Open Hypermedia Systems Working Group

DCOM

DCOM Recommendations

In Ken Anderson's evaluation of CORBA, three current requirements of the Open Hypermedia Protocol were identified. These are the ability to:

• specify a set of related components
• develop and run the components cheaply (e.g. for free) and on multiple platforms.
• exist in an Internet environment (e.g. with associated scale and security issues)

DCOM fulfills all these requirements. Firstly DCOM is an extension to COM, a tried and tested technology with a very large market penetration. COM's ability to specify a set of related components and use them reliably, is proven. It is firmly integrated with both NT and Win95, and all MS products (and most non-MS business products) now use COM in some shape or form. There are thus a lot of skilled programmers who are used to working with COM. If OHP is to become a successful standard this large skill base is worthy of consideration.

The second point is an interesting one, both COM and CORBA fall slightly short. While CORBA may yet prove expensive to develop, DCOM has yet to prove that it is truly multi-platform, although it is certainly very cheap to develop. Multi-platform versions of DCOM are an important objective and work is progressing on several UNIX variations (with Solaris already published) through Software AG's EntireX. Further investigation is needed to prove that these systems truly interoperate. It is certainly proven that COM applications can be developed in a variety of languages, with packages from many venders readily available..

Lastly, again similarly to CORBA, DCOM does address the internet issue but is not proven. DCOM is specifically designed to meet this challenge and I think there is no doubt as to its ability to use internetworked components. How well it does this and with what efficiency is yet to be seen as I known of no large distributed systems yet developed with DCOM.

I think that the biggest argument over this framework will be based on Microsoft itself. It is DCOM's biggest strength and greatest weakness. Being attached to such a large industry force means that DCOM will be used extensively. Even if only half the machines that run a Microsoft operating system used DCOM, the other half using some other framework, it would still mean that one third of all the machines in the world used DCOM above all other frameworks! The probability is that DCOM will become standard on virtually all the Microsoft platforms, based on current figures that's a 70% market share!

Unfortunately there is also a downside. Firstly COM as a technology is moving very quickly. Already Microsoft has announced a new standard COM+, although details are sketchy at the present. Designing a hypermedia framework for the future based on technology that could be completely revised within five years is obviously going to cause problems. Any OHP DCOM design would have to move with the technology as well as evolve as a system in its own right.

In summary the OHS WG should consider the following issues when choosing DCOM as the component technology for the OHP. DCOM is a very strong technology with large market support. It is as flexible and as powerful as CORBA and is constantly evolving. But it is a Microsoft technology. Design-wise, bound to Microsoft systems. Cheap and easy to develop on those systems, with support for other platforms through EntireX. We must balance the probability of a CORBA future, multi-platform, multi-vender. With the DCOM future, where Windows is the dominant platform and where the native DCOM implementation is superior to other frameworks on Windows machines.