Distributed Computing Challenges

The core of a distributed architecture is the ability to invoke a method on an object or communicate with services that reside in a different process and possibly on a different computer. Although this does not sound difficult, you must address a surprisingly long list of issues:

  • How do you instantiate a remote object?

  • If you want to invoke a method on an existing object, how do you obtain a reference to this object?

  • Network protocols transport only byte streams, not objects. How can you invoke a method over a byte stream?

  • What about security? Can anyone invoke a method on the remote object?

  • Most networks are inherently unreliable. What happens if the remote object is unreachable? What if the remote object receives the method invocation but cannot send the response because of network problems?

  • Calling a remote object can be much slower than invoking a local method. Do you want to invoke the remote method asynchronously so that you can continue processing locally while the remote object processes the request?

The list of questions continues. Fortunately, the features in the .NET Framework take care of most of these issues, allowing developers to create distributed applications without having to deal with many of the nasty details. These features make remote invocation almost transparent to the programmer, at least at the syntactic level. This simplicity can be deceiving, however, because developers still must understand some of the underlying principles of remote communication to write robust and efficient distributed applications. The Distributed Systems patterns cluster helps developers make informed design decisions when implementing distributed applications.

Enterprise Solution Patterns Using Microsoft. NET 2003
Enterprise Solution Patterns Using Microsoft. NET 2003
Year: 2004
Pages: 107

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net