue to judge whether to receive the main heartbeat message of other MRCP resource management servers, if do not receive, then wait for the main heartbeat message of only surplus next MRCP resource management server in the network; If receive, whether the IP address of then judging other MRCP resource management servers send main heartbeat message is less than the IP address of this MRCP resource management server, if judged result is for denying, then wait for the main heartbeat message of only surplus next MRCP resource management server in the network, otherwise, stop to send main heartbeat message, wait for the main heartbeat message of only surplus next MRCP resource management server in the network.
9. the equalization methods of distributed MRCP server SiteServer LBS according to claim 7, it is characterized in that: after described MRCP resource management server statement oneself is auxilliary MRCP resource management server, continue to judge whether to receive the auxilliary heartbeat message of other MRCP resource management servers, if do not receive, then wait for the auxilliary heartbeat message of only surplus next MRCP resource management server in the network; If receive, whether the IP address of then judging other MRCP resource management servers send auxilliary heartbeat message is less than the IP address of this MRCP resource management server, if judged result is for denying, then wait for the auxilliary heartbeat message of only surplus next MRCP resource management server in the network, otherwise, stop to send auxilliary heartbeat message, wait for the auxilliary heartbeat message of only surplus next MRCP resource management server in the network.
10. according to Claim 8 or the equalization methods of 9 described distributed MRCP server SiteServer LBSs, it is characterized in that: when in waiting until network, only remaining the main heartbeat message of next MRCP resource management server, the MRCP resource management server judges whether the IP address of this main heartbeat message is identical with the IP address of oneself, if identical, then determine it oneself is main MRCP resource management server; Otherwise, wait for the auxilliary heartbeat message of only surplus next MRCP resource management server in the network, when in waiting until network, only remaining the auxilliary heartbeat message of next MRCP resource management server, the MRCP resource management server judges whether the IP address of this auxilliary heartbeat message is identical with the IP address of oneself, if identical, then determine it oneself is auxilliary MRCP resource management server; Otherwise, determine it oneself is backup MRCP resource management server.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200910185900 CN101753558B (en) | 2009-12-11 | 2009-12-11 | Balancing method of distributed MRCP server load balancing system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200910185900 CN101753558B (en) | 2009-12-11 | 2009-12-11 | Balancing method of distributed MRCP server load balancing system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101753558A true CN101753558A (en) | 2010-06-23 |
| CN101753558B CN101753558B (en) | 2013-03-27 |
Family
ID=42479964
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200910185900 Active CN101753558B (en) | 2009-12-11 | 2009-12-11 | Balancing method of distributed MRCP server load balancing system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101753558B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2012159355A1 (en) * | 2011-07-29 | 2012-11-29 | 华为技术有限公司 | Service distribution method and device in communication system |
| CN105515839A (en) * | 2015-11-30 | 2016-04-20 | 上海斐讯数据通信技术有限公司 | System and method for promoting Radius service performance |
| CN105635199A (en) * | 2014-10-28 | 2016-06-01 | 广州睿哲网络科技有限公司 | Method and device for implementation of self-organization cluster server supporting load balancing |
| CN107567706A (en) * | 2015-05-05 | 2018-01-09 | 意大利电信股份公司 | Session of subscriber in communication network is distributed again |
| CN107749931A (en) * | 2017-09-29 | 2018-03-02 | 携程旅游信息技术（上海）有限公司 | Method, system, equipment and the storage medium of interactive voice answering |
| CN108243170A (en) * | 2016-12-27 | 2018-07-03 | 青岛融贯汇众软件有限公司 | Data access system and method based on socket frames |
| WO2019001092A1 (en) * | 2017-06-30 | 2019-01-03 | 华为技术有限公司 | Load balancing engine, client, distributed computing system, and load balancing method |
| CN111158915A (en) * | 2019-12-31 | 2020-05-15 | 厦门快商通科技股份有限公司 | Master-slave relationship switching method, slave server, master server and system |
| CN112217859A (en) * | 2020-08-28 | 2021-01-12 | 厦门快商通科技股份有限公司 | Load balancing method, system, mobile terminal and storage medium |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1495634A (en) * | 2002-06-27 | 2004-05-12 | 上海汉唐科技有限公司 | Server clustering load balancing method and system |
| CN101431532A (en) * | 2008-12-15 | 2009-05-13 | 中国电信股份有限公司 | Content routing method, load balancing equipment and resource management equipment |
-
2009
- 2009-12-11 CN CN 200910185900 patent/CN101753558B/en active Active
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2012159355A1 (en) * | 2011-07-29 | 2012-11-29 | 华为技术有限公司 | Service distribution method and device in communication system |
| CN103053144A (en) * | 2011-07-29 | 2013-04-17 | 华为技术有限公司 | Service distribution method and device in communication system |
| CN103053144B (en) * | 2011-07-29 | 2016-01-06 | 华为技术有限公司 | The method and apparatus of communication system traffic distribution |
| CN105635199A (en) * | 2014-10-28 | 2016-06-01 | 广州睿哲网络科技有限公司 | Method and device for implementation of self-organization cluster server supporting load balancing |
| CN105635199B (en) * | 2014-10-28 | 2019-03-15 | 睿哲科技股份有限公司 | A kind of self-organizing cluster server of holding load equilibrium |
| CN107567706A (en) * | 2015-05-05 | 2018-01-09 | 意大利电信股份公司 | Session of subscriber in communication network is distributed again |
| CN105515839A (en) * | 2015-11-30 | 2016-04-20 | 上海斐讯数据通信技术有限公司 | System and method for promoting Radius service performance |
| CN108243170A (en) * | 2016-12-27 | 2018-07-03 | 青岛融贯汇众软件有限公司 | Data access system and method based on socket frames |
| CN109218355A (en) * | 2017-06-30 | 2019-01-15 | 华为技术有限公司 | Load equalizing engine, client, distributed computing system and load-balancing method |
| WO2019001092A1 (en) * | 2017-06-30 | 2019-01-03 | 华为技术有限公司 | Load balancing engine, client, distributed computing system, and load balancing method |
| CN107749931A (en) * | 2017-09-29 | 2018-03-02 | 携程旅游信息技术（上海）有限公司 | Method, system, equipment and the storage medium of interactive voice answering |
| CN111158915A (en) * | 2019-12-31 | 2020-05-15 | 厦门快商通科技股份有限公司 | Master-slave relationship switching method, slave server, master server and system |
| CN112217859A (en) * | 2020-08-28 | 2021-01-12 | 厦门快商通科技股份有限公司 | Load balancing method, system, mobile terminal and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101753558B (en) | 2013-03-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101753558B (en) | Balancing method of distributed MRCP server load balancing system | |
| CN101925045B (en) | Method and system for providing user service data | |
| US20070136469A1 (en) | Load Balancing and Failover of Distributed Media Resources in a Media Server | |
| EP1788759B1 (en) | Wireless network device and resource allotting method for the same | |
| US7293085B2 (en) | Data processing system, information processing apparatus, data processing method, and computer program | |
| US20080205327A1 (en) | Methods, Apparatuses, and Computer Program Products for Providing Dynamic Assignment of Slot Cycle Index for Network Nodes | |
| WO2004114633A1 (en) | Systems and methods for event semantic binding in networks | |
| CN103634490A (en) | A gateway for the survivability of an enterprise network using SIP | |
| CN102457628B (en) | Calling method, system and voice server for Internet protocol phone | |
| WO2009155771A1 (en) | Resource allocation method, server, network device and network system | |
| CN101599924A (en) | Communication system | |
| CN101395891A (en) | Method and system for characterising heterogeneous communication nodes | |
| EA012519B1 (en) | Method for allocation at least one user data link to at least one multiplex connection | |
| US8892751B2 (en) | Method, system and network entity for negotiating the session description protocol version and obtaining the session description protocol version information | |
| US7483369B2 (en) | Method and apparatus for migrating to an alternate call controller | |
| US8539051B2 (en) | Method for the transmission of program updates for program-controlled devices in a communication network | |
| CN106970843B (en) | Remote calling method and device | |
| CN100486329C (en) | Method for interface adaptation between IPTV and stream media equipment | |
| CN101159910A (en) | Synchronous present state method and cluster server and terminal for implementing the method | |
| US20080043723A1 (en) | Arrangement and method for reducing required memory usage between communication servers | |
| CN101325564B (en) | Method, apparatus and system for selecting virtual medium gateway | |
| CN101321139A (en) | Resource management method, bearing equipment and bearing control equipment | |
| CN110474781B (en) | Method and device for forwarding multicast data | |
| CN101459631A (en) | Virtual media gateway selection method, apparatus and system | |
| CN101741873B (en) | Terminal name changing method and fusion service 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 | ||
| ASS | Succession or assignment of patent right |
Owner name: USTC IFLYTEC (BEIJING) CO., LTD. Free format text: FORMER OWNER: ANHUI USTC IFLYTEK CO., LTD. Effective date: 20120516 |
|
| C41 | Transfer of patent application or patent right or utility model | ||
| COR | Change of bibliographic data |
Free format text: CORRECT: ADDRESS; FROM: 230088 HEFEI, ANHUI PROVINCE TO: 100000 HAIDIAN, BEIJING |
|
| TA01 | Transfer of patent application right |
Effective date of registration: 20120516 Address after: 100000, Beijing Haidian District Yongfeng industrial base, North Wing Road, No. 3, the standard workshop A block 4, room 418 Applicant after: Zhongke Xunfei Internet (Beijing) Information Technology Co.,Ltd. Address before: 230088 No. 616, Mount Huangshan Road, hi tech Zone, Anhui, Hefei Applicant before: ANHUI USTC IFLYTEK Co.,Ltd. |
|
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address |
Address after: Room 311-2, 3 / F, building 5, east yard, No. 10, northwest Wangdong Road, Haidian District, Beijing 100094 Patentee after: iFLYTEK (Beijing) Co.,Ltd. Address before: 100000 Room 418, Floor 4, Building A, Standard Plant, No. 3, Yongjie North Road, Yongfeng Industrial Base, Haidian District, Beijing Patentee before: Zhongke Xunfei Internet (Beijing) Information Technology Co.,Ltd. |
|
| CP03 | Change of name, title or address |