CN102571686B - Method for realizing cloud conference system - Google Patents

Method for realizing cloud conference system Download PDF

Info

Publication number
CN102571686B
CN102571686B CN201010582161.8A CN201010582161A CN102571686B CN 102571686 B CN102571686 B CN 102571686B CN 201010582161 A CN201010582161 A CN 201010582161A CN 102571686 B CN102571686 B CN 102571686B
Authority
CN
China
Prior art keywords
cloud
meeting
media
conference system
node
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201010582161.8A
Other languages
Chinese (zh)
Other versions
CN102571686A (en
Inventor
李俊超
郭锐锋
孙建伟
李鸿彬
杨海波
林镜华
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenyang Institute of Computing Technology of CAS
Original Assignee
Shenyang Institute of Computing Technology of CAS
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shenyang Institute of Computing Technology of CAS filed Critical Shenyang Institute of Computing Technology of CAS
Priority to CN201010582161.8A priority Critical patent/CN102571686B/en
Publication of CN102571686A publication Critical patent/CN102571686A/en
Application granted granted Critical
Publication of CN102571686B publication Critical patent/CN102571686B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Abstract

The invention relates to a method for realizing a cloud conference system, comprising the following steps of: establishing the cloud conference system; connecting various types of clients with the cloud conference system from a cloud access point; and entering an assigned conference room to attend a conference, wherein the cloud conference system comprises a resource service layer, a conference service layer, an application service layer and a backup service layer; each layer comprises a server cluster formed by a great amount of ordinary PCs (Personnel Computer); the quantity of servers in each layer is increased or decreased according to requests; nodes are logic units which are classified according to functions; and each server comprises one node or more than one nodes with different functions, which are progresses or function modules on the server. The method disclosed by the invention has very strong flexibility and expandability, so that the conference scale can be greatly expanded, the cost is reduced, the processing efficiency is improved, and the process delay is reduced. A great scale of the server clusters which can be distributed around the world are utilized to reduce the network delay between a user and the servers. The requirements for conference scenes and instantaneity can be met.

Description

The implementation method of cloud conference system
Technical field
The present invention relates to computer network and the communications field, specifically a kind of implementation method of the service-oriented cloud conference system based on cloud computing.
Background technology
The IP communication technology, particularly VoIP (Voice over Internet Protocol, the technology that the digitlization of simulated sound signal the form with data packet are transmitted on network) are widely used in people's routine work life.Along with the progress of network technology and the growth of bandwidth, the scope of IP communication constantly expands, its form comprises the traditional communication such as phone, fax, mail, note, audio/video multimedia communication, and the data communication such as instant message, electronic whiteboard, desktop are shared, application is shared, webpage is shared.
Video conference, as a kind of application model of IP communication, has been broken the tradition that the personnel that converge just can hold a meeting, and has saved a large amount of time, money and energy, has improved operating efficiency.Its essence is the Unified Communication mode that VoIP communication is end to end evolved into multiparty conference, has realized the application model that transmits in real time live image, voice and application data information at multi-point.
Existing video conference mainly can be divided into centralized and distributed two kinds.The feature of centralized conference is Single-Server, and existing corresponding standard, product and patent, be mainly used in the small-scale meeting in local area network (LAN) and wide area network at present.The centralized conference structure of Single-Server is comparatively simple, strategy that can be based on certain or keeper's real-time selection operate meeting and the response user request controlled, its shortcoming is that autgmentability is poor, the disposal ability of individual server and bandwidth cause in practical application can people with a part in a conference's number very little.The feature of distributed meeting is that multiserver is distributed in different location, there is no at present standard, but existing procucts and patent, be mainly used in wide area network in extensive conference service.The distributed meeting of multiserver mainly contains tree type, net type and P2P (Peer-to-Peer; point-to-point) etc. structure distribution pattern; shortcoming is to postpone greatlyr, arranges that hardware facility costs dearly, participant number more than centralized conference but also cannot hold ultra-large meeting.
Meanwhile, cloud computing concept is being suggested in recent years, and its technology develops rapidly.Cloud computing has the features such as scale is very big, disposal ability is strong, extensibility is strong, cost is cheap, therefore can utilize its these advantages to solve the problem that existing video conference exists.Realize product without any video conference application cloud computing at present, also do not have standard and patent to occur, and MapReduce (mapping and reduction) model in existing cloud computing platform also cannot meet the requirement of real-time of video conference.
Summary of the invention
For the video conference small scale existing in prior art, postpone large, the weak point such as cost dearly, the technical problem to be solved in the present invention is to provide and a kind ofly uses service-oriented thought, utilizes the peculiar advantage of cloud computing to solve the implementation method of the cloud conference system of existing video conference problem.
For solving the problems of the technologies described above, the technical solution used in the present invention is:
The implementation method of a kind of cloud conference system of the present invention comprises the following steps:
Set up cloud conference system, comprise resource service layer, conference service layer, application service layer and backup services layer; Every layer comprises the cluster of servers being made up of a large amount of common PC, and every stratum server quantity increases as required or reduces; Node is the logic unit of dividing by function, and the node that every station server comprises one or more difference in functionalitys, is process or the functional module on server.
Various types of clients are connected to cloud conference system from cloud access point, and described cloud access point is the overall merit mechanism that adopts route selection algorithm, load balancing algorithm, the best Conference server that the user who obtains can connect;
User enters and specifies cloud meeting room to participate in a conference.
Described resource service layer is managed all scheduling of resource situations of all nodes in other layer of service on the whole, comprises three parts, is respectively name manager, scheduling person and network monitoring person; Wherein, all metadata of name administrator administers file system; Scheduling person is responsible for the Mission Operations of dispatching management whole system, is the core that system task is distributed; Network monitoring person is responsible for the distribution of grid resource, connect monitoring, network scheduling etc., and with heartbeat message periodically with each server communication, instruction definite state.
Described conference service layer comprises one or more meeting nodes, managerial meeting, conference service layer carries out Signalling exchange by Session Initiation Protocol and participant member, specify member's behavior, preservation state information, is the focus of meeting, and media information is separated, be transferred to application service layer processing, Media Stream is directly mutual between the interdependent node of member and application service layer.
Described application service layer comprises multimedia type node, application type node and type of webpage node, processes the application task of various meetings, adds at any time the application node of new kind.
Described backup services layer has been contained whole service process, for other each service layers back up.
Described participating in a conference comprises that establishment conference flow and user add two processes of conference flow, wherein creates conference flow to be:
First user enters cloud meeting web interface and logs in, and after checking, enters system;
If have the authority that creates meeting, create meeting template;
If create the success of meeting template, meeting node according to requested session to resource service layer application resource;
If the success of application resource, creates meeting room.
If the failure of application resource, returns to meeting node and again applies for resource according to requested session to resource service layer application resource step; If or created the failure of meeting template, would return to establishment meeting template step and re-create meeting template.
Described establishment meeting template comprises:
Which template type inquiry cloud conference system allows to use under current authority;
In the template type that allows use, revise template parameter;
Template after amendment template parameter is submitted to cloud conference system, complete the establishment of meeting template.
Described user adds conference flow to be:
User enters cloud meeting web interface and logs in, and after checking, enters system;
User accepts to add meeting after invitation;
Carry out media negotiation, the rating of listening to of accepting Media Stream and carrying out meeting;
If need to make a speech in meeting, to host's request floor authority;
If allowed, and existing resource is enough, can obtain competence of speech;
If meeting adjourned, exit meeting room.
If allowed, but existing resource is inadequate, system application new resources;
If there is no new resources, user can not obtain competence of speech.
Described request floor authority comprises permission displaying video, audio frequency, operational applications.
Described user adds conference flow to be:
User enters cloud meeting web interface and logs in, and after checking, enters system;
User inquires about meeting room list, therefrom selects meeting room;
If open session chamber directly enters.
Described user adds conference flow to be:
If not open session chamber, by entering after password authentification.
Or, if not open session chamber, sending the request that enters, host enters after allowing.
The data storage of cloud conference system is as follows:
The data of cloud conference system are divided into meeting material data and minutes data, and wherein meeting material data are stored in the hard disk of cloud conference system, and writing mode is for appending operation; Real time data fault tolerant storage strategy for minutes data acquisition, comprising:
Add the media negotiation stage of conference process user, adopt the overall merit mechanism of route selection algorithm, media quality evaluation model and load balancing algorithm, connect the user to most suitable media server in cloud, and determine backup media server in addition;
Above-mentioned media server receives user's real-time data memory in the buffering area of this media server, is saved in this media server hard disk and backup media server hard disk simultaneously, obtains multiple backups;
In the time there is mistake in data, preferentially from this media server hard disk reading out data;
If this media server hardware error, connects the user to backup with on media server.
Cloud conference system adopts cloud computing distributed computing platform, comprises three steps:
Preprocessing process, after receiving Media Stream, to data amplify, denoising;
Media processing process, carries out mixed processing to pretreated media data, obtains by the audio frequency and video output effect of different sub-venues, different user requirement;
Post-processed and distribution procedure, remove source sound by the audio frequency and video after media processing, and send back to user.
In three steps, want requirement of real time, media data transmission process between adjacent two steps adopts instantaneous transmission, the intermediate object program directly preprocessing process being obtained is transferred to media processing process, media processing Period Process carries out the operation of audio frequency and video hybrid algorithm on request, and the transmitting procedure between media processing process and post-processed and distribution procedure as above.
In three steps, each step adopts fault-tolerant processing, that is:
In resource service layer, scheduling person checks each working node periodically by network monitoring person;
If scheduling person or network monitoring person itself were lost efficacy, from the backup of resource service layer, to recover, all the other tasks do not affect.
If certain preprocessing process task lost efficacy, failure node is restarted, meanwhile, the task transfers of preprocessing process is arrived to backup media server, process the data of this backup, and again carry out media negotiation with user, connection is transferred on this backup media server;
From Backup Data, take out after the data that need, start to process, re-establish simultaneously and carry out TCP with the server of media processing process and be connected;
In preprocessing process, obtain after intermediate object program, on the one hand intermediate object program is sent to media processing process, on the one hand intermediate object program is preserved a period of time in internal memory, if media processing process task lost efficacy, failure node is restarted, and open new media processing process, and to set up TCP with preprocessing process place server and be connected, preprocessing process sends to this new media processing process to process the intermediate object program of preserving;
If for same Media Stream, preprocessing process and media processing process lost efficacy simultaneously, and preprocessing process and media processing process re-execute;
The processing of post-processed and distribution procedure is identical with media processing process, restarts work and connect on new node.
In three steps, each step adopts fault-tolerant processing, if there is " stagnant the latter " in three step implementations, the slow node of individual work efficiency does not catch up with the speed of other node, has caused the reduction of whole efficiency, evaluates by progress appraisement system;
When a task run more than the stipulated time, and its evaluation score is during lower than a threshold value of average mark, it is stagnant the latter for mark;
Former task is proceeded, and starts a same task at other node simultaneously;
Get the result of first completing of task as the input data of next step.
Described evaluation score is:
The bag quantity of receiving in the bag quantity/unit interval of finishing dealing with in the completion rate=unit interval in the unit interval.
The present invention has following beneficial effect and advantage:
1. adopt Service-Oriented Architecture Based (Service-Oriented Architecture, SOA) thought, service is separated, have stronger flexibility and autgmentability, more adapt to upgrading and changes in demand; Utilize the greatly enlarged meeting scale of feature of the scale super large of cloud computing; Utilize the parallel mechanism of cloud computing, use ordinary PC just can complete a large amount of evaluation works, reduce costs, improve treatment effeciency, reduce the delay of processing procedure; Utilization can be distributed in the server zone of the great scale in scope all over the world, can reduce the network delay between user and server; The application of backup services makes video conference have stronger robustness; The storage policy and the distributed computing platform that improve cloud computing platform, make it meet conference scenarios demand and requirement of real-time.
Brief description of the drawings
Fig. 1 is the applied topological diagram of the inventive method;
Fig. 2 is system hlerarchy schematic diagram of the present invention;
Fig. 3 is the establishment conference flow figure of the inventive method;
Fig. 4 is the establishment meeting template sub-process figure of the inventive method;
Fig. 5 be the inventive method add conference flow figure.
Embodiment
The present invention is directed to existing video conference small scale, postpone large, the shortcoming such as cost dearly, provide a kind of and use service-oriented thought, utilize the peculiar advantage of cloud computing to solve cloud conference system and its implementation of existing video conference problem.
The inventive method comprises the following steps:
Set up cloud conference system, comprise resource service layer, conference service layer, application service layer and backup services layer; Every layer comprises the cluster of servers being made up of a large amount of common PC, and every stratum server quantity increases as required or reduces; Node is the logic unit of dividing by function, and the node that every station server comprises one or more difference in functionalitys, is process or the functional module on server.
Various types of clients are connected to cloud conference system from cloud access point, and described cloud access point is the overall merit mechanism that adopts route selection algorithm, load balancing algorithm, the best Conference server that the user who obtains can connect;
User enters and specifies cloud meeting room to participate in a conference.
The cloud conference system that the present invention is made up of a large amount of servers, as shown in Figure 1, the autgmentability of cloud is extremely strong, can and distribute from a corporate lan size variation to All Around The World scope according to the quantity of server, various types of clients comprise that the terminals such as computer, mobile phone, phone all can be connected on cloud from nearest cloud access point, then enter particular conference room and participate in a conference.
As shown in Figure 2, the structure of cloud conference system uses service-oriented thought to be divided into four layers of service, comprises resource service, conference service, application service and backup services.Wherein cloud computing has greatly feature of scale, and every layer of service all can comprise the cluster of servers being made up of a large amount of common PC.Cloud computing has very strong extensibility, and the number of servers in every layer of service can increase minimizing as required.Node is the logic unit of dividing by function, and every station server can comprise the node of one or more difference in functionalitys, and it can be process or functional module on server.
Resource service is the core of whole system, grasps the scheduling of resource of whole cloud conference system, comprises three parts, is respectively name manager, scheduling person and network monitoring person.Wherein, the metadata that name administrator administers file system is all, comprises name space, access control information, and file is to the map information of piece, and the position at the current place of piece.Scheduling person is responsible for the Mission Operations of dispatching management whole system, is the core that system task is distributed.Network monitoring person is responsible for the distribution of grid resource, connects monitoring, network scheduling etc., and with heartbeat message periodically with each server communication, indicate and determine state to them.Generally speaking resource service is all scheduling of resource situations of managing on the whole all nodes in other layer of service.
Conference service comprises one or more meeting nodes, major function is managerial meeting, use Session Initiation Protocol (Session Initiation Protocol, Session initiation Protocol, the signaling control protocol of application layer) and participant member carry out Signalling exchange, specify member's behavior, preservation state information, is the focus of meeting.Meanwhile, the authority of conference member, media behavior is also subject to the control and management of conference service.Under the control of resource service, conference service is by SDP (the Session DescriptionProtocol of Session Initiation Protocol, Session Description Protocol) media description, media information is separated, being transferred to application service processes, afterwards, Media Stream is directly mutual between the interdependent node of member and application service.
Application service comprises multimedia type node, application type node and type of webpage node, and its function is to process the application task of various meetings, and autgmentability is strong, can add at any time the application node of new kind, and the node of every type can have one or more.Multimedia reception, forwarding, mixing, broadcasting, the records etc. such as multimedia type node processing audio frequency, video; The use of application type node processing blank, share desktop, sharing application; The various web application of type of webpage node processing and management of webpage interface etc.
Backup services has contained whole service process, and its prerequisite is, in the time of the popularization of cloud, number of servers is extremely many, and therefore, it is unexpected that the inefficacy of node or server is no longer considered to, but be counted as normal phenomenon.Single master server strategy in existing cloud computing platform, realize easy and be convenient to manage, but its backup that can only ensure file and task recovers, and the weakness of master server itself does not solve, so the state information in management server internal memory is regularly saved to local hard drive and backups to backup server, and periodically establish checkpoint, once master server lost efficacy, by moving from up-to-date checkpoint, therefore there is stronger robustness than original system on backup server.
The flow process that participates in a conference of the present invention mainly comprises that establishment conference flow and user add two processes of conference flow.
Wherein, create conference flow as shown in Figure 3:
First user enters cloud meeting web interface and logs in, and after checking, enters system;
If have the authority that creates meeting, create meeting template;
If create the failure of meeting template, return to establishment meeting template step and re-create meeting template;
If create the success of meeting template, meeting node, according to requested session to resource service layer application resource, comprises meeting node resource, media node resource, application node resource etc.;
If the failure of application resource, returns to application resource step and again applies for resource;
If the success of application resource, creates meeting room.
The sub-process of the establishment meeting template in above-mentioned establishment conference flow process as shown in Figure 4, comprising:
Query template, which template type inquiry cloud conference system allows to use under current authority;
Select a template at the template type that allows to use;
Amendment template parameter, media and application type, the maximum speech number using as meeting duration, maximum number of user, regulation meeting, needs the personnel participating in the meeting that invites etc.;
Template after amendment parameter is submitted to cloud conference system, complete the establishment of meeting template.
As shown in Figure 5, user taking accept the passive mode of invitation add conference flow as:
User enters cloud meeting web interface and logs in, and after checking, enters system;
User accepts to add meeting after invitation;
Carry out media negotiation, the rating of listening to of accepting Media Stream and carrying out meeting;
If need to make a speech in meeting, to host's request floor authority;
If allowed, and existing resource is enough, can obtain competence of speech;
If meeting adjourned, exit meeting room.
If allowed, but existing resource is inadequate, system application new resources;
If there is no new resources, user can not obtain competence of speech.
Described request floor authority comprises permission displaying video, audio frequency, operational applications.
Or user taking the mode of participating in a conference initiatively add conference flow as:
User enters cloud meeting web interface and logs in, and after checking, enters system;
User inquires about meeting room list, therefrom selects meeting room;
If open session chamber directly enters.
If not open session chamber, by entering after password authentification;
Or, if not open session chamber, sending the request that enters, host enters after allowing; Afterwards step with passive add identical.
The data storage of described cloud conference system is as follows:
Data in cloud conference system can be divided into two kinds of forms, and one is meeting material, comprise video, picture, e-book etc., leave data center in, allow participant download as FTP; One is minutes, comprises the information that records in conference process, daily record, recording, video recording etc.
The data of meeting material and existing cloud storage policy (GFS, Google File System, extendible distributed file system, for application large-scale, distributed, that mass data is conducted interviews) conditional likelihood: file rapid development, capacity is very large, can be recorded in piece, main writing mode is for appending operation, and the file that therefore uses GFS mode to complete minutes reads the download of uploading with meeting material.
Minutes are real time datas, comprise active conference information, member's state information, audio stream, video flowing etc., their main feature is time changing, successively receive, need to process in real time and life span short.Existing video conference to the processing mode of these real time datas be receive after, they are saved in internal memory, then immediately process, do not consider the Fault-Tolerant Problems of server failure.
The present invention has increased fault-tolerant content for real-time data memory strategy, and implementation method is as follows:
Add the media negotiation stage of conference process user, adopt the overall merit mechanism of route selection algorithm, media quality evaluation model and load balancing algorithm, connect the user to most suitable media server in cloud, and determine backup media server in addition;
Because the Internet has unsteadiness, with reference to the method that used buffering before playing of existing media player, before media are synthetic, set up buffering area, eliminate the jitter problem that network brings;
Meanwhile, in the time of delay bringing in buffering, the media bag receiving is also not processed, makes full use of during this period of time, and by its preservation and backup, its object has two, and the one, improve robustness, the 2nd, prevent buffer overflow and packet loss; Backup have three parts, portion is saved in local hard drive, other two parts under the gerentocratic management of name, from the appropriate media server analyzing before, select two preserve copies;
Data decimation priority in recovery process is, first from local hard drive reading out data, secondly reads from other server, again, if this locality is hardware error, switches on the server that user is connected to backup;
Because server continuously obtains Media Stream from a large amount of clients, data volume expands very fast, and IMS conference medium is compared with online video display, its content integrity requires lower slightly, but requirement of real-time is higher, therefore, set their life span, abandon the long media bag of delay that exceedes this time limit;
Now, in the file of preservation, overtime media can be deleted or be covered by new data packets, only need to check unmarked data when from fault recovery.
Described cloud conference system adopts cloud computing distributed computing platform, comprises three steps:
Preprocessing process, after receiving Media Stream, to data amplify, denoising;
Media processing process, carries out mixed processing to pretreated media data, obtains by the audio frequency and video output effect of different sub-venues, different user requirement;
Post-processed and distribution procedure, remove source sound by the audio frequency and video after media processing, and send back to user.
The present invention is as follows to the improvement of cloud computing distributed computing platform (MapReduce, mapping and stipulations):
In existing cloud computing platform, MapReduce prototype is mainly tackled the extensive file that is kept on cloud storage system or the batch processing task of database, object is to complete as early as possible whole task, and in cloud conference system of the present invention, need to process the periodic task of long-time continuous, data input and output are all in network, and object is real-time and the fault-tolerance of guarantee task.
In order to adapt to the demand of cloud meeting, the present invention is improved to Map-Reduce-Dispatch (mapping, reduction and distribution) three step frameworks MapReduce:
Map is preprocessing process, is mainly to receive the media processing procedure such as amplification, denoising afterwards;
Reduce is media processing process, mainly comprises the various different mixed processing (as different sub-venues, synthetic audio-video source difference or the effect difference of different customer requirement) that require;
Dispatch is post-processed and distribution procedure, is mainly to remove source sound in case cause echo etc.
In the time having user to be authorized transmission media by host, increase a new Map task.In order to save the Internet resources in cloud, the position of the Map task of distribution is chosen as the node that is arranged in data model receiving media, or adjacent node.When there being new sub-venue to form, or the audio-video frequency content that user need to be different, as requested the Media Stream in specific several sources is mixed in a certain way, set up a new Reduce task.Usually, a corresponding Map task of Dispatch task, because Dispatch need to remove former sound, to different user output ground result difference.
Strategy in existing MapReduce is the Map stage to obtain inputting data from GFS, and processing formation intermediate object program is kept in local hard drive, all complete Deng the Map stage, obtain intermediate object program by Reduce machine by RPC and start to carry out Reduce process, and final result is outputed to GFS.In order to meet real-time and the continuity requirement of video conference, time of delay must be shortened, therefore, can not wait whole Map process to complete and just enter Reduce process, then Reduce process completes the Dispatch process that enters again (this situation just become whole meeting adjourned, participant just can receive and send the audio frequency and video of returning).Therefore, the present invention changes transmitting procedure as instantaneous transmission into.Because the machine in common cloud computing is in together in local area network (LAN) or express network, be connected to each other and be exceedingly fast and safety and stability, so obtaining setting up a TCP between the Reduce process of result in Map process and hope, the present invention connects, the intermediate object program directly Map being obtained is transferred to Reduce, and Reduce such as does not need to complete at all Map yet, only need periodically to carry out on request the operation of audio frequency and video hybrid algorithm, the transmitting procedure between Reduce and Dispatch is also like this.
The present invention is described as follows cloud computing distributed computing platform fault-tolerant strategy:
Scheduling person sends ping to each working node periodically by network monitoring person, if scheduling person or network monitoring person itself were lost efficacy, from the backup of resource service layer, recovers, and all the other tasks do not affect.If certain Map task lost efficacy, failure node is restarted, and meanwhile, Map task transfers, to the backup media data that receive, is processed to the data of this part backup, and is again carried out media negotiation with user, and connection is changed on this new node.Learn from scheduling person the progress that Map is present, then from Backup Data, take out the data that need and start to process, re-establish with the TCP of Reduce simultaneously and be connected.For the inefficacy of Reduce, after solution of the present invention is the intermediate object program of getting at Map, send it on the one hand Reduce, be kept at a period of time in internal memory on the one hand, so that Reduce fetched calculating after losing efficacy or makeing mistakes again.And for same Media Stream, the situation that the situation that Map and Reduce lost efficacy lost efficacy much smaller than single task role, if lost efficacy simultaneously, can only start to reform from whole flow process simultaneously.The mode of Dispatch is also identical with Reduce, need on new node, restart work and connect.
If there is the situation of " stagnant the latter " in the course of work, the slow node of individual work efficiency does not catch up with the speed of other node, has caused the reduction of whole efficiency.Settling mode is that a set of progress evaluation system is set, when a task run more than one minute, and its mark is during lower than a Δ value of average mark, it is stagnant the latter for mark, at this moment, former task is proceeded, and starts a same task at other node simultaneously, last which task first completes, and just adopts the result of which task.Making evaluation score is the bag quantity of receiving in the bag quantity/unit interval of finishing dealing with in the completion rate=unit interval in the unit interval.Due to for video conference, requirement of real-time is higher than fault-tolerance requirement, thus processing policy be, the bag that cannot on time complete processing is abandoned, instead of extends the processing time, the former can reduce frame coherence, the latter can cause picture to lag behind.For continuity task, on long terms, the former just reduces play quality in some moment, but the latter can accumulate, and causes lagging behind more and more of a specified duration, and therefore the present invention adopts front a kind of strategy.Therefore, if within the t time, if certain task completion rate lower than certain threshold value δ, it is stagnant the latter for mark, and re-establishes this task at new node.

Claims (18)

1. an implementation method for cloud conference system, is characterized in that comprising the following steps:
Set up cloud conference system, comprise resource service layer, conference service layer, application service layer and backup services layer; Every layer comprises the cluster of servers being made up of a large amount of common PC, and every stratum server quantity increases as required or reduces; Node is the logic unit of dividing by function, and the node that every station server comprises one or more difference in functionalitys, is process or the functional module on server;
Various types of clients are connected to cloud conference system from cloud access point, and described cloud access point is the overall merit mechanism that adopts route selection algorithm, load balancing algorithm, the best Conference server that the user who obtains can connect;
User enters and specifies cloud meeting room to participate in a conference;
Described resource service layer is managed all scheduling of resource situations of all nodes in other layer of service on the whole, comprises three parts, is respectively name manager, scheduling person and network monitoring person; Wherein, all metadata of name administrator administers file system; Scheduling person is responsible for the Mission Operations of dispatching management whole system, is the core that system task is distributed; Network monitoring person is responsible for the distribution of grid resource, connect monitoring, network scheduling etc., and with heartbeat message periodically with each server communication, instruction definite state.
2. by the implementation method of cloud conference system claimed in claim 1, it is characterized in that: described conference service layer comprises one or more meeting nodes, the meeting of described meeting node administration, conference service layer carries out Signalling exchange by Session Initiation Protocol and participant member, specifies member's behavior, preservation state information, it is the focus of meeting, and media information is separated, being transferred to application service layer processing, Media Stream is directly mutual between the interdependent node of member and application service layer.
3. by the implementation method of cloud conference system claimed in claim 1, it is characterized in that: described application service layer comprises multimedia type node, application type node and type of webpage node, the application task of the various meetings of above-mentioned node processing, adds the application node of new kind at any time.
4. by the implementation method of cloud conference system claimed in claim 1, it is characterized in that: described backup services layer has been contained whole service process, for other each service layers back up.
5. by the implementation method of cloud conference system claimed in claim 1, it is characterized in that: described in participate in a conference and comprise and create conference flow and user adds two processes of conference flow, wherein create conference flow to be:
First user enters cloud meeting web interface and logs in, and after checking, enters system;
If have the authority that creates meeting, create meeting template;
If create the success of meeting template, meeting node according to requested session to resource service layer application resource;
If the success of application resource, creates meeting room.
6. by the implementation method of cloud conference system claimed in claim 5, it is characterized in that: if the failure of application resource is returned to meeting node and again applied for resource according to requested session to resource service layer application resource step; If or created the failure of meeting template, would return to establishment meeting template step and re-create meeting template.
7. by the implementation method of cloud conference system claimed in claim 5, it is characterized in that: described establishment meeting template comprises:
Which template type inquiry cloud conference system allows to use under current authority;
In the template type that allows use, revise template parameter;
Template after amendment template parameter is submitted to cloud conference system, complete the establishment of meeting template.
8. by the implementation method of cloud conference system claimed in claim 5, it is characterized in that: described user adds conference flow to be:
User enters cloud meeting web interface and logs in, and after checking, enters system;
User accepts to add meeting after invitation;
Carry out media negotiation, receiving media flows to the rating of listening to of guild's view;
If need to make a speech in meeting, to host's request floor authority;
If allowed, and existing resource is enough, can obtain competence of speech;
If meeting adjourned, exit meeting room.
9. by the implementation method of cloud conference system claimed in claim 8, it is characterized in that:
If allowed, but existing resource is inadequate, system application new resources;
If there is no new resources, user can not obtain competence of speech.
10. by the implementation method of cloud conference system claimed in claim 8, it is characterized in that: described request floor authority comprises permission displaying video, audio frequency and operational applications.
11. by the implementation method of cloud conference system claimed in claim 5, it is characterized in that: described user adds conference flow to be:
User enters cloud meeting web interface and logs in, and after checking, enters system;
User inquires about meeting room list, therefrom selects meeting room;
If open session chamber directly enters.
12. by the implementation method of cloud conference system claimed in claim 5, it is characterized in that: described user adds conference flow to be:
If not open session chamber, by entering after password authentification;
Or, if not open session chamber, sending the request that enters, host enters after allowing.
13. by the implementation method of cloud conference system claimed in claim 1, it is characterized in that: the data storage of cloud conference system is as follows:
The data of cloud conference system are divided into meeting material data and minutes data, and wherein meeting material data are stored in the hard disk of cloud conference system, and writing mode is for appending operation; Real time data fault tolerant storage strategy for minutes data acquisition, comprising:
Add the media negotiation stage of conference process user, adopt the overall merit mechanism of route selection algorithm, media quality evaluation model and load balancing algorithm, connect the user to most suitable media server in cloud, and determine backup media server in addition;
Above-mentioned media server receives user's real-time data memory in the buffering area of this media server, is saved in this media server hard disk and backup media server hard disk simultaneously, obtains multiple backups;
In the time there is mistake in data, preferentially from this media server hard disk reading out data;
If this media server hardware error, connects the user to backup with on media server.
14. by the implementation method of cloud conference system claimed in claim 1, it is characterized in that: cloud conference system adopts cloud computing distributed computing platform, comprises three steps:
Preprocessing process, after receiving Media Stream, to data amplify, denoising;
Media processing process, carries out mixed processing to pretreated media data, obtains by the audio frequency and video output effect of different sub-venues, different user requirement;
Post-processed and distribution procedure, remove source sound by the audio frequency and video after media processing, and send back to user.
15. by the implementation method of the cloud conference system described in claim 14, it is characterized in that: in three steps, want requirement of real time, media data transmission process between adjacent two steps adopts instantaneous transmission, the intermediate object program directly preprocessing process being obtained is transferred to media processing process, media processing Period Process carries out the operation of audio frequency and video hybrid algorithm on request, and the transmitting procedure between media processing process and post-processed and distribution procedure is as above-mentioned instantaneous transmission.
16. by the implementation method of the cloud conference system described in claim 14, it is characterized in that: in three steps, each step adopts fault-tolerant processing, that is:
In resource service layer, scheduling person checks each working node periodically by network monitoring person;
If scheduling person or network monitoring person itself were lost efficacy, from the backup of resource service layer, to recover, all the other tasks do not affect;
If certain preprocessing process task lost efficacy, failure node is restarted, meanwhile, the task transfers of preprocessing process is arrived to backup media server, process the data of this backup, and again carry out media negotiation with user, connection is transferred on this backup media server;
From Backup Data, take out after the data that need, start to process, re-establish simultaneously and carry out TCP with the server of media processing process and be connected;
In preprocessing process, obtain after intermediate object program, on the one hand intermediate object program is sent to media processing process, on the one hand intermediate object program is preserved a period of time in internal memory, if media processing process task lost efficacy, failure node is restarted, and open new media processing process, and to set up TCP with preprocessing process place server and be connected, preprocessing process sends to this new media processing process to process the intermediate object program of preserving;
If for same Media Stream, preprocessing process and media processing process lost efficacy simultaneously, and preprocessing process and media processing process re-execute;
The processing of post-processed and distribution procedure is identical with media processing process, restarts work and connect on new node.
17. by the implementation method of the cloud conference system described in claim 14, it is characterized in that: in three steps, each step adopts fault-tolerant processing, if occur in three step implementations " stagnant the latter ", be the speed that the slow node of individual work efficiency does not catch up with other node, cause the reduction of whole efficiency, evaluated by progress appraisement system;
When a task run more than the stipulated time, and its evaluation score is during lower than a threshold value of average mark, it is stagnant the latter for mark;
Former task is proceeded, and starts a same task at other node simultaneously;
Get the result of first completing of task as the input data of next step.
18. by the implementation method of the cloud conference system described in claim 17, it is characterized in that: described evaluation score is:
The bag quantity of receiving in the bag quantity/unit interval of finishing dealing with in the completion rate=unit interval in the unit interval.
CN201010582161.8A 2010-12-09 2010-12-09 Method for realizing cloud conference system Active CN102571686B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201010582161.8A CN102571686B (en) 2010-12-09 2010-12-09 Method for realizing cloud conference system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201010582161.8A CN102571686B (en) 2010-12-09 2010-12-09 Method for realizing cloud conference system

Publications (2)

Publication Number Publication Date
CN102571686A CN102571686A (en) 2012-07-11
CN102571686B true CN102571686B (en) 2014-09-03

Family

ID=46416175

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201010582161.8A Active CN102571686B (en) 2010-12-09 2010-12-09 Method for realizing cloud conference system

Country Status (1)

Country Link
CN (1) CN102571686B (en)

Families Citing this family (32)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10735486B2 (en) 2012-12-28 2020-08-04 Qualcomm Incorporated Device timing adjustments and methods for supporting dash over broadcast
CN104284132A (en) * 2013-07-05 2015-01-14 中兴通讯股份有限公司 Video communication method and device
CN103338348B (en) * 2013-07-17 2016-08-10 天脉聚源(北京)传媒科技有限公司 Implementation method, system and the server of a kind of network audio-video meeting
CN103581300B (en) * 2013-10-10 2018-06-19 国家电网公司 The interactive online security method of conference system information material
US10261834B2 (en) * 2013-12-18 2019-04-16 Telefonaktiebolaget Lm Ericsson (Publ) Method and network node for selecting a media processing unit based on a media service handling parameter value
CN105281924A (en) * 2014-06-30 2016-01-27 中兴通讯股份有限公司 Conference creating method and system, and conference server
CN104301126A (en) * 2014-09-25 2015-01-21 深圳市京华科讯科技有限公司 Synergetic virtual meeting room system and implementation method thereof
WO2016086373A1 (en) * 2014-12-03 2016-06-09 深圳市融资城网络服务中心有限公司 Protocol signing method and server
CN104579713A (en) * 2015-01-16 2015-04-29 四川联友电讯技术有限公司 Method for adding conference participants to conference by host for synchronous voice conference system
CN104579714A (en) * 2015-01-16 2015-04-29 四川联友电讯技术有限公司 Conference initiating method for fragmentation asynchronous conference system
CN104579712A (en) * 2015-01-16 2015-04-29 四川联友电讯技术有限公司 Method for conference member to issue character information in fragmentation asynchronous conference system
US9948786B2 (en) * 2015-04-17 2018-04-17 Cisco Technology, Inc. Handling conferences using highly-distributed agents
CN106303360A (en) * 2015-06-10 2017-01-04 中兴通讯股份有限公司 Terminal meeting load-balancing method, Apparatus and system
CN107295290A (en) * 2016-04-05 2017-10-24 中兴通讯股份有限公司 A kind of recorded broadcast control method and device
CN107276772A (en) * 2016-04-06 2017-10-20 中兴通讯股份有限公司 The application process and payment services plateform system of a kind of cloud conference system
CN107295054A (en) * 2016-04-12 2017-10-24 中兴通讯股份有限公司 The management method and device of cloud conferencing resource
CN105721178A (en) * 2016-04-25 2016-06-29 四川联友电讯技术有限公司 Pushing method of conference messages of fragmented asynchronous conference system
CN107547822A (en) * 2016-06-23 2018-01-05 南京中兴软件有限责任公司 Video session method and device
CN106230608A (en) * 2016-07-21 2016-12-14 山东共达信息技术有限公司 A kind of cloud regards sound intelligent interactive system
CN106507031A (en) * 2016-11-23 2017-03-15 中国联合网络通信集团有限公司 A kind of update method of meeting room template and device
CN107302679A (en) * 2017-06-13 2017-10-27 成都全云科技有限公司 A kind of video conference accesses access system
CN107508790A (en) * 2017-07-05 2017-12-22 融音网络技术(上海)有限公司 The conference system and its cut-in method for supporting heterogeneous terminals to be accessed by cloud platform
CN107508689A (en) * 2017-08-24 2017-12-22 太仓安顺财务服务有限公司 A kind of conference flow integrated management approach
CN108566531A (en) * 2017-12-29 2018-09-21 广州雅广信息科技有限公司 A kind of meeting management system and conference management method
CN108063832B (en) * 2018-02-06 2021-04-30 上海地面通信息网络股份有限公司 Cloud storage system and storage method thereof
CN108364083A (en) * 2018-02-11 2018-08-03 中企动力科技股份有限公司 Conference reservation method and system Internet-based
CN108199858A (en) * 2018-03-22 2018-06-22 无锡城市职业技术学院 A kind of distributed meeting management system and system generation method based on radio-frequency card
CN108810449A (en) * 2018-07-06 2018-11-13 合肥明高软件技术有限公司 A kind of long-distance cloud video service system and method for servicing applied to party-building in base
CN110955602B (en) * 2018-11-30 2023-07-07 中国航空工业集团公司沈阳飞机设计研究所 Distributed embedded software testing system based on resource sharing
CN110971864B (en) * 2019-11-29 2021-05-11 北京航天星桥科技股份有限公司 Quick conference holding system in cloud video conference
CN114024694B (en) * 2020-07-16 2023-07-04 深圳瞬玩科技有限公司 Remote conference method and system based on cloud mobile phone
CN114296850A (en) * 2021-12-24 2022-04-08 中船重工(武汉)凌久高科有限公司 Method for realizing customized conference desktop special terminal based on virtual cloud

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101635635A (en) * 2009-08-25 2010-01-27 北京原力创新科技有限公司 Cloud mode streaming media service platform

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101635635A (en) * 2009-08-25 2010-01-27 北京原力创新科技有限公司 Cloud mode streaming media service platform

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
Junchao Li 等.Study on Service-Oriented Cloud Conferencing.《Computer Science and Information Technology (ICCSIT), 2010 3rd IEEE International Conference》.2010,
Pedro Rodriguez 等.VaaS: Videoconference as a Service.《Collaborative Computing: Networking, Applications and Worksharing, 2009》.2009,
Study on Service-Oriented Cloud Conferencing;Junchao Li 等;《Computer Science and Information Technology (ICCSIT), 2010 3rd IEEE International Conference》;20100711;第21-25页,图4-6 *
VaaS: Videoconference as a Service;Pedro Rodriguez 等;《Collaborative Computing: Networking, Applications and Worksharing, 2009》;20091114;全文 *

Also Published As

Publication number Publication date
CN102571686A (en) 2012-07-11

Similar Documents

Publication Publication Date Title
CN102571686B (en) Method for realizing cloud conference system
WO2021213184A1 (en) Distributed election-based end-to-end content distribution network system and distribution method
CN108197329A (en) It is synchronized by the Real-time document demonstration data of generic service
CN104247380B (en) CRUD type agreements are bound in distributed agreement protocol
US20050234943A1 (en) Method, system, and apparatus for enabling near real time collaboration on an electronic document through a plurality of computer systems
CN108280080A (en) A kind of method of data synchronization, device and electronic equipment
CN101163117B (en) Packet management method, packet resource sharing method and instant communication equipment
JP2022534179A (en) DATA TRANSMISSION METHOD, DEVICE, COMPUTER PROGRAM AND COMPUTER DEVICE
CN104092746B (en) A kind of O&M auditing system and the network load balancing method in its system
CN105610947A (en) Method, device and system for realizing high-available distributed queue service
CN111083177B (en) Cross-domain collaborative interaction method based on collaborative gateway
CN110196843A (en) A kind of document distribution method and container cluster based on container cluster
CN114328432A (en) Big data federal learning processing method and system
Xhafa et al. Supporting scenario-based online learning with P2P group-based systems
EP3902279A1 (en) Audio and video stream distribution method in audio and video stream distribution system, and dynamic parent node
WO2022183518A1 (en) Cloud-computing-oriented high-performance blockchain architecture method
CN109347906A (en) A kind of data transmission method, device and server
CN103442034B (en) A kind of stream media service method based on cloud computing technology and system
CN102387062B (en) Method for improving transmission rate of peer-to-peer (p2p) node in crossing network by dynamic bridging contact
US20150172136A1 (en) Apparatus of mapping logical point-of-delivery to physical point-of-delivery based on telecommunication information networking
CN103685497A (en) On-line storing and sharing method and system
CN101645872B (en) Business service system of Internet large-scale application environment and working method thereof
CN114205233A (en) Intelligent contract self-adaptive configuration and execution framework oriented to data management and control
CN108737480A (en) Conference cascading method, server based on w media server and storage medium
US11258834B2 (en) System and method for recording online collaboration

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