To fully understand the importance of web services, you need to understand the requirements of distributed computing. Distributed computing is the partitioning of application logic into units that are executed on two or more computers in a network. The idea of distributed computing has been around a long time, and numerous communication technologies have been developed to allow the distribution and reuse of application logic. Many reasons exist for distributing application logic. Some of the most important include the following:
High scalability: By distributing the application logic, the load is spread out to different machines. This usually won’t improve the performance of the application for individual users (in fact, it may slow it down), but it will almost always improve the scalability, thereby allowing the application to serve a much larger number of users at the same time.
Easy deployment: Pieces of a distributed application may be upgraded without upgrading the whole application. A centrally located component can be updated without needing to update hundreds (or event thousands) of clients.
Improved security: Distributed applications often span company or organization boundaries. For example, you might use distributed components to let a trading partner query your company’s product catalog. It wouldn’t be secure to let the trading partner connect directly to your company database. Instead, the trading partner needs to use a component running on your servers, which you can control and restrict appropriately.
The Internet has increased the importance and applicability of distributed computing. The simplicity and ubiquity of the Internet makes it a logical choice as the backbone for distributed applications. Before web services, the dominant protocols were COM (which is called DCOM, or Distributed COM, when used on a network) and CORBA. Although CORBA and DCOM have a lot in common, they differ in the details, making it hard to get the protocols to interoperate.
High scalability: By distributing the application logic, the load is spread out to different machines. This usually won’t improve the performance of the application for individual users (in fact, it may slow it down), but it will almost always improve the scalability, thereby allowing the application to serve a much larger number of users at the same time.
Easy deployment: Pieces of a distributed application may be upgraded without upgrading the whole application. A centrally located component can be updated without needing to update hundreds (or event thousands) of clients.
Improved security: Distributed applications often span company or organization boundaries. For example, you might use distributed components to let a trading partner query your company’s product catalog. It wouldn’t be secure to let the trading partner connect directly to your company database. Instead, the trading partner needs to use a component running on your servers, which you can control and restrict appropriately.
The Internet has increased the importance and applicability of distributed computing. The simplicity and ubiquity of the Internet makes it a logical choice as the backbone for distributed applications. Before web services, the dominant protocols were COM (which is called DCOM, or Distributed COM, when used on a network) and CORBA. Although CORBA and DCOM have a lot in common, they differ in the details, making it hard to get the protocols to interoperate.
No comments:
Post a Comment