Virtualization Grid Computing

The concept of the grid deals critically with the virtualization, which means that at all levels there exist a layer of the abstraction as to what is requested and what is provided(service).Grid computing makes it possible to create a single IT infrastructure that can be shared by multiple business processes.

This can be better  understand by considering an Oracle"s Grid environment, where the end users as for an application service and the grid works out as to which clustered J2EE application server can best provide it.

Consider an Oracle application server model in which an Application server asks for the database service from an RAC database and Grid work out from which RAC instance the data service can best be provided. Within the Grid there is indexing of all the possible services to available service provider and the algorithms for assinging the worload and resource appropriately. Due to this the end user neither needs nor have the capacity to know from where the computing resources are actually being provided.

The Grid is not only exclusive to Oracle but at the physical level, some O.S and hardware vendors have the capabilities to partition the server into virtual machines, which can be alloted RAM and CPUs dynamically according to the demand. Thus it is conceptually similar to the Oracle"s approach of dynamically assinging application server and database server resources. Thus there is no reason for which the two approaches can"t be combined and work towards the same goal and can work together. The result of a such a approach would be an environment where adequate resources will always be available on demand, without facing the issues of excess capacity or underperformance at times. It should be possible to design a Grid environment with no single point of failure, and thus achieving 100% uptime, which is demanded by many organizations.

The Oracle"s Application Server

Understanding Need
The need for application servers has come with the potential realization of the Web as the standard platform for delivering the applications to end users. An application server has replaced the client-side software which is traditionally installed on end-user computers: it runs applications centrally on the server and presents it to users by displaying it locally in web browsers like mozilla, internet explorer. These applications make use of the data stored in one or more database servers.

 
Model
Oracle Application Server provides a platform for developing, deploying, and managing web applications. A web application is any application with which
users communicate via HTTP. Web applications usually run in at least three tiers environment:First is a database tier which manages access to the data, the client tier which is often implemented as a web browser and handles the local window management(web browser here) for communications with the users, and an application tier in the middle of the two which executes the program logic that generates the user interface transferred to the browser of user and the SQL calls to the database. This is the model by which corporate companies provides its services to the users over internet and thus you can realize  the need of an application server.


Advantages, Conforming standards and contribution to grid
For developing Web applications there are a number of technologies, among which most dominant today is Java. Applications which are written in Java should conform to the J2EE (Java 2 Enterprise Edition) standard, which defines how the java applications should be packaged and deployed. J2EE standards are controlled by Sun Microsystems(currently takeover by Oracle corporation) and accepted virtually by all major software developers. Oracle Application
Server is a fully J2EE-compliant application server. Oracle’s implementation of the standards has allowed the automatic load balancing and fault tolerance across the multiple application servers on multiple machines though J2EE clustering. The clustering  virtualizes the provision of the application service i.e. users ask for an application, which may be available from a number of locations, the cluster works out from where the session or request can best be serviced. If one location fails due to some error, others can take up the load, and thus the more resources can be made available for an application as necessary. Oracle Application Server’s has majorly contributed to the Grid with the ability to separate the request for a service from the location provision and adding or removing the J2EE servers from a cluster dynamically.


Global Approach
 Oracle has commitment to the international standards i.e. Applications which are running in the Oracle Application Server environment can connect to any databases which has the  Java-compliant drivers which means it is not necessary to use an Oracle database only. Applications which are developed with the Oracle Application Server toolkits can be deployed to a third-party J2EE-compliant application server. Although, the Oracle product set is particularly powerful and will often be the best choice.


Connection pooling model
For the connection with the database generally connection pooling model is used instead a one-for-one mapping of end-user session as it is more efficient. In connection pooling, the application server establishes a small number of persistent database sessions and makes them available on demand (queuing requests if necessary) to a relatively large number of end-user sessions against the application server.