Summary of the invention
Technical problem to be solved by this invention is to provide a kind of Large Copacity perception cloud computing platform system, can process a large number of users Concurrency Access, ensures the stability of platform.
The technical solution adopted for the present invention to solve the technical problems is: provide a kind of Large Copacity perception cloud computing platform system, comprise the cloud computing cluster management layer from bottom to top distributed, basic cloud computing platform service layer and cloud service race, described cloud computing cluster management layer is used for carrying out unified monitoring and management to the hardware resource in cloud computing cluster and software resource, for described basic cloud computing platform service layer provides the base layer support of resource management and scheduling; Described basic cloud computing platform service layer is used for providing the service of mass data distributed storage, Distributed Parallel Computing service, distributed application integration service, the service of configurationization soft plug-in management, distributed flow scheduling and management service and vehicle intelligent terminal information gathering to serve; Described cloud service race is used for realizing specific service function.
Described cloud computing cluster management layer comprises cluster system hardware state monitoring section and cluster system cloud computing platform service management section, and described cluster system hardware state monitoring section is used for carrying out unified monitoring and management to the hardware resource in cloud computing cluster; Described cluster system cloud computing platform service management section is used for carrying out unified monitoring and management to the software service resource in cloud computing cluster.
The service of described mass data distributed storage by Namespace virtualization technology, load-balancing technique, data Large-scale Migration and fault recovery technology and data trnascription administrative mechanism, for cloud platform user the Organization And Management of file system is provided, the management of distributed storage Virtual File System, storage node network state show, data block storage and load-balancing function.
Described Distributed Parallel Computing service is carried out abstract to the resource in cluster virtual computation environmental, adopt and execute the task map reduction algorithm, extract and perform parallel computational model, realizing the parallel processing of distributed data in cluster.
Described distributed application integration service is communicated for platform user provides by route, addressing and the communication technology, the function of service interaction, Services Integration and mapping and protocol conversion.
The service of described configurationization soft plug-in management carries out unified management to soft insert receptacle, build one support Dynamic expansion, software environment when dynamically changing the operation of system action.
Described distributed flow scheduling and management service, by the automatic flow of realization flow and scheduling, for platform user provides analysis to mission bit stream, processing and processing capacity, and formulate the function that corresponding scheduling strategy completes the scheduling controlling to task.
Described vehicle intelligent terminal information gathering service provides based on the monitor data model management function of vehicle intelligent terminal, devices communicating management function and device drives management function for platform user.
It is integrated that Large Copacity perception cloud computing platform system adopts between Heterogeneous Data Integration Middleware implementation platform system, described Heterogeneous Data Integration Middleware is the succession middleware mutually accessed based on supporting database and the XML metadata of XML standard design, is mainly divided into configure base layer and integrated data switching layer; Described configure base layer is the configuration component supporting middleware, and generating configuration file, as the medium of exchanges data, realizes the input and output function of data respectively; Described integrated data switching layer is for realizing from heterogeneous data source to XML file and by the data transaction work of XML format file to target database.
Beneficial effect
Owing to have employed above-mentioned technical scheme, the present invention compared with prior art, there is following advantage and good effect: the present invention can process a large number of users Concurrency Access, ensure the stability of platform, make user obtain the Consumer's Experience identical with when using low capacity service platform.
Embodiment
Below in conjunction with specific embodiment, set forth the present invention further.Should be understood that these embodiments are only not used in for illustration of the present invention to limit the scope of the invention.In addition should be understood that those skilled in the art can make various changes or modifications the present invention, and these equivalent form of values fall within the application's appended claims limited range equally after the content of having read the present invention's instruction.
Embodiments of the present invention relate to a kind of Large Copacity perception cloud computing platform system, as shown in Figure 1, comprise the cloud computing cluster management layer from bottom to top distributed, basic cloud computing platform service layer and cloud service race, described cloud computing cluster management layer is used for carrying out unified monitoring and management to the hardware resource in cloud computing cluster and software resource, for described basic cloud computing platform service layer provides the base layer support of resource management and scheduling; Described basic cloud computing platform service layer is used for providing the service of mass data distributed storage, Distributed Parallel Computing service, distributed application integration service, the service of configurationization soft plug-in management, distributed flow scheduling and management service and vehicle intelligent terminal information gathering to serve; Described cloud service race is used for realizing specific service function.Wherein, cloud computing cluster management layer and basic cloud computing platform service layer exist as the core of whole Large Copacity perception cloud computing platform system.
Cloud computing cluster management layer comprises cluster system hardware state monitoring section and cluster system cloud computing platform service management section.Platform carries out unified monitoring and management by these two aspects to the hardware resource in cloud computing cluster and software service resource, based on cloud computing platform service the base layer support of resource management and scheduling is provided.
Described basic cloud computing platform service layer is used for providing the service of mass data distributed storage, Distributed Parallel Computing service, distributed application integration service, the service of configurationization soft plug-in management, distributed flow scheduling and management service and vehicle intelligent terminal information gathering to serve.
The service of mass data distributed storage by Namespace virtualization technology, load-balancing technique, data Large-scale Migration and fault recovery technology, data trnascription administrative mechanism etc., for cloud platform user the Organization And Management of file system is provided, the function such as the management of distributed storage Virtual File System, storage node network state show, data block storage and load balancing.
Mass data storage and administrative skill can provide a stable shared storage and analytical system, realize the storage and management to mass data, each back end has two backup nodes, when this back end goes wrong, can according to the strategy use backup machine pre-set, thus improve the reliability of data, ensure the service quality of platform.
Wherein, mass data distributed memory system, comprises large data memory module and small data memory module, and the large data that described large data memory module is uploaded for vehicle intelligent terminal carry out store and management; The small data that described small data memory module is uploaded for vehicle intelligent terminal carries out store and management.This distributed memory system provides the storage and management service of mass data, is the basis that cloud computing system is run.
Large data memory module is the module that the large data uploaded for vehicle intelligent terminal carry out store and management, and as the key sound video data etc. after haulage vehicle warning, its data volume is comparatively large, then can store by this memory module.
This module is primarily of index data node and back end composition, and the metadata store of generic-document is at index data node, and data then disperse to be stored on different back end.In this module, each number, according to there being its copy data, ensures the consistency between them by distributed information log system between master data and copy data.Only have master data to provide access at ordinary times, when after the node failure of master data place, access automatically switches on copy data.
The small data uploaded for vehicle intelligent terminal during small data memory module carries out the module of store and management, and as positional information and the vehicle sensory information of haulage vehicle, its data are usually less, then can store by this memory module.
This module adopts storage virtualization technology that organic combination is carried out in storage hardware, storing software and storage administration, defines one and supports that virtualized extendible dawn cloud stores, and provide unified management platform, simplify system manager's work.Pass through storage management software, adopt Intel Virtualization Technology that numerous memory node is formed storage pool, can support more than 10,000 virtual volume, these virtual volumes can be distributed to different users and use, user can share these hardware, but cannot access the data of other users, and the data of each user are mutually independently.
Described mass data distributed memory system also comprises calamity for memory module, and the daily critical data of vehicle that described calamity is used for storing for memory module carries out disaster-tolerant backup.When disaster occurs, calamity can be recovered the daily critical data of vehicle with Backup Data for memory module.
Distributed Parallel Computing service is carried out abstract to the resource in cluster virtual computation environmental, adopt executing tasks parallelly mapping/reduction algorithm, extract and perform parallel computational model, realize the parallel processing function of distributed data in cluster, for platform user provides parallel computation, reasonable layout data, calls dispatching algorithm to improve computational efficiency, to ensure to export the function of result of calculation certainty and support atomicity task.
High performance distributed computing technology can manage large-scale high performance distributed computing cluster, thus Concurrency Access and the inquiry request of a large number of users can be processed, the mistake of computing node can be tolerated simultaneously, make the inefficacy of part of nodes can not the normal operation of influential system, therefore can ensure the service quality of platform.
Distributed application integration is served by route, addressing, the communication technology etc., for platform user provides communication, service interaction, Services Integration and the function such as mapping, protocol conversion.
The soft plug-in management service of configurationization is intended to solve the durability of platform software, componentization and plug and play.By the soft plug-in management service of configurationization, platform carries out unified management to soft insert receptacle, build one support Dynamic expansion, software environment when dynamically changing the operation of system action, for the software function in platform or service provide the support of pluggableization, simultaneously for platform user provides the dynamic installation of plug-in unit, startup, stopping, renewal and offloading functions.
The soft insert receptacle of configurationization is the important component part of basic cloud platform, be mainly used in that Dynamic expansion supported by structure one, software environment when dynamically changing the operation of system action, for software function component (or plug-in unit) basic in system provide plug-in unit, the insert receptacle of micro-kernel mechanism, it is made up of following 4 modules:
The mode developed plug-in unit, organize plug-in unit, dispose soft plug-in unit in the soft plug-in definition module definition soft insert receptacle of configurationization.The soft insert receptacle of configurationization realizes dynamic soft plug-in management mechanism by the class load mechanism of the bag shared mechanism between the metadata definition mode of soft plug-in unit, soft plug-in unit and soft plug-in unit.
Life cycle management modular structure, on the basis of the soft insert receptacle of configurationization, makes the soft plug-in management service of configurationization to manage the life cycle of plug-in unit dynamically.The function of life cycle management module is the definition to soft plug-in unit life cycle on the one hand, and the life cycle state of soft plug-in unit comprises: install, resolve, start, activate, stop and unloaded; Meanwhile, the life cycle of life cycle management module in charge to soft plug-in unit is changed.
The function of soft plug-in unit communication management module is that to set up with service be the soft plug-in unit communication mechanism of core.Communication between soft plug-in unit and sharing functionality is completed by the bag shared mechanism provided to the registration service of soft plug-in unit communication management module and the soft insert receptacle of configurationization.
For other subsystems and component, the application in the soft insert receptacle of configurationization provides the infrastructure of disposing and managing to soft plug-in security administration module, ensures that the soft insert receptacle of configurationization can control the soft plug-in unit assigned in wherein.
Distributed flow scheduling and management service, by the automatic flow of realization flow and scheduling, for platform user provides analysis to mission bit stream, processing and processing capacity, formulate the function that corresponding scheduling strategy completes the scheduling controlling to task simultaneously.Distributed flow scheduling and management provide flow scheduling engine management function, flow monitoring management function, procedural model management function.
For the related service of carrier of CNPC and group of CNPC can be supported better, cloud computing support platform provides distributed flow scheduling and ADMINISTRATION SUBSYSTEM, by abstract for the links in logistics transportation processing procedure and emergency scheduling be workflow, and carry out dispatching and manage to the implementation of workflow.This subsystem is made up of following 3 modules:
Flow definition module carries out modeling by Process Modeling Language to service link, describes the various actions of flow process.The Process Modeling Language that this module is now supported is BPMN2.0.In flow definition process, user is by Process Modeling Language definitim flow, and product process definition configuration file, for other module.
Flow engine module is the nucleus module of flow performing, and it is with flow definition configuration file for input, and defined by process of analysis, Establishing process example also runs.In running, flow engine module can regularly by the intermediateness of flow process stored in database, be convenient to user on the one hand by management tool monitoring flow process practice condition, prevent the loss of flow process intermediate data after system crash on the other hand.
The flow performing state of workflow management tool model by preserving in acquisition database, for user provides visual management and monitoring function.User sees the current each flow state (comprising flow process progress, the Output rusults etc. of each activity of flow process), the current all flow definitions that are running by management tool module, and distributes to the task of each user.
Vehicle intelligent terminal information gathering service provides based on the monitor data model management function of vehicle intelligent terminal, devices communicating management function and device drives management function for platform user.Terminal equipment is realized to the functions such as unified control and centralized watch.
In order to be undertaken integrated by existing different vehicle operation platform, Heterogeneous Data Integration Middleware technology can be adopted, solve the difference of the data structure, interface mode etc. existed between different system platform.For supporting the distributed data translative mode of different vehicle platform, the succession middleware that supporting database and XML metadata access mutually based on XML standard design, it is mainly divided into configure base layer and integrated data switching layer two large divisions.
A) configure base layer
For supporting the configuration component of middleware, generating configuration file, as the medium of exchanges data, realizes the input and output function of data respectively.Therefore this partial function realizes the mapping definition of isomeric data source data to XML format, and XML format defines to the conversion map of target database.Major function comprises data format mapping definition, data encryption/decryption, data compression/decompression contracting etc., forms the flow process of configuration file as shown in Figure 2.
B) integrated data switching layer
Realize from heterogeneous data source to XML file and by the data transaction work of XML format file to target database.Obtain data from the heterogeneous data source of different platform and adopt WebServices technology to realize to the function of target database write data.The Data Integration Middleware realized under this framework, effectively can connect the application system of different platform, and each system is source data or destination data source each other, achieves the data integration between heterogeneous platform, and system architecture as shown in Figure 3.