CN103559104B - A kind of distributed redundancy real-time database framework based on mixed communication - Google Patents
A kind of distributed redundancy real-time database framework based on mixed communication Download PDFInfo
- Publication number
- CN103559104B CN103559104B CN201310549119.XA CN201310549119A CN103559104B CN 103559104 B CN103559104 B CN 103559104B CN 201310549119 A CN201310549119 A CN 201310549119A CN 103559104 B CN103559104 B CN 103559104B
- Authority
- CN
- China
- Prior art keywords
- real
- time database
- time
- database
- main
- 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
Links
Landscapes
- Hardware Redundancy (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
A kind of distributed redundancy real-time database framework based on mixed communication of disclosure, this framework is in rail traffic monitoring system, data scale is huge, there is inherent semantic association in monitoring data, work station is based on features such as real time data inquiries, distribution in redundancy real-time database, the each side such as synchronization and failure tolerant have been respectively adopted various applicable strategy, unicast communication and cast communication real-time Transmission data of different nature respectively are used by mixing, the problems such as the synchronous communication data volume that solution brings due to scale enlargement increases rapidly, by adopting rational strategy to meet the temporal consistency requirement that real-time database is synchronized by monitoring data inherence semantic association, effectively achieve the failure tolerant under distributed many redundancies real-time database framework, more effective achieve many redundancies of distributed real-time database in monitoring system, the reliabilty and availability of real-time database compared to conventional dual redundant framework more sound assurance, stable operation for rail traffic monitoring system lays the foundation。
Description
Technical field
This technological invention relates to track traffic automatic field, particularly relates to distributed monitoring system field, and this technological invention is widely portable to electric railway and urban track traffic each specialty application such as monitoring system and comprehensive monitoring system。
Background technology
Along with developing rapidly of computer science and automatic technology, every field in track traffic, the such as field such as electric railway and urban track traffic, monitoring system is from traditional small table system to modern distributed complex system development, the territorial scope of its control, flow process scale etc. all increases sharply。And due to conventionally employed discrete monitoring system model, namely there is signal, PSCADA(power monitoring), BAS(ring control) etc. multiple independent monitoring systems be respectively arranged with assistant director and be operated, there is operation cost height, the information of multiple systems is difficult to problems such as sharing, and the comprehensive monitoring system integrating each subsystem function is increasingly becoming new development trend。In comprehensive monitoring system, subsystems is by integrated or interconnection, it is achieved is monitored by same Upper system, so that information can be effectively shared, and reduces workload, cut operating costs。
Distributed redundancy real-time database: no matter being discrete monitoring system or comprehensive monitoring system, real-time data base is all the data core of monitoring system, real-time data base, often again referred to as real-time database, wherein houses the various real time datas that rail transportation operation is on-the-spot。The monitoring system related data by gathering or I/O module obtains up-to-date presence state data real-time update real-time database from equipment or subsystem；HMI(man machine interface) module obtains real time data from real-time database, is presented to operator in the way of hommization；Alarm module is reported to the police accordingly according to real-time status data-triggered in real-time database；Trend module timing obtains latest data from real-time database and presents with curve form；And operator is according to various practical situations, is sent instruction by real-time database timely, control in real time。Real-time database realizes the unified management of real time data as the real time data maincenter of monitoring system and other module for system provides consistent real-time data access service, so that monitoring system is as the continual and steady reliability service of an organic whole, correctly realize the control function of rail transportation operation。
Due to various application in track transportation, it is no matter electric railway or subway or light rail, it is directed to extensive personal safety, therefore monitoring system has strict reliability requirement, and as the real-time database of system real time data core, being that whole system is rely the basis of reliability service, its reliability is undoubtedly the most important thing。Redundancy is to ensure the effective technology of reliability。Therefore, the monitoring system of track traffic adopts redundancy generally in every respect, thus the reliabilty and availability of safeguards system as much as possible。Such as, in conventional configuration, in the LAN of monitoring system, will not only run a station server, but run two redundant servers standby each other simultaneously, and configure some work stations, realize redundancy by reasonably communication, as shown in Figure 1。Equally, redundancy is also the effective technology ensureing real-time database reliability。In conventional configuration, on dual redundant server in LAN together with time run identical real-time database, guarantee when the real-time database on a station server delays the reason inefficacies such as machine because of server, real-time database on another station server still can provide normal real-time data access service for system, thus realizing the dual redundant of real-time database, to ensure the reliabilty and availability of real-time database。What this meant that real-time database in monitoring system adopts is distributed redundant shelf。Owing to real-time database housing the last state of real time data, and be used for providing real-time data access service, therefore, all real-time database in distributed redundant shelf must flow through rational real-time Communication for Power and keep data realtime uniform, such guarantee is all realtime uniform from the real time data numerical value that arbitrary real-time database accesses, this is called synchronization, otherwise is called step-out。
Recently, some major accidents of track traffic industry, such as Wenzhou rear end collision of motor train accident, and some have a strong impact on the large scale disasters of rail transportation operation, such as Wenchuan earthquake etc., the safe and reliable of application in track transportation is more paid close attention to by people, thus the reliabilty and availability of rail traffic monitoring system be it is also proposed higher requirement。This makes, and monitoring system is more crucial at some, need more reliable occasion to have employed many redundant shelf than conventional dual redundant with higher reliability。Such as in the Dispatch and Command Center of the high ferro maincenters such as Wuhan, monitoring system often configures the server that multiple stage is standby each other, as in figure 2 it is shown, so that it is guaranteed that when breaking down more than a station server system still can correctly run；And in a lot of cities, particularly being located in the city of disaster prone district such as earthquake zone, subway monitoring system, except being configured with dual redundant server and some work stations in operation control centre, has Disaster Preparation Center simultaneously outside certain distance, and identical configuration is provided, as shown in Figure 3。Maintain data consistent by communicating at ordinary times, once operation control centre cannot use because there is large scale disasters, be then transferred to Disaster Preparation Center and maintain control function。In the configuration of many redundancies, system exists a master server and the standby server of multiple stage, run identical real-time database at every station server simultaneously, and by rational real-time Communication for Power maintenance real-time synchronization。Even if thus ensureing the faults such as machine that have real-time database on multiple servers because server is delayed and losing efficacy, as long as it is effective for also having the real-time database at least one station server, just can provide real-time data access service for system, thus realizing the failure tolerant of real-time database to a greater extent, the more reliability of sound assurance monitoring system。
Summary of the invention
For problem existing in prior art, the present invention proposes a kind of distributed redundancy real-time database framework suitable in rail traffic monitoring system, this framework is in rail traffic monitoring system, data scale is huge, there is inherent semantic association in monitoring data, work station is based on features such as real time data inquiries, distribution in redundancy real-time database, the each side such as synchronization and failure tolerant have been respectively adopted various applicable strategy, unicast communication and cast communication real-time Transmission data of different nature respectively are used by mixing, the problems such as the synchronous communication data volume that solution brings due to scale enlargement increases rapidly, by adopting rational strategy to meet the requirement that real-time database is synchronized by monitoring data inherence semantic association, and effectively achieve the failure tolerant under many redundant distributions formula real-time database framework in a suitable manner, many redundancies of real-time database in monitoring system are achieved thus more effective, the reliabilty and availability of real-time database compared to conventional dual redundant more sound assurance, for system stability, reliable continuous service provides more solid foundation。
The technical scheme is that a kind of distributed redundancy real-time database framework based on mixed communication, including: real-time data access module；Distributed many redundancies real-time database framework；Redundancy real-time database synchronization module；The failure tolerant module of distributed many redundancies real-time database。
Described real-time data access module: in monitoring system, real-time database saves the latest data of the on-the-spot status information of rail transportation operation, and constantly real-time update is to ensure the present situation that reflection is on-the-spot in time, thus other module offer real-time data access service being system as real time data maincenter；Real-time data access is divided into queried access and updates access；Obtaining the up-to-date numerical value of real time data when described queried access refers in storehouse, to meet application demand, queried access will not change the numerical value of real time data in real-time database；Update access to refer to according to application demand, update the numerical value of real time data in real-time database, update access and can change the numerical value of the real time data in real-time database。
Described distributed many redundancies real-time database module: in order to ensure the reliability of real-time database in monitoring system, generally runs identical real-time database, to realize the redundancy of real-time database on the server of all redundancies simultaneously；Work station then can be divided into two ways: one is not run real-time database on work station, and all of real-time data access, all passing through to communicate is provided remote real-time data to access service by the real-time database on server；Another kind is then run local real-time database on a workstation。Distributed redundancy real-time database framework in literary composition have employed the second way, namely runs real-time database on a workstation, so directly can provide local service to these queried accesses, without nationality by the offer remote service that communicates。This avoids the heavy network traffic load thus brought on the one hand, the more important thing is, owing to entirely without communication delay, the local real-time performance providing queried access service to be substantially better than remote service, such that it is able to more timely, smooth the present situation to operator's displaying scene。
Described redundancy real-time database synchronization module: the real-time database in monitoring system saves the real time data of the on-the-spot current state of all operations in internal memory, and all real-time database in distributed redundant shelf must maintain real-time synchronization all the time。Here mainly considering when renewal accesses the data that have modified in real-time database, in redundant shelf, how the distributed real-time database of original synchronization maintains synchronization, including the workflow updating access and redundancy real-time database updating access, standby real-time database of main real-time database。
The failure tolerant module of described distributed many redundancies real-time database: adopt the mode of many redundancies by can the reliability of more sound assurance real-time database than dual redundant, even if ensure that in framework and losing efficacy more than the real-time database of a station server, as long as also having the real-time database on the server no less than properly functioning, just remain able to provide real-time data access service into system, thus more effectively realize the failure tolerant of real-time database than dual redundant real-time database framework, possesses higher reliability。
Accompanying drawing explanation
Fig. 1 is dual redundant server configures schematic diagram of the prior art。
Fig. 2 is many redundant servers configuration schematic diagram of the prior art。
Fig. 3 is many redundant servers configuration schematic diagram in prior art with Disaster Preparation Center。
Fig. 4 is distributed many redundancies real-time database block schematic illustration of the present invention。
Fig. 5 is the synchronization policy schematic diagram of the renewal access request of the main real-time database of the present invention。
Fig. 6 is the present invention synchronization policy schematic diagram for the renewal access request of real-time database。
Fig. 7 is that the present invention is for real-time database operational process schematic diagram。
Detailed description of the invention
Below in conjunction with relevant drawings and specific embodiment, the present invention is further elaborated。
1) real-time data access module: in monitoring system, the real-time data access of real-time database totally can be divided into two classes: a class is queried access unit, the up-to-date numerical value of real time data is obtained from real-time database, to meet application demand, such as: HMI module inquires about latest data from real-time database, and Dynamic Announce is to operator；Alarm module inquires about current data from real-time database, then reports to the police as abnormal；Trend module is periodic queries data from real-time database, and present with curve form。Another kind of is update to access unit, according to application demand, update the numerical value of real time data in real-time database, such as: gather or I/O module obtains from equipment or subsystem and the related data real-time database can be carried out real-time update after the last state data that operation is on-the-spot, and update its acquisition time simultaneously, and the control instruction that operator makes, usually also first write real-time database, and notify control module execution via real-time database。
2) distributed many redundancies real-time database module: in order to ensure the reliability of real-time database in monitoring system, generally runs identical real-time database, to realize the redundancy of real-time database on the server of all redundancies simultaneously。Work station then can be divided into two ways, and one is not run real-time database on work station, and all of real-time data access, all passing through to communicate is provided remote real-time data to access service by the real-time database on server；Another kind is then run local real-time database on a workstation。Owing to work station is mainly used in providing interface and operator mutual, therefore such as HMI module, alarm interface module, trend module etc. are required for frequent visit real-time database and obtain last state numerical value and refresh interface so that the situation that operation is on-the-spot is presented to operator in real time, these mainly depend on queried access, only when necessary, when operator makes suitable control operation, can relate to updating and access, visible, in the real-time data access of work station, most for queried access, thus the distributed redundancy real-time database framework in the present embodiment have employed the second way, namely real-time database is run on a workstation, so directly can provide local service to these queried accesses, without nationality by the offer remote service that communicates, this avoids the heavy network traffic load thus brought on the one hand, the more important thing is, due to entirely without communication delay, the local real-time performance providing queried access service to be substantially better than remote service, such that it is able to it is more timely, smooth the present situation to operator's displaying scene。Certainly, in order to realize this point, applicable strategy must be adopted to make the real-time database of work station also keep real-time synchronization with the real-time database of server, and use rational communication mode to avoid the problem that the synchronous communication data volume caused due to the synchronous communication of work station real-time database increases severely。
As shown in Figure 4, Fig. 4 is the schematic diagram of distributed many redundancies real-time database module。In distributed many redundancies real-time database module of the present embodiment, master server, all of standby server and work station all run real-time database, and keep real-time synchronization by the real-time Communication for Power mode being suitable for, for the ease of describing, respectively the real-time database on active and standby server and work station is called main real-time database, standby real-time database and client's real-time database, uses R successivelya, RsAnd RcRepresent；Real-time database on each station server has oneself unique priority, uses Pt, t=1,2,3 ... representing, t is the numbering of the real-time database on server；Wherein, dotted line represents cast communication, in all of real-time database, connect including setting up cast communication between main real-time database, whole standby real-time database and whole client's real-time database to realize cast communication, otherwise varied, main real-time database and each standby real-time database both connected transmission data by multicast, also connected from multicast and received data, and all of client's real-time database only connects from multicast and receives data, do not connect to multicast and send data；Solid line represents that unicast communication, all of standby real-time database and all of client's real-time database are set up independent unicast communication respectively and connected with main real-time database, to realize respective unicast communication。
3) redundancy real-time database synchronization module: the real-time database in monitoring system saves the real time data of the on-the-spot current state of all operations in internal memory, is herein referred to as the current state of real-time database, uses Si, i=1,2,3 ... representing, all real-time database in distributed redundant shelf must maintain real-time synchronization all the time, does not revise the real time data in real-time database due to queried access, and thus without causing real-time database Status Change, only local real-time database performs just can complete。Here main consideration is when updating the data that access have modified in real-time database, and in redundant shelf, how the original distributed real-time database synchronized maintains synchronization。
The renewal of main real-time database accesses: as it is shown in figure 5, the synchronization policy schematic diagram of the renewal access request that Fig. 5 is main real-time database。Assuming that in redundant shelf, in all real-time database, real time data is the last state data that operation is on-the-spot, real-time database has been carried out synchronizing, and state is SiThen the data acquisition module on master server obtains new status data, and submit to update access request D, to update related data in real-time database to up-to-date, synchronize to maintain all real-time database in distributed redundant shelf, adopt such a strategy here, first performed to update access request D by main real-time database, self related data is updated, thus the state of main real-time database is from SiIt is changed to Si+1；Then by synchronizing information Mi=< Ra,Pt,i+1,Di> gives all of standby real-time database and client's real-time database, wherein R by communications distributionaRepresent that role is main real-time database, PtFor the priority of current main real-time database, i+1 is main real-time database last state numbering, DiFor updating access request, standby real-time database and client's real-time database are after receiving synchronizing information, by performing renewal access request DiAlso from state SiIt is altered to Si+1, thus maintaining and the synchronization of main real-time database。Real-time database on all of server, namely main real-time database and all standby real-time database, preserve the last k performed individual more newly requested, i.e. Di-kTo Di。
Synchronous communication is the Primary communication between the real-time database in distributed redundant shelf。Owing to application in track transportation Regional Distribution is extensive, the real-time database data volume of monitoring system is bigger, particularly comprehensive monitoring system integrated and interconnected all multi-specialized, the data volume of its real-time database is the vertical monitoring system of remote oversubscription especially, accordingly, the amount of communication data that the synchronous communication of the real-time database of rail traffic monitoring system relates to is bigger；Owing to distributed many redundancies real-time database framework existing multiple standby real-time database, and all of work station all exists client's real-time database, if employing unicast communication, so each is required for a synchronous communication data for real-time database and client's real-time database, this can cause that synchronous communication data increase with the real-time database scale in distributed redundant shelf and increase rapidly, thus being easy to exhaust communication bandwidth, the transmission of significant data in serious delay self and other communication, this allows for communication bandwidth becomes the severe constraints of real-time database Expansion；It is same data in fact due to what synchronous communication was sent to standby real-time database and client's real-time database, therefore, sharp increase problem in order to avoid synchronous communication data, substituting unicast communication used here as cast communication, namely main real-time database uses cast communication to send synchronizing information to all of standby real-time database and client's real-time database simultaneously；Owing to the traffic of multicast will not increase rapidly along with the increase of data receiver's quantity, thus avoiding communication bandwidth to the constraint of real-time database quantity in framework, real-time database quantity according to the increase freely of the demand of application, can be provided with good extensibility；Owing to main real-time database sends synchronizing information to all of standby real-time database and client's real-time database by cast communication simultaneously, therefore, if a standby real-time database or client's real-time database fail and receive or correctly process synchronizing information, just cannot require that main real-time database sends this synchronizing information again by cast communication, in order to avoid other real-time database repeats to receive and process, for this this real-time database by reacquiring, for labelling from main real-time database, the synchronous communication information preserved with information encoding separately through unicast communication and processing, thus maintaining and the synchronization of main real-time database。
Renewal for real-time database accesses: as shown in Figure 6, and Fig. 6 is the synchronization policy schematic diagram of the renewal access request of standby real-time database。Renewal access request on a standby server whether can adopt with main server class like pattern, first by local real-time database, namely corresponding standby real-time database performs, and is then distributed to main real-time database, and other performs for real-time database and all of client's real-time database to maintain synchronization？This cannot in the monitoring system of track traffic。Because do so is likely to result in redundant shelf in each real-time database, the update sequence of data is inconsistent, and the data in monitoring system exist inherent semantic association, and therefore this is unallowed。For the schematic example of power supply profession in an application in track transportation; the current break data of electric supply installation produce one at master server and update access request; and the trip signal of the protection device of correspondence is engraved in a standby server when much the same and produces a renewal access request; if the active and standby real-time database of correspondence is all taked first to locally execute the strategy then distributed; so due to the existence of communication delay; two more newly requested is distributed to the other side when performing after local real-time database has performed, and more newly requested the locally executing of the other side completes before this。Therefore, although in final the two real-time database, data mode is consistent, but the information of main real-time database reflection is the tripping operation after current break causes, and that is for the situation contrast of real-time database reflection, is current break occur after tripping operation。Due to communication complexity, other for real-time database and client's real-time database are probably in two kinds of situations any one。Both of these case is different types of fault for the operator monitoring system, it is necessary to adopt different counter-measures, and confirmation of responsibility is also different。Therefore, this is unacceptable in the monitoring system of application in track transportation。In order to avoid this problem, the renewal access request for standby real-time database have employed such a synchronization policy here。It is accessed for standby real-time database and the renewal access request obtained is transmitted to the execution of main real-time database by unicast communication, then perform result and return also by unicast communication。This role in fact having acted as one " setter " for real-time database has passed the requests to main real-time database, itself does not locally execute, until when synchronizing information is distributed to all standby real-time database and client's real-time database by cast communication after performing by main real-time database, this just performs to maintain synchronization for real-time database。This just forces all of renewal access request to must first be in main real-time database order and performs then to distribute successively, so that it is guaranteed that the update sequence of the data in all real-time database in redundant shelf is on all four。
Do so will not dramatically increase the communication pressure of monitoring system, this is because main real-time database distributes the traffic of synchronizing information mainly in main real-time database to the direction of standby real-time database to standby real-time database, and standby real-time database updates in the traffic baseset of access request in the opposite direction to the transmission of main real-time database, owing to the communication port in monitoring system is usually full duplex, obvious communication pressure therefore so can't be caused to increase。One advantage of this synchronization policy is that standby real-time database delivers renewal access request by communication is transparent, so that the real-time data access on standby server completes just as on master server, thus relatively uniform for the module accessing real-time database can be distributed in master server and multiple standby server operation。For the large scale system that this function of rail traffic monitoring system is complicated, so can effectively equally loaded, for instance data acquisition module is in master server operation；At a standby server, data maintenance module operates to ensure that the CPU etc. of abundance calculates resource；And alarm module and interlocking module run at another standby server, thus reply in time will not be affected because of resource-constrained when emergency occurs。
The process strategy of the renewal access request of client's real-time database is consistent with standby real-time database。。
The workflow of redundancy real-time database: as it is shown in fig. 7, Fig. 7 is standby real-time database operational process schematic diagram。Real-time database continues to provide the basis that real-time data access service is the monitoring system stability reliability service of application in track transportation, in order to realize the redundancy of real-time database on multiple servers, real-time database on each station server initially sets up multicast and connects when bringing into operation, by monitoring whether multicast communication check has had main real-time database running, without then self running as main real-time database, if existed, no matter whether its priority is higher than self, oneself all runs as standby real-time database, this is to maintain stablizing of the synchronous communication set up between real-time database in distributed redundant shelf。When real-time database on server is run as standby real-time database, initially set up and the clean culture of main real-time database connects, and obtain all status datas of main real-time database by unicast communication thus realizing and the synchronization of main real-time database, then the synchronizing information of main real-time database distribution is being continuously carried out, while maintaining synchronization, to provide for local various real-time data access and to service。As it was previously stated, queried access is locally executing, and update access and transparent pass to main real-time database。
Although it should be noted that standby real-time database is also the same with main real-time database to preserve k last renewal access request, but can't be distributed by cast communication after having performed each renewal access request, in order to avoid other real-time database repeats to receive and perform。Time properly functioning, main real-time database and all of standby real-time database all can periodically through cast communication Distribution status information M '=R, Pt, i+1, D ', to other real-time database, wherein R represents role, is Ra in the status information of main real-time database, represent that role is main real-time database, the status information of standby real-time database is Rs, represent that role is standby real-time database, t numbers for this real-time database, Pt is the priority of this real-time database, and i+1 is this real-time database last state numbering, and D ' is status information, thus other real-time database in distribution of notifications formula redundant shelf, this real-time database is still properly functioning。The flow process of client's real-time database is similar with standby real-time database, but it should be noted that client's real-time database only connects from multicast and receive data, and do not connected by multicast and send any data, neither distribute synchronizing information, also not Distribution status information, and client's real-time database does not preserve any more newly requested yet。
4) the failure tolerant module of distributed many redundancies real-time database: the failure tolerant module of distributed redundancy real-time database is one of key of rail traffic monitoring system redundancy framework, as the real-time database of real time data maincenter of monitoring system once not be provided that real-time data access service, whole system will sink into paralysis。And adopt the mode of many redundancies by can the reliability of more sound assurance real-time database than dual redundant, in distributed many redundancies real-time database framework, master server and all standby servers run respective active and standby real-time database simultaneously and keep real-time synchronization, after main real-time database lost efficacy, a conversion in multiple standby real-time database substitutes original main real-time database as main real-time database and runs, and all the other keep for real-time database and new main real-time database synchronizes；If after a period of time, new main real-time database because failure and other reasons lost efficacy, is then run as main real-time database by a replacement in other properly functioning standby real-time database again again, and standby real-time database and up-to-date main real-time database synchronize。Lost efficacy more than the real-time database of a station server even if this ensures that thering in framework, as long as also having the real-time database on the server no less than properly functioning, just remain able to provide real-time data access service into system, thus more effectively realize the failure tolerant of real-time database than dual redundant real-time database framework, possesses higher reliability。Owing to, in redundant shelf, client's real-time database keeps Tong Bu with main real-time database, standby real-time database lost efficacy the impact on redundant shelf not quite, here the main failure tolerant that main real-time database is discussed。
After the main real-time database in distributed redundant shelf lost efficacy, to take over rapidly for one in multiple standby real-time database, continue to provide Real-time Data Service as main real-time database, and remaining to realize and to maintain for real-time database and all of client's real-time database and the synchronization of new main real-time database。Owing in rail traffic monitoring system, the structure of communication system is complicated, fault is various, therefore, when former main real-time database lost efficacy, its last synchronizing information, it is likely to have been distributed to all of standby real-time database and client's real-time database, it is likely to and has been distributed only to a part therein, or do not come and distribute at all, so after a standby real-time database is run as new main real-time database, remaining is probably synchronization for real-time database and client's real-time database with new main real-time database, it is also likely to be step-out, in order to ensure to synchronize, remaining all status data that must flow through the unicast communication new main real-time database of acquisition for real-time database and all of client's real-time database realizes synchronizing again, but, one problem of do so is, owing to the data volume of the real-time database of rail traffic monitoring system is very big, obtain whole status datas in real-time database and can cause very big traffic load, and all standby real-time database and client's real-time database obtain whole status data by unicast communication and extremely heavy offered load will be brought to impact the communication network of the system of monitoring in the same time period in failure tolerant process simultaneously, thus seriously incuring loss through delay the transmission of significant data in self and other communication, even cause communication blocking。
In order to avoid this situation, it is employed herein such a strategy, time properly functioning, active and standby real-time database in distributed redundancy real-time database framework can pass through cast communication Distribution status information, may thereby determine that the role of active and standby real-time database currently properly functioning in framework and correspondence thereof and priority。If main real-time database is delayed due to master server, the reason such as machine lost efficacy, the standby real-time database that so distributed redundant shelf medium priority is the highest will be promoted as main real-time database, but role before changing, the last k that this real-time database is preserved to the distribution of all real-time database of other in framework by cast communication is individual more newly requested, by synchronizing information Mj=< Rs,Pt,j+1,Dj>, wherein j=i '-k-1, i '-k, ..., i '-1 is distributed to other for real-time database and all client's real-time database by cast communication, and wherein i' is that this is converted to the end-state numbering before main real-time database for real-time database, the original state numbering namely run as main real-time database；RsRepresent that role is about to be transferred to main real-time database by standby real-time database；PtFor this priority for real-time database；DjFor the renewal access request in synchronizing information；J+1 performs after this access request this for the status number of real-time database；Sent preservation more newly requested after, this is converted into main real-time database for real-time database and brings into operation from the state being numbered i', after other receives these synchronizing informations for real-time database and all of client's real-time database from cast communication, oneself state numbering after former main real-time database being lost efficacy compares with it, for other one of them for real-time database and all client's real-time database, such as this real-time database end-state be numbered i ", generally have i '-k≤i ' '≤i '。Now, by multicast reception to being numbered i '-k-1 to i " the synchronous communication message of-1 comprises be this real-time database obtained from former main real-time database and performed more newly requested, directly abandon。Be numbered i " comprise in the synchronous communication message of i '-1 be this real-time database lost when former main real-time database lost efficacy more newly requested; update access request make the state of this real-time database be updated to numbering i' gradually by performing these successively; so that this real-time database and the initial running status of new main real-time database synchronize; then this real-time database and new main real-time database are set up unicast communication and connected, and maintain the synchronization with new main real-time database based on mixed communication。
If being unsatisfactory for i '-k≤i ' '≤i ', illustrate this real-time database due in lucky longer period before fault the cause specific such as communication failure and cannot realize synchronizing with new main real-time database in this way, now can only after setting up clean culture connection with new main real-time database, whole status datas are obtained from new main real-time database by unicast communication, again realize synchronizing, it is then based on mixed communication and maintains synchronization, but this situation seldom occurs。So, the standby real-time database that final former priority is the highest is run as new main real-time database, and remaining connects for the newly-built independent clean culture of real-time database and all of client's real-time database and new main real-time database, realize according to above-mentioned process and maintain and the synchronization of new main real-time database, it is achieved failure tolerant。Adopt this strategy, avoid other as far as possible again to realize synchronizing with new main real-time database and the burdensome communication load that produces and huge time overhead also by obtaining whole status datas when there is no need for real-time database and all client's real-time database, thus reducing in this process as far as possible, a large amount of real-time database obtain the impact of heavy network load in short-term that whole status data communication network to monitoring system brings simultaneously, effectively achieve the failure tolerant of many redundancies real-time database, ensure the reliabilty and availability of real-time database in distributed redundant shelf fully。
Above example is described only for the partial function of the present invention, but embodiment and accompanying drawing are not for limiting the present invention's。Without departing from the spirit and scope of the invention, any equivalence change done or retouching, also belong to the protection domain of the present invention, the content that therefore protection scope of the present invention should define with claims hereof is for standard。
Claims (2)
1. the distributed redundancy real-time database framework based on mixed communication, it is characterised in that including: real-time data access module；Distributed many redundancies real-time database module；Redundancy real-time database synchronization module；The failure tolerant module of distributed many redundancies real-time database；
Described real-time data access module is divided into queried access unit and updates access unit；Described queried access unit refers to the up-to-date numerical value obtaining real time data from real-time database, and to meet application demand, queried access will not change the numerical value of real time data in real-time database；The described access unit that updates refers to according to application demand, updates the numerical value of real time data in real-time database, updates access and can change the numerical value of the real time data in real-time database；
If a standby real-time database or client's real-time database fail and receive or correctly process synchronizing information, just cannot require that main real-time database sends this synchronizing information again by cast communication, in order to avoid other real-time database repeats to receive and process, for this this real-time database by reacquiring, for labelling from main real-time database, the synchronous communication information preserved with information encoding separately through unicast communication and processing, thus maintaining and the synchronization of main real-time database；
Described distributed many redundancies real-time database module refers to ensure the reliability of real-time database in monitoring system, runs identical real-time database, to realize the redundancy of real-time database on the server of all redundancies simultaneously；
In described distributed many redundancies real-time database module, master server, all of standby server and work station all run real-time database, in all of real-time database, connect including setting up cast communication between main real-time database, whole standby real-time database and whole client's real-time database to realize cast communication, all of standby real-time database and all of client's real-time database are set up independent unicast communication respectively and are connected with main real-time database, to realize respective unicast communication；
Described redundancy real-time database synchronization module refers to that all real-time database in distributed redundant shelf must maintain real-time synchronization all the time；In described redundancy real-time database synchronization module, main real-time database uses cast communication to send synchronizing information to all of standby real-time database and client's real-time database simultaneously；
When real-time database on server is run as standby real-time database, initially set up and the clean culture of main real-time database connects, and obtain all status datas of main real-time database by unicast communication thus realizing and the synchronization of main real-time database, then the synchronizing information of main real-time database distribution is being continuously carried out, while maintaining synchronization, to provide for local various real-time data access and to service；
Even if the failure tolerant module of described distributed many redundancies real-time database refers in framework lost efficacy more than the real-time database of a station server, as long as also having the real-time database on the server no less than properly functioning, just remain able to provide real-time data access service into system, thus more effectively realize the failure tolerant of real-time database than dual redundant real-time database framework, possesses higher reliability。
2. the distributed redundancy real-time database framework based on mixed communication according to claim 1, it is characterized in that: in described distributed many redundancies real-time database module, master server, all of standby server and work station all run real-time database, in all of real-time database, connect including setting up cast communication between main real-time database, whole standby real-time database and whole client's real-time database to realize cast communication, all of standby real-time database and all of client's real-time database are set up independent unicast communication respectively and are connected with main real-time database, to realize respective unicast communication；
In described redundancy real-time database synchronization module, it is accessed for standby real-time database and client's real-time database and the renewal access request obtained is transmitted to the execution of main real-time database by unicast communication, then perform result and return also by unicast communication；
In the failure tolerant module of described distributed many redundancies real-time database, time properly functioning, main real-time database, standby real-time database can pass through cast communication Distribution status information, so that it is determined that the role of main real-time database currently properly functioning in framework, standby real-time database and correspondence thereof and priority；If main real-time database due to master server delay machine reason lost efficacy, role is before changing, synchronizing information is distributed the more newly requested of preservation by cast communication to all real-time database of other in framework and all client's real-time database by this main real-time database, the standby real-time database that priority is the highest will be promoted as main real-time database, so that this main real-time database and the initial running status of new main real-time database synchronize, then this main real-time database and new main real-time database set up unicast communication connection, the synchronization with new main real-time database is maintained based on mixed communication, remaining connects for the newly-built independent clean culture of real-time database and all of client's real-time database and new main real-time database, realize and maintain and the synchronization of new main real-time database, realize failure tolerant；
If a standby real-time database or client's real-time database fail and receive or correctly process synchronizing information, just cannot require that main real-time database sends this synchronizing information again by cast communication, in order to avoid other real-time database repeats to receive and process, for this this real-time database by reacquiring, for labelling from main real-time database, the synchronous communication information preserved with information encoding separately through unicast communication and processing, thus maintaining and the synchronization of main real-time database。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310549119.XA CN103559104B (en) | 2013-11-07 | 2013-11-07 | A kind of distributed redundancy real-time database framework based on mixed communication |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310549119.XA CN103559104B (en) | 2013-11-07 | 2013-11-07 | A kind of distributed redundancy real-time database framework based on mixed communication |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103559104A CN103559104A (en) | 2014-02-05 |
| CN103559104B true CN103559104B (en) | 2016-06-22 |
Family
ID=50013356
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310549119.XA Active CN103559104B (en) | 2013-11-07 | 2013-11-07 | A kind of distributed redundancy real-time database framework based on mixed communication |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103559104B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104102741A (en) * | 2014-07-31 | 2014-10-15 | 浪潮通信信息系统有限公司 | Method for scheduling real-time data in monitoring system |
| CN107660281B (en) * | 2015-05-19 | 2021-06-08 | 华为技术有限公司 | System and method for synchronizing distributed computing runtime |
| CN105553819B (en) * | 2015-12-10 | 2019-09-10 | 交控科技股份有限公司 | Rail traffic maintenance support system and method |
| CN106656960B (en) * | 2016-09-28 | 2019-12-17 | 北京辰极国泰科技有限公司 | hilscher-based credible data acquisition system and method |
| CN106528871B (en) * | 2016-12-06 | 2019-07-26 | 爱普（福建）科技有限公司 | A kind of method of online updating industrial control system project data point information |
| CN107071351B (en) * | 2017-03-30 | 2019-11-05 | 杭州瑞网广通信息技术有限公司 | A kind of station multistage disaster tolerance framework and method |
| CN107395406B (en) * | 2017-07-12 | 2020-12-25 | 广州市百果园信息技术有限公司 | Online state data processing method, device and system of online system |
| CN111369794B (en) * | 2020-02-28 | 2022-01-25 | 腾讯科技（深圳）有限公司 | Method, device and equipment for determining traffic participation information and storage medium |
| CN113987079B (en) * | 2021-12-27 | 2022-04-19 | 四川旷谷信息工程有限公司 | Data synchronization method for urban rail transit security system |
| CN114665609B (en) * | 2022-05-20 | 2022-09-02 | 中铁电气化勘测设计研究院有限公司 | Data intercommunication construction method of BIM power supply system comprehensive maintenance management platform |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1780201A (en) * | 2004-11-17 | 2006-05-31 | 中兴通讯股份有限公司 | Method for realizing synchronous information between multiple nodes |
| US20090106357A1 (en) * | 2007-10-17 | 2009-04-23 | Marvin Igelman | Synchronized Media Playback Using Autonomous Clients Over Standard Internet Protocols |
| CN101459618A (en) * | 2009-01-06 | 2009-06-17 | 北京航空航天大学 | Data packet forwarding method and device for virtual machine network |
| CN102254031A (en) * | 2011-08-03 | 2011-11-23 | 无锡浙潮科技有限公司 | Batch processing request-based Microsoft SQL server database cluster |
| CN102945195A (en) * | 2012-11-26 | 2013-02-27 | 国电南瑞科技股份有限公司 | Master-standby redundancy replication method based on SQLite databases |
-
2013
- 2013-11-07 CN CN201310549119.XA patent/CN103559104B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1780201A (en) * | 2004-11-17 | 2006-05-31 | 中兴通讯股份有限公司 | Method for realizing synchronous information between multiple nodes |
| US20090106357A1 (en) * | 2007-10-17 | 2009-04-23 | Marvin Igelman | Synchronized Media Playback Using Autonomous Clients Over Standard Internet Protocols |
| CN101459618A (en) * | 2009-01-06 | 2009-06-17 | 北京航空航天大学 | Data packet forwarding method and device for virtual machine network |
| CN102254031A (en) * | 2011-08-03 | 2011-11-23 | 无锡浙潮科技有限公司 | Batch processing request-based Microsoft SQL server database cluster |
| CN102945195A (en) * | 2012-11-26 | 2013-02-27 | 国电南瑞科技股份有限公司 | Master-standby redundancy replication method based on SQLite databases |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103559104A (en) | 2014-02-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103559104B (en) | A kind of distributed redundancy real-time database framework based on mixed communication | |
| CN109451065B (en) | Soft load balancing and shunting automation system and operation method thereof | |
| CN106790436B (en) | Traffic system monitoring method based on cloud architecture and control center cloud server | |
| CN103199972A (en) | Double machine warm backup switching method and warm backup system achieved based on SOA and RS485 bus | |
| CN112565415B (en) | Cross-region resource management system and method based on cloud edge cooperation | |
| CN105471960A (en) | Information interaction system and method between private clouds and public cloud | |
| CN102664747B (en) | Cloud calculating platform system | |
| CN106100894A (en) | A kind of highly reliable cluster operation management method | |
| WO2012145963A1 (en) | Data management system and method | |
| CN103036719A (en) | Cross-regional service disaster method and device based on main cluster servers | |
| CN101989903A (en) | Dual-machine redundancy by-mouth switching method of comprehensive monitoring pre-communication controller | |
| CN110719282B (en) | Authentication dual-active system based on unified authority | |
| CN102207978A (en) | Database access method and system | |
| CN105354111A (en) | Redundancy backup method suitable for wide-area distributed real-time database | |
| CN108964986A (en) | Cooperative office system application layer dual-active disaster recovery and backup systems | |
| CN107612960A (en) | Integral control system in dispatching of power netwoks | |
| CN104580502A (en) | Method for achieving load balance dual-unit hot standby | |
| CN111541599B (en) | Cluster software system and method based on data bus | |
| CN105302670B (en) | A kind of method and device that station is monitored using multi-computer Redundancy mode | |
| CN202870563U (en) | Distributed comprehensive monitoring system | |
| CN109367586A (en) | A kind of city rail traffic signal system clock system and method | |
| KR101358995B1 (en) | Method and system for managing high availability | |
| CN209258153U (en) | A kind of city rail traffic signal system clock system | |
| CN104468829A (en) | Enterprise IT system structure implementation method based on virtualization environment | |
| CN102843265B (en) | Not resetting without the hot standby dual link method of perturbation of a kind of communication protocol |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |