CN101969391B - Cloud platform supporting fusion network service and operating method thereof - Google Patents

Abstract

The invention relates to a cloud platform supporting fusion network service and an operating method thereof. The cloud platform is provided with a plurality of hardware and software resources, and is connected with three access networks of a telecommunication network, an interconnection network and a broadcasting and television network respectively through three core networks of operators of the three networks and the corresponding communication protocol to allow platform users (including the operators of the three networks or service providers and the like) to rent platform resources according to respective requirements to deploy respective services and operational capabilities. The cloud platform has a layered structure, and is provided with a platform management layer, a service execution layer, a resource virtualization and management layer and a hardware resource layer respectively from top to bottom, and in an interactive mode among the layers, an upper-layer module uses the function provided a lower-layer module in a mode of interface call. The cloud platform supports the platform users to adjust the rent resources dynamically and conveniently according to actual requirements, and simultaneously, provides the access capability on the resources of the three networks and convenience for the development and operation of the fusion network service by the platform users.

Description

A cloud platform supporting converged network services and its working method

technical field

The invention relates to a cloud platform supporting integrated network services and a working method thereof, and belongs to the technical field of integrated network service supporting equipment and systems thereof.

Background technique

With the development of network technology, the three networks of telecommunication network, Internet and broadcasting network are increasingly showing the trend of mutual integration and mutual opening. In a converged network, on the one hand, services need to be converged so as to better utilize various network capabilities to jointly provide more types of new services; on the other hand, more flexible service provision and operation methods are required to quickly and efficiently provide new services business.

At present, when third-party service providers deploy and operate value-added services, they usually deploy and maintain corresponding software and hardware service systems by themselves, which increases the deployment and maintenance costs of service providers; at the same time, it also increases It is difficult for network operators to manage service providers.

Under the background of triple play integration, the operation of network services is more and more favored by telecom operators. However, the traditional telecom application server does not provide other services with access to Internet resources, for example, it does not support Internet protocols, which makes the development and operation of converged network-oriented telecom services very difficult.

The request load of telecommunication services is not stable, but fluctuates up and down according to different times, which is characterized by a large difference between the peak and trough loads; moreover, the load will increase sharply during holidays or emergencies. In order to meet the needs of telecom users, operators often deploy network elements such as servers to meet the maximum load requirements. However, doing so will cause two consequences: when the volume of business requests is at a low ebb, a large number of servers are idle, which wastes a lot of labor costs for hardware and maintenance. However, when the user business volume increases sharply during a special time period, the existing computer room and server capacity may still not be able to meet the maximum load demand, and the user experience is still poor.

Typical products of traditional telecom business support platforms include: Puyuan EOS Northern Telecom Business Support System, Guangdong Telecom Integrated Business Support System and IBM's next-generation mobile value-added service platform. in,

Puyuan EOS platform is an application software platform based on J2EE, which provides complete systematic support from development, debugging, operation, maintenance and management, and has strong secondary development capabilities. Each application system of the EOS-based North Telecom business support system is built on the basis of the application carrying platform, and the core components of the application carrying platform are the EAI platform, core data objects and system management modules. The system adopts the hierarchical structure of presentation layer, business logic layer and data layer.

Guangdong Telecom's integrated business support system includes 13 subsystems including business acceptance, scheduling management, resource management, construction management, customer service management, charging management, interface management, marketing management, product management, customer management, system management, basic data management and statistics management , using a layered system architecture of public information management platform and interface platform layer, basic service layer, business framework layer, and business application layer. The public information management platform defines the public information shared by the IBSS system and other related systems. At present, it mainly provides customer management to meet the needs of each system for customer information; the interface platform provides the system's external data exchange mechanism and handles the data exchange between the IBSS and external systems Protocol; the basic service layer defines system public services, which have nothing to do with business processing, and all systems can handle such basic services, such as database access, name service, public API (such as date/character processing), etc.; business framework layer Define the basic business framework of the system, which is the basis for system business realization, such as authority management, product management, etc.; the business application layer is the business realization layer, which realizes the specific business of IBSS by invoking the services of the business framework, which includes the subsystems of each subsystem accomplish. Each subsystem of the business application layer is independent of each other and accessed through the interface. The business processing sequence and business process requirements between the various subsystems are coordinated and controlled through the business process system.

The smooth evolution solution of IBM's mobile value-added service platform and its key supporting technologies can realize the transformation from a single vertical system with business separation and repeated functions to a horizontal service platform that supports cross-network and unified functions; it provides a system that can be in harmony with legacy systems A business platform integration framework for work, and a set of systems science methods. It can evaluate the capability maturity of the operator's existing separated business system according to new business requirements, determine the capability gap and system integration mode, and help design and generate a new system architecture; the horizontal business platform provides business channel integration, business logic orchestration and Capabilities for business delivery management.

Although the above three typical telecom service support platform products can meet the service support requirements in the telecom field, however, from the perspective of triple play, these three platforms only provide support for telecom services and are not compatible with broadcasting and network services. and Internet services; from the perspective of platform capabilities, these three platforms do not have the advantages brought by the cloud platform, such as service providers renting network equipment on demand and dynamically scaling the required resources, thereby improving equipment utilization. Solve the problems encountered in the service support technology in the telecommunications field mentioned above.

With the rapid development of the Internet field, cloud computing technology has attracted extensive attention from the industry due to its on-demand service, flexible scalability, high reliability, and low price. Users who access the "cloud" can obtain much higher computing power than local resources, and can flexibly change their ordered cloud computing power according to actual needs. Since the service is in the cloud, and the cloud provides high reliability guarantee, users do not need to perform any maintenance operations locally, which can significantly save a lot of hardware investment and maintenance costs.

Cloud platform, or Platform as a Service (PaaS, Platform as a Service), as one of the ways cloud computing provides services, can be considered as the core of the entire cloud computing system, which provides Internet business development, deployment and operation environment of. Business developers can publish their own developed business to the cloud platform for business debugging and operation, and in the process, the platform provides all the resources needed by the business, and users do not need to worry about how to manage these resources at all. At present, cloud platform products are relatively mature: Google's Google App Engine, Microsoft's Azure and Apache's Hadoop, etc.

Google App Engine is essentially a platform composed of application server farms, BigTable structured data distribution storage system and GFS data storage services. It provides developers with a set of application programming interfaces. Developers use Python or Java programming language to write Web applications. program, and use these interfaces in the program to access services such as space, database storage, e-mail and memcache provided by Google.

Windows Azure Platform is a cloud computing platform running in Microsoft's data center, which includes a cloud computing operating system and a set of services for developers. Applications created by developers can either run directly on the platform or use the services provided by the cloud computing platform. The Windows Azure platform includes Windows Azure, SQL Azure, and Windows Azure platform AppFabric.

Hadoop mainly implements the idea of GFS and the MapReduce model. As an open source software platform, Hadoop makes it easier to write and run applications that process massive amounts of data. Hadoop mainly includes three parts: Hadoop Distributed File System (HDFS), MapReduce implementation and HBase (Google Bigtable implementation). When HDFS stores data, it divides files into data blocks of a set size, and each block is distributed to each node in the cluster. In order to ensure reliability, HDFS will create multiple copies of data blocks according to the configuration, and place them on the computing nodes of the cluster respectively. MapReduce divides the application into multiple small tasks for execution, and each small task only processes the data blocks stored locally on the computing node.

Although the above cloud platform products have their own advantages, they are only oriented to Internet services, which makes it difficult to promote the cloud computing provided by them to converged networks. Taking the telecommunications network in the converged network as an example, the main reasons for analyzing the bottlenecks in it are:

(1) In terms of network protocols, the Internet mainly uses the HTTP protocol, which is relatively simple; the communication protocols of the telecommunication network are numerous and complex, and the definitions of each protocol are completely different from HTTP.

(2) From the perspective of the network conversation model, Internet services are mainly stateless non-conversational services; telecom network services include both stateful conversational services (such as call services) and non-conversational services (such as short message services). business). Since conversational services are stateful, their processing is more complicated than that of non-conversational services. Internet-oriented cloud platforms cannot be applied to telecommunication networks because they cannot support basic telecommunication protocols and cannot meet the maintenance requirements of call sessions unique to telecommunication services.

A similar situation holds for other networks in the converged network. Therefore, various existing cloud platforms are not suitable for converged networks.

Chinese patent application (CN200710002988.5) proposes a general service platform supporting multiple services based on multi-network integration. The general service mainly includes: a service access unit connected to the communication network of each network operator for access Input various communication services, provide protocol support and perform data interaction with the message processing unit; the message processing unit accepts the service request of the service access unit, judges the type of request, calls the service instance in the service engine unit and receives the service engine unit to send service instructions and forward them to the service access unit; the service engine unit is used to provide operation support for each communication service, execute the service according to the service invocation request sent by the message processing unit, trigger a new service request according to the execution result, and send the new The request is returned to the message processing unit; the management unit is respectively connected to the above units and completes the coordination and management of them. Although the general service platform provided by this patent application can provide a variety of network access protocols to support different services of heterogeneous network operators; however, it cannot provide equipment rental, on-demand purchase of computing resources and The ability to dynamically change and adjust resources for business load cannot solve the inherent shortcomings of traditional telecom service support platforms.

Therefore, the prior art has not yet solved a working platform capable of supporting converged network services, and this has become a focus of attention of technical personnel in the industry.

Contents of the invention

In view of this, the purpose of the present invention is to overcome the deficiencies of the Internet cloud platform and traditional telecom service support platform in the prior art, and provide a cloud platform supporting converged network services and its working method. The way of resources provides platform users (including triple network operators or service providers, etc.) The platform can provide access to the resources of the three networks, and provide convenience for platform users to develop and operate converged network services.

In order to achieve the above-mentioned purpose of the invention, the present invention provides a cloud platform supporting converged network services, which is characterized in that: the cloud platform is respectively connected through three core networks provided by three network operators and corresponding different communication protocols The three access networks of the telecommunications network, the Internet and the radio and television network; the cloud platform is equipped with a variety of hardware and software resources including CPU, hard disk, memory, bandwidth and communication protocols for various other network operators and service providers These resources are leased according to their respective needs to deploy and operate their respective services; the cloud platform adopts a layered structure, with platform management layer, business execution layer, resource virtualization and management layer, and hardware resource layer respectively from top to bottom. The interaction between layers adopts the function provided by the upper layer module in the form of interface call; each layer has the following functional modules:

The cloud platform management module is located in the management layer of the platform and provides the management and operation functions of the platform. It is used to provide access interfaces for platform administrators and platform users. It costs four units and provides corresponding functions respectively; this module uses the data storage capacity provided by the cloud platform resource pool to store the registration information of platform users and related information of ordered resources, the platform users and business status to which the business belongs, and the number of business engine instances The information and information of the platform including service node information, platform logs and alarms are distributed and stored in each resource node of the underlying data storage server;

Business execution module, the core component of business execution of the cloud platform at the business execution layer, has multiple business execution spaces that represent set resource capabilities and are essentially logical entities composed of multiple virtual machines, and are responsible for saving and managing business The business execution space management module of the execution space; this business execution module uses the virtualized hardware resources provided by the cloud platform resource pool, responds to the resource ordering requests of platform users, and divides multiple business execution spaces with different resource capabilities for different platform users, And match the resource capabilities of each execution space with the resource capabilities ordered by platform users;

The cloud platform resource pool is located in the resource virtualization and management layer. As a key component to support the operation of the cloud platform, this module is equipped with a resource access management unit, a data storage management unit, a resource virtualization unit and a resource access management unit; It provides access to a large number of server hardware, and maintains and manages the accessed hardware resources after virtualization for calls by upper-layer modules; virtualized resources are divided into two categories: computing resources including CPU resources, memory resources, and bandwidth resources Resources and hard disk database storage resources;

Computing server resources and data storage server resources composed of a large number of underlying physical servers are located at the hardware resource layer and provide the most basic hardware resource support for the cloud platform; by running resource agent programs on these underlying physical servers, the underlying physical The server becomes a computing server resource or a data storage server resource, and then by registering the underlying physical server as a computing resource and data storage resource to the resource pool access management unit of the upper-level cloud platform, the upper-level module can be provided with the ability to access local resources .

In order to achieve the above object of the invention, the present invention also provides a working method of a cloud platform supporting converged network services, which is characterized in that: comprising the following steps:

(1) Deployment and startup of the cloud platform: after adopting the multi-server cluster distributed structure to deploy the cloud platform, start the cloud platform according to the sequence from the bottom layer to the upper layer of the cloud platform architecture components; this step (1) includes the following operations:

(11) Start and run the service program of the cloud platform resource pool on some servers first, so that these servers form a cloud platform resource pool server cluster, and then let the cloud platform resource pool server cluster open the access interface of the resource pool resources, and wait Access to underlying hardware resources;

(12) Select a large number of physical servers at the three levels of cross-region, cross-computer room and cross-rack, and start and run the underlying resource service program, so that these physical servers form the underlying resource server cluster; if the underlying resource server runs computing resources If the proxy program is used as a computing resource server, it will register and access to the resource pool server; if the data storage agent program runs on the underlying resource server, it will register and access as a data storage resource server;

(13) The cloud platform resource pool server cluster performs real-time monitoring on the connected underlying resource server cluster, and virtualizes the connected server capabilities to facilitate the dynamic adjustment of resources and serve as virtual resources in the cloud platform resource pool Provided for use by upper-layer services; due to the redundant storage of physical resources and the ability to divide them into small blocks, it can increase the availability of the cloud platform and significantly improve the utilization of physical resources;

(14) Start and run business execution and management service programs on some servers, so that these servers form a business execution and management server cluster; after the business execution and management server cluster is connected to the cloud platform resource pool server cluster, obtain the cloud The virtual resources used by the platform itself, and then initialize the resources including system data;

(15) Start and run the service program managed by the cloud platform on some servers, so that these servers can be used as cloud platform management server clusters, and after connecting the business execution and management server clusters and cloud platform resource pool server clusters respectively, perform corresponding initialization ;

(16) After the cloud platform operator administrator performs the initial configuration operation on the entire cloud platform system, the access interface is opened to the cloud platform users, and the cloud platform startup is completed;

(2) The cloud platform executes platform user registration, resource application, and loading and deployment of new services: after the cloud platform starts up and receives the platform user's registration and purchase of the required resources, it creates resources for the platform user according to the number of resources ordered by the platform user. Its unique virtual business execution space, and create a business instance including a business execution engine and a load balancer for it; this step (2) includes the following operations:

(21) Cloud platform users including service providers or network operators access the user registration function provided by the user management unit of the platform management module, fill in user information, and complete registration;

(22) Platform users order various cloud platform resource capabilities including CPU resources, data storage resources and network bandwidth from the resource allocation and scheduling unit of the cloud platform according to their own needs;

(23) The resource allocation scheduling unit sends a request to the execution space management module to create a business execution space, and the execution space management module applies to the platform resource manager for the virtual resources required by the platform user, and creates a business execution space and business execution based on the resources. virtual machine instances in the space;

(24) The execution space management module sends a virtual machine file backup request to the platform resource manager, and the platform resource manager performs regular backup and synchronization according to the redundant backup strategy; so far, the cloud platform completes the platform user resource ordering operation;

(25) The platform user accesses the business management unit of the platform management module, and uploads the business to be deployed on the cloud platform; the business management unit sends a service loading request to the execution space management module;

(26) The execution space management module responds to the service loading request, creates a corresponding number of service execution engine instances, loads the business and returns a successful loading response;

(3) The cloud platform performs business-related operations and dynamically adjusts resources according to the platform user's instructions: during the operation process of the platform user using the cloud platform, the cloud platform dynamically adjusts the resources used by it according to the actual needs of the business; this step (3) includes The following operations:

(31) After the cloud platform completes the service loading, according to the instructions issued by the platform user through the service lifecycle management interface of the service management unit, perform service activation, deactivation or offline operations;

(32) After the business has been running for a period of time, if the platform user finds that the existing platform resource capacity is not enough to support the load of business access or needs to reduce the rented platform resources, he will log in to the resource allocation and scheduling unit to reorder resources and dynamically adjust the number of required resources ;

(33) The resource allocation scheduling unit sends a resource re-application request to the execution space management module; the execution space management module applies to the platform resource manager for adjusted resources, and changes the configuration of the original business execution space, and adjusts the number of virtual machine instances and business The number of execution engine instances; so far, the cloud platform completes the dynamic adjustment of resources according to the platform user's request;

(4) The cloud platform supports the operation of converged network services: after the platform users who register and purchase resources upload and activate the service, the service is in the activated running state on the cloud platform and can accept traffic from various operator networks outside the cloud platform. business request; and during the operation of the business, the cloud platform can simultaneously support three different network resources and related protocols of the Internet, telecommunication network and broadcasting network involved in the business; this step (4) is processing the business request reported by the network Or when a business request is initiated by the business itself, the following operations are included:

(41) The service request arrives at the load balancer, and according to the service trigger strategy, the service request is forwarded by the load balancer to a specific service execution engine instance;

(42) The corresponding protocol stack adapter of the business execution engine instance receives the forwarded business request, analyzes and processes the business request, encapsulates it into a platform internal message event, and sends the internal message event according to the subscription relationship of the business to resources to the corresponding business instance management unit;

(43) The service instance management unit searches for the corresponding service instance according to the session identifier in the message event, if found, then sends the message event to the service instance for business processing, otherwise a new service instance will be created and linked to the message The session ID in the event is saved together in the business instance mapping list of the business instance management unit itself;

(44) The message event is sent to the event queue of the corresponding business instance. The business instance is encapsulated into a Task and placed in the task execution queue in the business execution engine for processing. For stateless business, after the business processing is completed, the business instance The service instance management unit will be notified to delete the service instance and its corresponding session identifier from the service instance mapping list; for stateful services, the service instance will still be stored in the service instance mapping list, waiting for subsequent service requests;

(45) The follow-up service request of the stateful service will be directly sent to the corresponding protocol stack adapter of the corresponding service execution engine. After step (42), the follow-up service request is parsed into an internal message event and sent to the corresponding service instance management unit. Through the session ID in the event, the business instance management unit finds the corresponding business instance for corresponding business processing; after the stateful business instance processes the entire business logic, it notifies the business instance management unit to delete the business instance itself and its session ID mapping ;

(46) The system processing result of the service is transmitted to the corresponding protocol stack adapter and sent to the network.

Compared with the closest prior art, the present invention has the following advantages: the cloud platform supporting converged network services in the present invention is a platform that provides unified service deployment and operation, and has the ability to simultaneously support the execution of conversational services and non-conversational services mechanism, and provides a large number of protocol stack resources in the field of triple play, so that the triple play service can be deployed and operated on this platform at the same time, which has changed the way that each service provider (SP) connected its own equipment to the network operator in the past, thus The network structure and service provision mode of their respective services are provided separately, which also overcomes the limitation that the Internet cloud platform can only support Internet services. In this way, on the one hand, it saves the financial resources and labor costs of service providers and network operators for maintaining multiple service servers; on the other hand, it also simplifies the management of service providers by network operators, and makes the provision of services more Openness means that service providers no longer need to have their own equipment to provide and operate services, which greatly reduces the threshold for service provision and is conducive to the innovation and diversification of converged network services.

The invention supports the cloud platform of the converged network service to make full use of the virtualization technology, and conveniently provides the resources of the three networks for service use. Platform users can rent the platform environment for business operations. At the same time, they can rent and customize resource capabilities according to their own needs, so that the provided services have elastic and scalable functions to meet the requirements of on-demand services, which not only improves equipment utilization, but also makes individual business Ability to achieve higher server performance. Compared with the traditional telecommunications service support platform, the present invention adopts the cloud platform of virtualization technology, and also adopts distributed service provision and data storage, and adopts redundancy mechanisms of different granularities such as cross-node, cross-rack, and cross-machine room, etc., avoiding single Nodes provide business and the hidden dangers of service downtime and data loss caused by single-node data storage, so as to ensure the high reliability of the business platform and the business running on the platform. Not only is it improved through distributed technology, but it also makes the management of service providers easier, and solves the shortcomings of the low utilization rate of the telecom application server and the inability to flexibly adjust resources.

The technical innovation of the present invention is: a cloud platform that supports converged network services is proposed, which changes the network structure and service provision method of service providers in the past, and adopts the form of renting platform resources on demand to deploy and operate various service providers Business. In terms of resource management, the cloud platform provides businesses with the ability to flexibly configure resources, which not only increases the upper limit of service capabilities, but also greatly improves the utilization rate of equipment. The cloud platform can provide rich protocol stack resources in the fields of Internet, telecommunication network and broadcasting network to facilitate the development of integrated network services. In addition, the platform adopts cloud data storage, shields the underlying details of the database from users, and backs up data files with cross-domain redundancy, which has high availability and high security, and can ensure long-term and stable operation of the business. The platform also has session management capabilities and provides support for stateful converged network services. In a word, the present invention has a very good prospect of popularization and application.

Description of drawings

FIG. 1 is a schematic diagram of a network location of a cloud platform supporting converged network services according to the present invention.

FIG. 2 is a schematic diagram of the overall architecture composition of the cloud platform supporting converged network services according to the present invention.

Fig. 3 is a schematic diagram of the structure composition of the service execution space of the cloud platform of the present invention.

Fig. 4 is a schematic diagram of the composition of the layered structure of the service execution engine of the cloud platform of the present invention.

Fig. 5 is a flow chart of operation steps of the working method of the cloud platform supporting converged network services according to the present invention.

Fig. 6 is a schematic diagram of a cloud platform deployment embodiment in the cloud platform working method of the present invention.

Fig. 7 is a sequence interaction diagram of user registration, resource ordering, and business operation performed by new platform users in the cloud platform working method of the present invention.

Fig. 8 is an interaction diagram of dynamic resource adjustment performed by platform users in the cloud platform working method of the present invention.

Fig. 9 is a schematic flowchart of an implementation example of a third-party calling service in the working method of the cloud platform of the present invention.

Fig. 10 is a schematic diagram of the service execution engine processing the Http request in the cloud platform working method of the present invention.

Fig. 11 is a schematic diagram of the service execution engine processing WebService requests in the cloud platform working method of the present invention.

Fig. 12 is a schematic diagram of processing a SIP message (INVITE) by the service execution engine in the working method of the cloud platform of the present invention.

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments.

Referring to Fig. 1, the location and structure of the cloud platform supporting converged network services of the present invention deployed in the network system are introduced: on the basis of the three access networks of telecommunications, Internet, and radio and television, the respective cores provided by the operators of the three networks are respectively provided. The network and the corresponding communication protocol are connected upward to the converged network service cloud platform. Other network operators or service providers need to rent the service execution engine resources of the cloud platform to deploy and operate services in their respective fields, and the cloud platform provider is responsible for the management and maintenance of the cloud platform.

Referring to FIG. 2 , the overall structure of the cloud platform supporting converged network telecommunication services of the present invention is introduced in detail. The one-way arrows in the figure represent interface calls, the two-way arrows represent data reading and writing, the rectangles with four corners arcs represent virtual spaces, and the rectangles with four right angles represent modules or units.

The cloud platform of the present invention is to respectively connect three access networks of the telecommunication network, the Internet and the radio and television network through three core networks and corresponding respective communication protocols provided by the operators of the three networks; A variety of hardware and software resources, memory, bandwidth and communication protocols, for other network operators and service providers to rent these resources to deploy and operate their own services according to their own needs. The cloud platform adopts a layered structure, which consists of hardware resource layer, resource virtualization and management layer, business execution layer and platform management layer from bottom to top. The interactive relationship between layers is that the upper layer modules use the functions provided by the lower layer modules in the form of interface calls. Each layer has the following functional modules:

The cloud platform management module 01 located in the platform management layer provides management and operation functions of the platform. This module is used to provide access interfaces to platform administrators and platform users. It is equipped with platform user management, business management, resource scheduling and It costs four units and provides corresponding functions respectively. This module uses the data storage capacity provided by the cloud platform resource pool 03 to combine the registration information of platform users and related information of ordered resources, the information of platform users and business status to which the business belongs, the number of business engine instances, and the platform's service node information. , platform log and alarm information are distributed and stored in each resource node 05 of the underlying data storage server. The following is a detailed introduction to the structure and functions of the four units in the cloud platform management module 01:

The user management unit 011 is responsible for providing functions including platform user registration and login, platform user basic information management and platform user authority distribution. After the platform user accesses, they must first register to apply for the permission to use the cloud platform. User registration information is stored in the resource node 05 of the underlying data storage server through the cloud storage capability provided by the cloud platform resource pool 03.

The business management unit 012 is used to provide business life cycle management and related configuration functions. When platform users perform various business operations on their own business, including loading, trial operation, activation, deactivation and offline, etc., they can Its business sets the number of instances of the execution engine to determine the maximum concurrent capability of the business. This unit is to call the relevant interface of the business execution space management module 021 to support the above-mentioned various business operations.

The resource scheduling allocation unit 013 is responsible for providing and configuring the required resources of platform users. Platform users determine the hardware resources they need, such as the number of CPUs, memory, hard disk capacity, and network bandwidth, through resource ordering. The business execution module 02 can maintain a business execution space for the user according to the subscription resources of the user.

The billing unit 014 is used to complete the billing function for platform users, that is, network operators and service providers. This module generates the user's billing bill according to the number of resources applied by the user and the time used, and stores it in the storage resources provided by the cloud platform resource pool 03.

The business execution module 02 located at the business execution layer is the core business execution component of the cloud platform. It is equipped with multiple business execution spaces that represent set resource capabilities and are essentially logical entities composed of multiple virtual machines, and is responsible for saving and managing business The business execution space management module of the execution space. This module uses the virtualized hardware resources provided by the cloud platform resource pool 03, responds to the resource ordering requests of platform users, divides multiple business execution spaces 022 with different resource capabilities for different platform users, and makes the resource capabilities of each execution space Match the resource capabilities ordered by platform users. The business execution space representing a certain resource capability is a logical entity, which is saved and managed by the business execution space management module 021 . The business execution space actually consists of several virtual machines, and its internal structure is detailed in Figure 3.

The cloud platform resource pool 03 located in the resource virtualization and management layer is a key component to support the operation of the cloud platform. This module provides access to a large number of server hardware downwards, and maintains and manages the accessed hardware resources after virtualization for calling by upper-layer modules. Virtualized resources are divided into two categories: computing resources (including CPU resources, memory resources, and bandwidth resources, etc.) and storage resources (hard disk database). This module is provided with a resource access management unit 031 , a data storage management unit 032 , a resource virtualization unit 033 and a resource access management unit 034 . The structure and functions of these four units are described in detail below:

The functions and structures of the resource access management unit, data storage management unit, resource virtualization unit and resource access management unit in the cloud platform resource pool are;

The resource access management unit 034 is used to establish connections with each server in the form of the latter registering with the former; the accessed server resources include two types: computing server resources 04 and data storage server resources 05; the distinction between these two types of resources It is only logical, representing a certain computing or storage function, not a physical concept, that is, the same physical server host can be both a computing server resource and a data storage server resource. The unit can obtain the ability to access the local resources of the server through registration, and the accessed server resources can send and receive heartbeat messages with the unit to ensure that the unit can monitor the running status of the server nodes in real time.

The resource virtualization unit 033 is used to virtualize the resources of the resource access management unit 034. Through the description of access resources and the encapsulation of access capabilities, it is abstracted into unified cloud platform computing resources and data storage resources, so as to facilitate resource management. Management and use of resources by upper-layer interfaces.

The resource access management unit 031 is used for allocating virtualized computing resources in response to computing resource application requests issued by the business execution space management module. This module adopts a redundant strategy to allocate resources, and saves virtual machine files in one main address and multiple backup addresses. If the server at the main address works abnormally, the server at the backup address runs the virtual machine file to ensure that the service will not be caused by individual servers. Discontinued due to problems.

The data storage management unit 032 is used to provide the database cloud storage capability of the cloud platform. This unit utilizes virtualized storage resources, responds to the request of the data storage access interface of the platform, and redundantly saves and reads the contents of the database in blocks; this unit also saves the index information and write protection lock of each block data to ensure that the data It can read and write successfully, and at the same time modify the record information according to the index, and regularly synchronize the backup data.

Computing server resources 04 and data storage server resources 05, which consist of a large number of underlying physical servers, are located at the hardware resource layer, providing the most basic hardware resource support for the cloud platform. By running the resource agent program on these underlying physical servers, these underlying physical servers can be made into computing server resources 04 or data storage server resources 05, and then by using the underlying physical servers as computing resources and data storage resources to the cloud platform The resource pool access management unit 034 registers to provide the upper layer module with the ability to access local resources.

Referring to Fig. 3, the internal structure of the business execution space of the business execution module in the cloud platform of the present invention is introduced:

Each business execution space is equipped with at least one business execution space management agent unit, a load balancing process and multiple business execution engine instance processes, wherein the load balancing process and the business execution engine instance processes run in their respective business execution space virtual machines middle. in,

The business execution space management agent unit is used as the interface of the business execution space management module 021. On the one hand, this unit should respond to the command request of the business execution space management module, and the execution includes creating a new virtual machine and execution engine instance, loading and activating the business, etc. On the other hand, it is necessary to dynamically monitor the load balancer and execution engine instances in the space, and judge the survival status of each process through the heartbeat mechanism. Once the process is found to be abnormal, restart the process in time, and generate an alarm log to the business execution space management module 021.

The load balancing process is the connection portal between the cloud platform and the operator's network in the present invention. The services provided by different network operators need the support of different protocols, and the cloud platform supports multiple protocols in different networks, thereby realizing the support for different network services. When a service request is sent from the operator network to the open port of the load balancing process, the service request is forwarded to the current execution engine instance with a lighter load according to the service triggered by the service request. The execution engine instance is the actual container where the business instance runs. Figure 4 illustrates its structure and method in detail.

Referring to Figure 4, the structural composition of the business execution engine is introduced in detail: from bottom to top, there are resource layer, business instance management layer, business execution layer, and business management layer, as well as a statistical monitoring and alarm module that monitors the execution of each layer. in,

The resource layer is provided with protocol stack adaptation resources and data storage access interface resources. The former encapsulates a variety of protocol stack adapters of the telecom network, the Internet, and the radio network, such as HTTP, SIP, Parlay/ParlayX, WebService, SMS, MMS, GIS, LBS, and Mail, etc., which are used to interpret various protocol events and encapsulate them into internal Message events facilitate resource expansion and unified management. The latter provides the operations of adding, deleting, modifying and querying the database, so as to call the cloud storage capability provided by the data storage management unit in the cloud platform resource pool.

The business instance manager of the business instance management layer is responsible for the management and maintenance of business instances and the subscription and maintenance of business resources. It is characterized by being able to support conversational services; it has a business instance management unit and a resource subscription unit. The business instance management unit is used to maintain and manage business instances. It realizes support for stateful conversational services by maintaining the service instances in the service execution engine and their corresponding service session ID mappings; When a request event arrives from the underlying protocol stack adapter, if there is already a service instance corresponding to the session ID in the request event in the business instance management unit, the new request event will be forwarded to the event queue maintained by the service instance , and encapsulate the business instance into a processing task TASK and deliver it to the business execution queue in the business execution core; otherwise, create a new business instance, and the new business instance and its business session ID mapping will be maintained in pairs In the business instance management unit. Each time a complete business process ends, the business instance will notify the business instance management unit to delete the business instance itself and its business session ID mapping.

The resource subscription module is used to maintain the subscription relationship between business and resources. There is not a one-to-one correspondence between each resource instance and business instance. Some resources, such as SIP protocol stack and web service protocol stack, are between resources and services One-to-many relationship. When the business is deployed, the resource subscription unit maps the business and the resources according to all the resources required by the business listed in the business configuration file, and when the event that the business registers with the corresponding resource arrives at the resource, the resource subscription unit passes Find the subscription relationship and deliver related resources to the correct business instance.

The business execution core unit of the business execution layer has a task execution queue, and each business instance and an internal event are put into the task execution queue as a task, waiting to be processed by the execution thread. The business execution engine adopts a concurrency mechanism to ensure that each engine instance can run concurrently. Heartbeat events can be encapsulated into execution tasks and placed in the execution queue. When the execution thread processes the task, the business execution engine generates and sends a corresponding heartbeat message to the business execution space management module. The advantage is that each heartbeat is issued to ensure that the execution thread is working normally, preventing the occurrence of an abnormal execution thread but still being able to send a heartbeat; thus realizing the real-time monitoring of the status of each engine instance by the cloud platform and ensuring high fault tolerance.

The business container of the business management layer is provided with a business management unit, a business trigger rule management unit and an engine management unit. When each business is loaded and executed, the business management unit is responsible for parsing the configuration file of the business and performing the initialization operation of the business; the business logic provided by the developer is compiled into a dynamic link library, which will be linked when the execution engine executes an instance of the business dynamic link library. Each business has a set of triggering rules for judging whether it is triggered. The service trigger rule management unit is used to store and maintain the corresponding relationship between service trigger rules and services; after each service is loaded, its service trigger rules will be included in the telecommunication system including the softswitch entity and IP multimedia subsystem in the next generation network NGN Acquired by the Serving Call Session Control Function S-CSCF entity in the IMS. The engine management unit is used to provide various scripts including Python and BPEL, and website language execution engines including JSP and ASP and their corresponding management interfaces. These website language execution engines are used to support corresponding services, and the JSP engine is used for platform support. Internet JSP web page business.

The statistical monitoring and alarm module is responsible for monitoring the running status of the execution engine: counting the number of currently surviving business instances of each business, and periodically sending the statistical data to the business execution space management agent unit; periodically generating heartbeat events, and then packaging them into execution tasks and adding them to the execution The head of the engine team enables the execution engine to periodically send heartbeat data to the business execution space management agent unit to realize the heartbeat mechanism; provides alarm tools for each component of the execution engine, and transmits the generated alarm information to the upper management node when an exception occurs.

Referring to Fig. 5, the working method of the cloud platform supporting the converged network service of the present invention is introduced, including the following steps:

Step 1. Deployment and startup of the cloud platform: After deploying the cloud platform with a multi-server cluster distributed structure, start the cloud platform in the order from the bottom layer to the top layer of the cloud platform architecture components.

Step 2: The cloud platform executes platform user registration, resource application, and loading and deployment of new services: After the cloud platform starts up and receives the platform user's registration and purchase of the required resources, it creates resources for the platform user according to the number of resources ordered by the platform user. Its unique virtual business execution space, and create business instances including business execution engine and load balancer for it.

Step 3: The cloud platform performs business-related operations and dynamically adjusts resources according to the platform user's instructions: When the platform user uses the cloud platform to operate the business, the cloud platform dynamically adjusts the resources used by it according to the actual needs of the business.

Step 4, the cloud platform supports the operation of converged network services: after the platform users who register and purchase resources upload the service and activate the service, the service is in the activated running state on the cloud platform and can accept traffic from various operator networks outside the cloud platform business request; and during the operation of the business, the cloud platform can simultaneously support the three different network resources and related protocols of the Internet, telecommunication network and broadcasting network involved in the business.

Below in conjunction with the embodiment of the present invention (the cloud platform service execution engine supports multiple sip business instances of the telecommunication network), with the basic sequence of time, the specific method of the above-mentioned steps of the cloud platform of the present invention is described in detail:

Referring first to the deployment diagram of the cloud platform embodiment in Fig. 6, introduce step 1, deployment and startup of the cloud platform: the cloud platform of the present invention adopts a distributed structure of multi-server clusters, and its deployment and startup are all based on the platform architecture from bottom to bottom The upper sequence is accomplished in the following specific steps:

(11) Start and run the service program of the cloud platform resource pool on some servers first, so that these servers form a cloud platform resource pool server cluster, and then let these cloud platform resource pool server clusters open the access interface of the resource pool resources, and wait Access to underlying hardware resources.

(12) Select a large number of physical servers at the three levels of cross-region, cross-computer room and cross-rack, and start and run the underlying resource service program, so that these physical servers form the underlying resource server cluster. If the underlying resource server runs a computing resource agent program, it registers and accesses to the resource pool server as a computing resource server; if the underlying resource server runs a data storage agent program, it registers and accesses as a data storage resource server.

(13) The cloud platform resource pool server cluster performs real-time monitoring on the connected underlying resource server cluster, and virtualizes the connected server capabilities to facilitate the dynamic adjustment of resources and serve as virtual resources in the cloud platform resource pool Provided for use by upper layer services.

As shown in Figure 6, in each resource server distributed at the bottom layer, physical resources are divided into small resource blocks, and each small block marked with a different label represents the resources provided by four cloud platform users A, B, C, and D respectively. The physical resources of the service are connected to the cloud platform resource pool server cluster, and these physical resource blocks are virtualized and provided to users with integrated virtual resource capabilities. The virtual resources are represented by the four dotted boxes A, B, C, and D in the server cluster of the cloud platform resource pool, and their sizes represent the capacity of the virtual resources. For example, the resource capacity of the cloud platform required by user A is greater than that of the other three users. Through the virtualization of physical resources, dynamic adjustment of resources can be easily realized. At the same time, due to the redundant storage of physical resources, the availability of the cloud platform can be increased; the resource capability of dividing physical resources into small blocks also greatly improves the utilization of physical resources.

(14) Start and run service programs for business execution and management on some servers, so that these servers form a cluster of business execution and management servers. After these business execution and management server clusters are connected to the cloud platform resource pool server clusters, they obtain the virtual resources used by the cloud platform itself, and then initialize the resources including system data.

(15) Start and run the service program managed by the cloud platform on some servers, so that these servers can be used as cloud platform management server clusters, and after connecting the business execution and management server clusters and cloud platform resource pool server clusters respectively, perform corresponding initialization .

(16) The administrator of the cloud platform operator performs an initial configuration operation on the entire cloud platform system, and then opens the access interface to the cloud platform users, and the start of the cloud platform is completed.

Referring to Figure 7, step 2 is introduced, the cloud platform user performs user registration, resource application and new service loading:

After the platform is launched, platform users must first register with the platform, fill in the basic information and purchase the required resources, and the platform will create its unique virtual execution space for the user according to the number of resources ordered by the user, and initially create a platform for it Execution engine, load balancer and other instances. This step includes many interactive actions, so it will be described in the interactive form shown in FIG. 7 .

(21) Cloud platform users (including service providers or network operators) log in to the user registration page provided by the user management unit of the access platform management module, fill in user information, and complete the registration.

(22) Platform users complete ordering resource capabilities; according to their own needs, order various cloud platform resource capabilities (including CPU resources, data storage resources, network bandwidth, etc.) required by the resource allocation and scheduling unit of the cloud platform.

(23) The resource allocation scheduling unit sends a request to the execution space management module to create a business execution space, and the execution space management module applies to the platform resource manager for the virtual resources required by the platform user, and creates a business execution space and business execution based on the resources. A virtual machine instance in a space.

(24) The execution space management module sends a virtual machine file backup request to the platform resource manager, and the platform resource manager performs regular backup and synchronization according to the redundancy backup strategy; so far, the cloud platform completes the platform user resource ordering operation.

(25) The platform user accesses the business management unit of the platform management module, and uploads the business to be deployed on the cloud platform; the business management unit sends a service loading request to the execution space management module.

(26) The execution space management module responds to the service loading request, creates a corresponding number of service execution engine instances, loads the service and returns a loading success response.

Referring to Figure 8, it is introduced that after the cloud platform completes service loading, it performs service-related operations and dynamically adjusts leased resources according to platform user instructions. This step includes the following operations:

(31) After the cloud platform completes the service loading, it executes service activation, deactivation or offline operations according to the instructions issued by the platform user through the service lifecycle management interface of the service management unit.

(32) After the business has been running for a period of time, if the platform user finds that the existing platform resource capacity is not enough to support the load of business access or needs to reduce the rented platform resources, he will log in to the resource allocation and scheduling unit to reorder resources and dynamically adjust the number of required resources .

(33) The resource allocation scheduling unit sends a resource re-application request to the execution space management module; the execution space management module applies to the platform resource manager for adjusted resources, and changes the configuration of the original business execution space, and adjusts the number of virtual machine instances and business The number of execution engine instances; so far, the cloud platform completes the dynamic adjustment of resources according to the platform user's request.

Step 4, the cloud platform supports the operation of converged network services: when services are executed in the service execution engine of the cloud platform, the protocol stack adapters and corresponding protocols adopted by the cloud platform will be different for different network services; and, According to the characteristics of different network services, the support functions of the cloud platform for the service state are also different; however, they are basically the same. For example, this step includes the following operations when processing a service request reported by the network or a service request initiated by the service itself:

(41) The service request arrives at the load balancer, and according to the service trigger policy, the service request is forwarded by the load balancer to a specific service execution engine instance.

(42) The corresponding protocol stack adapter of the business execution engine instance receives the forwarded business request, analyzes and processes the business request, encapsulates it into a platform internal message event, and sends the internal message event according to the subscription relationship of the business to resources to the corresponding business instance management unit.

(43) The service instance management unit searches for the corresponding service instance according to the session identifier in the message event, if found, then sends the message event to the service instance for business processing, otherwise a new service instance will be created and linked to the message The session ID in the event is stored together in the business instance mapping list of the business instance management unit itself.

(44) The message event is sent to the event queue of the corresponding business instance. The business instance is encapsulated into a Task and placed in the task execution queue in the business execution engine for processing. For stateless business, after the business processing is completed, the business instance The service instance management unit will be notified to delete the service instance and its corresponding session identifier from the service instance mapping; for stateful services, the service instance will still be stored in the service instance mapping, waiting for subsequent service requests.

(45) The follow-up service request of the stateful service will be directly sent to the corresponding protocol stack adapter of the corresponding service execution engine. After step (42), the follow-up service request is parsed into an internal message event and sent to the corresponding service instance management unit. Through the session ID in the event, the business instance management unit finds the corresponding business instance for corresponding business processing; after the stateful business instance processes the entire business logic, it notifies the business instance management unit to delete the business instance itself and its session ID mapping .

(46) The system processing result of the service is transmitted to the corresponding protocol stack adapter and sent to the network.

Finally, an embodiment is described in detail: the implementation process of the cloud platform running the three-party calling service of the converged network. This service can be completed: the user logs in the webpage provided by the service from the Internet browser, clicks to dial, and realizes the video call between the mobile phone terminal A and the cable digital TV terminal B. The whole business process involves the Internet terminal, the telecommunication network terminal and the radio network terminal, reflecting the support of the cloud platform of the present invention for the three network resources. The entire business process is shown in Figure 9:

The whole business involves the support of the cloud platform for three protocols, namely the HTTP protocol, the Soap protocol and the SIP protocol. The following describes in detail how the service execution engine supports the three protocols.

Step A, when the user uses the Internet browser to log in to the business dial-up webpage, what is used is the support of the service to the HTTP protocol. Since the development of webpages in different languages requires corresponding engines to support, the webpage of this embodiment is developed using the JSP language . Fig. 10 describes the processing method of the service engine to the HTTP request sent by the user from the browser.

Referring to Figure 10, since Http requests are stateless, in the cloud platform business execution engine pair, business instances are created each time a business request arrives, and are deleted when the business request is processed, and will not be saved in the business In the instance management unit, the specific process is described as follows:

(1) The user logs in to the business website from the browser, and an Http Get request for the webpage is sent to the load balancer corresponding to the business in the cloud platform;

(2) The load balancer selects the target business execution engine, and sends the received Http Get request to the target address;

(3) After the Http protocol stack of the target business execution engine monitors the newly arriving Http Get request, it is encapsulated into an HttpEvent event, and the HttpEvent event is sent to the corresponding business instance management unit according to the service resource order information;

(4) The service instance management unit searches for a service instance matching the session ID in the HttpEvent event from the service instance and session ID mapping;

(5) Since the corresponding service instance cannot be found, request the service container to create a new service instance;

(6) The JSP engine creates a new business instance and maintains it in the business instance and session ID mapping list of the business instance management unit;

(7) The HttpEvent event is added to the processing queue of the new business instance, encapsulated into a Task, and enters the queue of the cloud platform execution engine to wait to be processed;

(8) The newly generated Task is processed by the execution thread to extract the relevant information of HttpEvent inside the Jsp engine, generate HttpRequest, and perform corresponding processing operations;

(9) After the Task is processed, an HttpResponse is generated and handed over to the Http protocol stack;

(10) The service instance notifies the service instance management unit to delete the mapping where it is located from the service instance and session ID mapping list;

(11) The Http protocol stack encapsulates the received HttpResponse into an Http response message, and then sends it to the user's browser, so that the content of the webpage is presented in front of the user.

Step B, after the above processing flow, the user opens the dialing page, enters the calling addresses of both parties in the video call, and sends a call request. The process is actually that the user sends a WebService call request to the business from the browser. Refer to how the cloud platform supports web services as shown in Figure 11.

The platform opens the Bpel service development interface for business developers to develop Bpel services, and then the Bpel services finally provide WebService services to the network. Each Bpel business actually includes two parts: Bpel script logic and Bpel script engine. The Bpel script logic is included in the Bpel script engine to be executed, as shown in the business container in Figure 10; the Bpel script engine is the last part of the business The outer layer presents to the platform. In this service, the open WebService call interface of the Bpel service is to accept the web page click-to-call request and establish a video call between the calling parties.

Each WebService message is actually implemented by carrying a Soap message on an Http message, so the cloud platform of the present invention needs to use the Http protocol stack and the Soap protocol stack to support the WebService message. The specific process is described as follows:

(1) The user enters the addresses of both parties on the click-to-call page, and clicks to call. The browser generates a corresponding Web Service request. The request is sent to the load balancer of the cloud platform through the Http protocol bearer;

(2) The load balancer accesses the business policy library to obtain the address and port of the target execution engine;

(3) The business policy library returns the target address and port to the load balancer;

(4) The load balancer forwards the Web Service request to the Http protocol stack adapter of the target execution engine. The Http protocol stack adapter of the cloud platform includes an Http server and a Web container, and the Soap protocol stack adapter is deployed in the container. When the Web Service request is received by the Http protocol stack, the Soap message body carried in the Http package can be extracted and handed over to the Soap protocol stack adapter for further analysis;

(5) The Soap protocol stack adapter parses the Soap message body and encapsulates it into a Soap event inside the platform. Since Internet business calls are mostly stateless calls, in order to support a small number of stateful Internet services and cloud platform telecom business processing methods consistent, the Soap protocol stack will assign a random session Id to the Soap event to identify the Soap message The session you are in. If the Soap message has a session Id, use the session Id in the Soap message to replace the randomly assigned Id. Soap events are sent to the call instance manager of the corresponding business;

(6) call the instance manager to search the call instance matched with the session Id in the Soap event in the instance mapping table;

(7) If there is no corresponding instance, call the method of the service container to generate the call instance;

(8) The service container generates an instance of the corresponding Bpel service, and stores the service instance and the session Id in pairs in the call instance mapping table corresponding to the service;

(9) The newly generated call instance is encapsulated into an execution task TASK that the execution engine can handle, and the generated TASK is inserted into the end of the execution engine queue;

(10) The execution thread obtains TASK from the head of the engine queue and executes it;

(11) In the TASK logic, the third-party call logic invokes the Sip protocol stack, and sends a session establishment request to both the calling party and the called party (Note: For the specific process of SIP of the third-party call logic, please refer to the relevant standards of ParlayX, which will not be specifically explained here);

(12) After TASK completes the relevant processing, it sends a WebService response to the Soap protocol stack, which records the result of the call;

(13) The Soap protocol stack and the Http protocol stack work together to generate an Http response and send it to the user's browser.

In step C, the service execution engine performs Sip signaling interaction with both calling parties to establish a video call. In this process, although the two parties are the terminal of the telecommunication network and the terminal of the radio network, since the terminal of the radio network (such as a digital TV set-top box) also accesses the core network through IP and supports the SIP protocol, there is no connection with the terminal of the radio network (such as a SIP phone). Essentially different, so this section only explains how the cloud platform service execution engine supports the SIP protocol.

Since the session establishment process involves the interaction of various SIP messages (including requests and responses), and the service execution engine has no difference in its processing mechanism, so here we only take how the service execution engine handles the INVITE message as an example to illustrate, see Figure 12 introduces the flow of the service execution engine processing SIP messages.

(1) The SIP terminal on the user side or a software device supporting the SIP protocol sends an INVITE message to the load balancer;

(2) The load balancer queries the business policy library;

(3) The service policy library returns a response and forwards the request to the destination service execution engine instance address and port;

(4) The load balancer forwards the request to the destination address;

(5) The sip protocol stack of the business execution engine instance obtains the INVITE message, encapsulates it into a platform internal event - INVITE event, locates which business the event belongs to through the business trigger rules, and sends the event to the business registered to the SIP adapter instance manager;

(6) When the instance manager receives an event, it usually extracts the session ID in the event, and queries the ID and the existing call instance mapping table to obtain the call instance. For this scenario, the INVITE message is an initial session request, and the instance manager directly invokes the service container to request a new call instance;

(7) The instance manager invokes the method of creating a new instance of the service container;

(8) The service container generates a new call instance of the corresponding service, returns it to the instance manager, and the new instance is added to the session ID and instance mapping table;

(9) The INVITE event is passed to a new call instance, which is encapsulated into an execution task TASK that can be processed by the execution engine, and the generated TASK is inserted into the end of the execution engine queue;

(10) The execution engine has a running execution thread, which obtains TASK from the head of the engine queue and executes it.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the present invention. within the scope of protection.

Claims (8)

1. A cloud platform supporting converged network services, characterized in that: the cloud platform is connected to the telecommunications network, the Internet and the radio and television network respectively through three core networks provided by three network operators and corresponding different communication protocols Three access networks; the cloud platform is equipped with a variety of hardware and software resources including CPU, hard disk, memory, bandwidth and communication protocols, for other network operators and service providers to rent these resources according to their own needs Deploy and operate their respective businesses; the cloud platform adopts a layered structure, with platform management layer, business execution layer, resource virtualization and management layer, and hardware resource layer respectively from top to bottom. The interaction between each layer adopts the upper layer The module uses the functions provided by the lower module in the form of interface call; each layer has the following functional modules: The cloud platform management module is located in the management layer of the platform and provides the management and operation functions of the platform. It is used to provide access interfaces for platform administrators and platform users. It costs four units and provides corresponding functions respectively; this module uses the data storage capacity provided by the cloud platform resource pool to store the registration information of platform users and related information of ordered resources, the platform users and business status to which the business belongs, and the number of business engine instances The information and information of the platform including service node information, platform logs and alarms are distributed and stored in each resource node of the underlying data storage server; Business execution module, the core component of business execution of the cloud platform at the business execution layer, has multiple business execution spaces that represent set resource capabilities and are essentially logical entities composed of multiple virtual machines, and are responsible for saving and managing business The business execution space management module of the execution space; this business execution module uses the virtualized hardware resources provided by the cloud platform resource pool, responds to the resource ordering requests of platform users, and divides multiple business execution spaces with different resource capabilities for different platform users, And match the resource capabilities of each execution space with the resource capabilities ordered by platform users; The cloud platform resource pool is located in the resource virtualization and management layer. As a key component to support the operation of the cloud platform, this module is equipped with a resource access management unit, a data storage management unit, a resource virtualization unit and a resource access management unit; It provides access to a large number of server hardware, and maintains and manages the accessed hardware resources after virtualization for calls by upper-layer modules; virtualized resources are divided into two categories: computing resources including CPU resources, memory resources, and bandwidth resources Resources and hard disk database storage resources; Computing server resources and data storage server resources composed of a large number of underlying physical servers are located at the hardware resource layer and provide the most basic hardware resource support for the cloud platform; by running resource agent programs on these underlying physical servers, the underlying physical The server becomes a computing server resource or a data storage server resource, and then by registering the underlying physical server as a computing resource and data storage resource to the resource pool access management unit of the upper-level cloud platform, the upper-level module can be provided with the ability to access local resources . 2. cloud platform according to claim 1, is characterized in that: the structure and function thereof of four units in the described cloud platform management module are: The user management unit is responsible for providing functions including platform user registration and login, platform user basic information management, and platform user authority distribution. After platform users access, they must first register to apply for cloud platform use authority; user registration information is passed through The cloud storage capability provided by the cloud platform resource pool is stored in the resource nodes of the underlying data storage server; The business management unit is used to provide business life cycle management and related configuration functions. When platform users perform various business operations on their own business, including loading, trial operation, activation, deactivation and offline, it is through their The business sets the number of instances of the execution engine to determine the maximum concurrent capability of the business; this unit calls the relevant interface of the business execution space management module to support the above-mentioned business operations; The resource scheduling and allocation unit is responsible for providing and configuring the resources required by the platform users. The platform users determine the hardware resources they need, including the number of CPUs, memory, hard disk capacity and network bandwidth through resource ordering; the business execution module The subscription resource situation maintains a business execution space for it; The billing unit is used to complete the billing function for platform users, that is, network operators and service providers; this module generates user bills according to the number of resources requested by users and the time used, and stores them in the cloud platform resource pool Provided storage resources. 3. The cloud platform according to claim 1, characterized in that: the internal structure of the business execution space in the business execution module is: one or more business execution space management agent units, a load balancing process and multiple The instance process of the business execution engine, where the load balancing process and the instance process of the business execution engine run in their respective business execution space virtual machines; among them, The business execution space management agent unit, as the interface of the business execution space management module, responds to the command request of the business execution space management module, executes the life cycle including creating a new virtual machine and execution engine instance, and loading and activating the business For management operations, it is also necessary to dynamically monitor the load balancer and execution engine instances in the space, and judge the survival status of each process through the heartbeat mechanism. Once the process is found to be abnormal, restart the process in time, and generate an alarm log to the business execution space management module; The load balancing process is the connection portal between the cloud platform and the operator's network. When a service request is sent from the operator's network to the open port of the load balancing process, the service request is forwarded to the current load balancing process according to the service triggered by the service request. A light execution engine instance; the execution engine instance is the actual container in which the business instance runs. 4. The cloud platform according to claim 3, characterized in that: the structural composition of the service execution engine located in the service execution space is as follows: from bottom to top, there are resource layer, business instance management layer, business execution layer and business The management layer, and the statistical monitoring and alarm module that monitors each layer; among them, The resource layer is equipped with protocol stack adaptation resources and data storage access interface resources. The former encapsulates multiple communication networks including HTTP, SIP, Parlay/ParlayX, Web Service, SMS, MMS, GIS, LBS and Mail, the Internet and the radio and television network. A protocol stack adapter, which is used to interpret various protocol events and encapsulate them into internal message events, which is convenient for resource expansion and unified management; the latter provides the operations of adding, deleting, modifying and querying the database, so as to call the data storage in the cloud platform resource pool The cloud storage capability provided by the management unit; The business instance manager of the business instance management layer is responsible for the management and maintenance of business instances and the subscription and maintenance of business resources, and can support conversational services; it has a business instance management unit and a resource subscription unit, the former is used to maintain and manage business instances; The latter is used to maintain the subscription relationship between services and resources. There is not a one-to-one correspondence between each resource instance and service instance. There is a one-to-many relationship between resources and services including the SIP protocol stack and web service protocol stack. relationship; when the business is deployed, the resource subscription unit maps the business and resources according to all the resources required by the business listed in the business configuration file, and when the business registers with the corresponding resource and the event arrives at the resource, the resource subscription unit Deliver relevant resources to the correct business instance by looking up the subscription relationship; The business execution core unit of the business execution layer has a task execution queue, and each business instance and an internal event are put into the task execution queue as a task, waiting for processing by the execution thread; the business execution engine adopts a concurrency mechanism to ensure that each engine Instances can run concurrently, and the business execution engine generates and sends heartbeat signals to the business execution space management module to realize real-time monitoring of the status of each engine instance by the cloud platform and ensure high fault tolerance; The business container of the business management layer is equipped with a business management unit, a business trigger rule management unit and an engine management unit; when each business is loaded and executed, the business management unit is responsible for parsing the configuration file of the business and performing the initialization operation of the business; the business trigger rule The management unit is used to store and maintain the corresponding relationship between service trigger rules and services. The service trigger rules are used to judge whether each service is triggered; Acquired by the softswitch entity in NGN and the service call session control function S-CSCF entity in IP multimedia subsystem IMS; the engine management unit is used to provide various scripts including Python and BPEL and website language execution engines including JSP and ASP And its corresponding management interface, these website language execution engines are used to support the corresponding business, and the JSP engine is used for the platform to support Internet JSP web page business; The statistical monitoring and alarm module is responsible for monitoring the running status of the execution engine: counting the number of currently surviving business instances of each business, and periodically sending the statistical data to the business execution space management agent unit; periodically generating heartbeat events, and then packaging them into execution tasks and adding them to the The head of the execution engine team enables the execution engine to periodically send heartbeat data to the business execution space management agent unit to realize the heartbeat mechanism; provides alarm tools for each component of the execution engine, and transmits the generated alarm information to the upper management node when an exception occurs. 5. The cloud platform according to claim 4, characterized in that: the service instance management unit in the service instance manager maintains and manages the service instances by maintaining the service instances and their corresponding ones that exist in the service execution engine. Business session ID mapping, to support stateful conversational business; that is, when each business request event arrives from the underlying protocol stack adapter, if the business instance management unit already has a session ID corresponding to the request event business instance, the request event is forwarded to the event queue maintained by the business instance, and the business instance is encapsulated into a processing task TASK and delivered to the business execution core unit; otherwise, a new business instance is created, and the The new business instance and its business session ID mapping will be maintained in the business instance management unit in pairs; each time a complete business process ends, the business instance will notify the business instance management unit to delete the business instance itself and its business session ID map. 6. The cloud platform according to claim 1, characterized in that: the functions and structures of the resource access management unit, the data storage management unit, the resource virtualization unit and the resource access management unit in the cloud platform resource pool are; The resource access management unit is used to establish a connection with each server in the form of the latter registering with the former. The server resources accessed include two types: computing server resources and data storage server resources; the distinction between these two types of resources is only logical Above, it represents a certain computing or storage function, not a physical concept, that is, the same physical server host can be both a computing server resource and a data storage server resource; this unit can obtain the ability to access the local resources of the server through registration , and the connected server resource and the unit send and receive heartbeat messages to ensure that the unit performs real-time monitoring of the running status of the server node; The resource virtualization unit is used to virtualize the resources of the resource access management unit. Through the description of access resources and the encapsulation of access capabilities, it is abstracted into unified cloud platform computing resources and data storage resources to facilitate resource management. and the use of resources by the upper interface; The resource access management unit is used to respond to the computing resource application request issued by the business execution space management module, and allocate virtualized computing resources; and adopt a redundancy strategy when allocating, and store virtual machine files in one primary address and multiple backup addresses, When the server at the main address is not working properly, the server at the backup address runs the virtual machine file to ensure that the service will not be stopped due to problems with individual servers; The data storage management unit is used to provide the database cloud storage capability of the cloud platform; this unit uses virtualized storage resources to respond to the request of the data storage access interface of the platform, and redundantly saves and reads the database content in blocks; this unit also Save the index information and write protection lock of each block data to ensure that the data can be read and written successfully, and at the same time modify the record information according to the index, and regularly synchronize the backup data. 7. A working method of a cloud platform supporting converged network services, characterized in that: comprising the following steps: (1) Deployment and startup of the cloud platform: after adopting the multi-server cluster distributed structure to deploy the cloud platform, start the cloud platform according to the sequence from the bottom layer to the upper layer of the cloud platform architecture components; this step (1) includes the following operations: (11) Start and run the service program of the cloud platform resource pool on some servers first, so that these servers form a cloud platform resource pool server cluster, and then let the cloud platform resource pool server cluster open the access interface of the resource pool resources, and wait Access to underlying hardware resources; (12) Select a large number of physical servers at the three levels of cross-region, cross-computer room and cross-rack, and start and run the underlying resource service program, so that these physical servers form the underlying resource server cluster; if the underlying resource server runs computing resources If the proxy program is used as a computing resource server, it will register and access to the resource pool server; if the data storage agent program runs on the underlying resource server, it will register and access as a data storage resource server; (13) The cloud platform resource pool server cluster performs real-time monitoring on the connected underlying resource server cluster, and virtualizes the connected server capabilities to facilitate the dynamic adjustment of resources and serve as virtual resources in the cloud platform resource pool Provided for use by upper-layer services; due to the redundant storage of physical resources and the ability to divide them into small blocks, it can increase the availability of the cloud platform and significantly improve the utilization of physical resources; (14) Start and run business execution and management service programs on some servers, so that these servers form a business execution and management server cluster; after the business execution and management server cluster is connected to the cloud platform resource pool server cluster, obtain the cloud The virtual resources used by the platform itself, and then initialize the resources including system data; (15) Start and run the service program managed by the cloud platform on some servers, so that these servers can be used as cloud platform management server clusters, and after connecting the business execution and management server clusters and cloud platform resource pool server clusters respectively, perform corresponding initialization ; (16) After the cloud platform operator administrator performs the initial configuration operation on the entire cloud platform system, the access interface is opened to the cloud platform users, and the cloud platform startup is completed; (2) The cloud platform executes platform user registration, resource application, and loading and deployment of new services: after the cloud platform starts up and receives the platform user's registration and purchase of the required resources, it creates resources for the platform user according to the number of resources ordered by the platform user. Its unique virtual business execution space, and create a business instance including a business execution engine and a load balancer for it; this step (2) includes the following operations: (21) Cloud platform users including service providers or network operators access the user registration function provided by the user management unit of the platform management module, fill in user information, and complete registration; (22) Platform users order various cloud platform resource capabilities including CPU resources, data storage resources and network bandwidth from the resource allocation and scheduling unit of the cloud platform according to their own needs; (23) The resource allocation scheduling unit sends a request to the execution space management module to create a business execution space, and the execution space management module applies to the platform resource manager for the virtual resources required by the platform user, and creates a business execution space and business execution based on the resources. virtual machine instances in the space; (24) The execution space management module sends a virtual machine file backup request to the platform resource manager, and the platform resource manager performs regular backup and synchronization according to the redundant backup strategy; so far, the cloud platform completes the platform user resource ordering operation; (25) The platform user accesses the business management unit of the platform management module, and uploads the business to be deployed on the cloud platform; the business management unit sends a service loading request to the execution space management module; (26) The execution space management module responds to the service loading request, creates a corresponding number of service execution engine instances, loads the business and returns a successful loading response; (3) The cloud platform performs business-related operations and dynamically adjusts resources according to the platform user's instructions: during the operation process of the platform user using the cloud platform, the cloud platform dynamically adjusts the resources used by it according to the actual needs of the business; this step (3) includes The following operations: (31) After the cloud platform completes the service loading, according to the instructions issued by the platform user through the service lifecycle management interface of the service management unit, perform service activation, deactivation or offline operations; (32) After the business has been running for a period of time, if the platform user finds that the existing platform resource capacity is not enough to support the load of business access or needs to reduce the rented platform resources, he will log in to the resource allocation and scheduling unit to reorder resources and dynamically adjust the number of required resources ; (33) The resource allocation scheduling unit sends a resource re-application request to the execution space management module; the execution space management module applies to the platform resource manager for adjusted resources, and changes the configuration of the original business execution space, and adjusts the number of virtual machine instances and business The number of execution engine instances; so far, the cloud platform completes the dynamic adjustment of resources according to the platform user's request; (4) The cloud platform supports the operation of converged network services: after the platform users who register and purchase resources upload and activate the service, the service is in the activated running state on the cloud platform and can accept traffic from various operator networks outside the cloud platform. business request; and during the operation of the business, the cloud platform can simultaneously support three different network resources and related protocols of the Internet, telecommunication network and broadcasting network involved in the business; this step (4) is processing the business request reported by the network Or when a business request is initiated by the business itself, the following operations are included: (41) The service request arrives at the load balancer, and according to the service trigger strategy, the service request is forwarded by the load balancer to a specific service execution engine instance; (42) The corresponding protocol stack adapter of the business execution engine instance receives the forwarded business request, analyzes and processes the business request, encapsulates it into a platform internal message event, and sends the internal message event according to the subscription relationship of the business to resources to the corresponding business instance management unit; (43) The service instance management unit searches for the corresponding service instance according to the session identifier in the message event, if found, then sends the message event to the service instance for business processing, otherwise a new service instance will be created and linked to the message The session ID in the event is saved together in the business instance mapping list of the business instance management unit itself; (44) The message event is sent to the event queue of the corresponding business instance. The business instance is encapsulated into a Task and placed in the task execution queue in the business execution engine for processing. For stateless business, after the business processing is completed, the business instance The service instance management unit will be notified to delete the service instance and its corresponding session identifier from the service instance mapping list; for stateful services, the service instance will still be stored in the service instance mapping list, waiting for subsequent service requests; (45) The follow-up service request of the stateful service will be directly sent to the corresponding protocol stack adapter of the corresponding service execution engine. After step (42), the follow-up service request is parsed into an internal message event and sent to the corresponding service instance management unit. Through the session ID in the event, the business instance management unit finds the corresponding business instance for corresponding business processing; after the stateful business instance processes the entire business logic, it notifies the business instance management unit to delete the business instance itself and its session ID mapping ; (46) The system processing result of the service is transmitted to the corresponding protocol stack adapter and sent to the network. 8. The method according to claim 7, characterized in that: in the step (4), when various services for different networks are executed in the service execution engine of the cloud platform, the cloud platform will Features, adopting different protocol stack adapters and corresponding protocols to support different functions for business status.

Images