Chapter 5. Distributed Programming in .NET

I l @ ve RuBoard

Applications need to be able to communicate with each other. This has always been possible using technologies such as shared memory and network socket communication, but using these technologies in a way that preserves types and type fidelity has always been a challenge for developers. Subsequent technologies such as COM and DCOM have provided Windows developers with some of the necessary functionality, but COM and DCOM don't play well in the world of the Internet. DCOM, especially , has suffered because most network administrators restrict the number of TCP/IP ports that Internet traffic can use to get through their firewalls. Unfortunately, the ports used by DCOM usually don't make the cut for security and performance reasons, so DCOM applications rarely work across the Internet (at least not with corporate customers).

With the .NET Framework, Microsoft has completely revised its interapplication communication. In place of DCOM are .NET Remoting and XML Web services. XML Web services seem to get more press, but .NET Remoting is an equally formidable technology that solves certain problems that XML Web services simply cannot. In this chapter, we'll look at both of these technologies and explore the advantages and disadvantages of each.

You've seen the basic demonstrations of XML Web services and how they're supposed to change the Web. You might also have heard of .NET Remoting. But how should these services be integrated into real applications? How do you enforce permissions and referential security across the enterprise and over the Internet? These are among the questions this chapter will address.

We need to begin, however, with a discussion of object serialization. The topic of serialization is relevant to both XML Web services and Remoting and is crucial to understanding distributed interapplication communication.

I l @ ve RuBoard