DataSpiderServer Architecture
The following sections discuss the key architectural concepts of DataSpiderServer.
Architectural composition
DataSpiderServer is a Java Application. Java application programs are executed inside a Java Virtual Machine, which is in turn executed inside an Operating System. From DataSpiderServer point of view, the JVM together with the OS, represents the execution environment where as from the OS point of view, it is just an instance of an application program.
Architectural Diagram of DataSpiderServer
Kernel and Manager
DataSpiderServer's architecture can be broadly separated into the kernel and the managers.Kernel is the core component of DataSpiderServer and is responsible for starting/stopping and registering/unregistering of each manager.
Each manager registered to DataSpiderServer controls its own service whose functionality are narrowly aligned with its responsibility.The most important ones are listed below, however there are other managers that are not described here.
Trigger Manager
Trigger Manager controls service that performs trigger management and operations, which include registering, modifying, deleting, starting, and stopping of triggers.
User Manager
User Manager controls service that performs user and group accounts management and operations, which include registering, modifying, and deleting of user and group accounts.
FileSystem Manager
FileSystem Manager controls service which performs file and directory manipulation such as viewing their attributes, copying and deleting them.
Data Processing Connection Manager(DPCon Manager)
FileSystem Manager controls service which performs file and directory manipulation such as viewing their attributes, copying and deleting them.It also controls resource pool.
Advantages of Managers
Each manager is a self-contained service provider and its implementation is concealed.
Services provided by managers are accessible and can be utilized only through kernel which controls them, which will yield advantages as described below.
- Loose coupling
As each service is self-contained and controlled by its distinct manager which exists independently from others,modifications can be as small as adding new services when scaling out the DataSpiderServer with new features.
- Transparency
Services are transparently accessible via kernel by adapters and other managers without knowing the specific details of their implementation.
For an example, when an adapter utilizes the DataSpider filesystem, all file system operations are performed through FileSystem Manager and the adapter does not need to beware of whether it is communicating with the local file system(platform dependent filesystem) or database filesystem.
Specification Limitations
- DataSpiderServer has a built-in servlet container to be used with its system. It is not intended to be used to run other web applications.