For your own free email subscription, send mail to
sub@objectwatch.com saying:
subscribe your-name,your-eMail
For example,
subscribe Roger Sessions,roger@objectwatch.com
To unsubscribe, send mail to sub@objectwatch.com saying
unsubscribe your-name,your e-mail
We are talking major hoopla. On September 9th, the OMG (Object Management Group) and bigwigs from BEA, IBM, and several other companies, "rallied to support and officially announce new CORBA 3.0 initiatives at a group-sponsored event... CORBA 3.0 will achieve a new level of capabilities for both experienced distributed computing programmers and less technical business programmers/analysts. The new version of CORBA will simplify the use of CORBA ORBs for the development of distributed object applications and include features that will facilitate CORBA's adoption in the enterprise." (Ref. 1)
Now it turns out that the presence of smoke doesn't necessarily imply the presence of fire. As we read through the press release, we realize that this major rally is not being held in honor of some new technology, but only in honor of the possibility that the OMG may someday in the future agree to a new CORBA specification, and that some companies may, even further in the future, actually implement that technology and that this new technology (when and if implemented) will offer the anti-Microsoft camp a new hope of stopping the Microsoft juggernaut. So not only doesn't the presence of smoke necessarily imply the presence of fire, it doesn't even necessarily imply the presence of more smoke.
Nevertheless, I am intrigued by the statement that CORBA 3.0 will "include features that will facilitate CORBA's adoption in the enterprise." I have been arguing for some time that the major problem with the current CORBA specification is that it offers no architectural standard for the middle tier (Ref. 2). The OMG's answer to this has been to attack my credibility without answering any of my criticisms (Ref. 3). This new press release made me wonder if the OMG was finally admitting that it had problems on the middle tier and was ready to address them.
Without a middle tier architecture, the OMG was essentially conceding the middle tier to Microsoft. Microsoft has a very effective middle tier architecture that I have been calling the Microsoft Distributed Component Architecture (MDCA) (Ref. 2). The heart of MDCA is COM, the Microsoft component model and MTS, the oddly named Microsoft Transaction Server. MTS provides a component runtime environment that makes use of highly efficient Transaction Processing Monitor (TPM) algorithms to host scalable n-tier component based applications. COM and MTS are rapidly merging into COM+.
Microsoft's COM/MTS as a combined architecture have been available since 1996. Sun responded to this threat with Enterprise Java Beans (EJB), which was finalized as a specification only a few months ago. Although EJB uses different language than does COM/MTS to describe what it does, it provides similar functionality. The major difference is that EJB is paperware whereas COM/MTS are actual products. But even as paperware, Sun was well ahead of the OMG, which was still refusing to admit that it had problems on the middle tier . The industry speculation has been that the OMG would eventually adopt the EJB standard. Bob Orfali and Dan Harkey, two of the best CORBA writers around, were adamant about this in their most recent book (Ref. 4).
Over the last two years the OMG, which started out life as a language neutral organization, has become increasingly dominated by the Java camp. Given that the Java camp was pushing EJB, it seemed likely the rest of the OMG would eventually follow suit. Based on this and on Bob Orfali and Dan Harkey's predictions, I have been assuming that the OMG would resolve its middle tier problems by adopting the EJB specification.
So I searched the news release for some mention of EJB or EJB-like functionality. Many capabilities of CORBA 3.0 was pre-announced, including a "plug-and-play framework" based on JavaBeans, a scripting language, mappings that would allow Java programmers to completely avoid having to work in OMG's Interface Definition Langauge (IDL), and several other useful capabilities. But not even a hint of EJB or any other middle tier technology.
It was hard to imagine that the OMG would release CORBA 3.0 without addressing this fundamental deficiency. I called several friends that were well connected with the OMG. Is EJB part of CORBA 3.0 or not? Nobody could give a definitive answer. In fact, the OMG bureaucracy seemed unable to differentiate between a client-tier component model, such as JavaBeans, and a middle-tier component model, such as EJB.
I had finally concluded that the OMG was going to ignore its middle tier problems when I was lucky enough to make contact with Larry Cable. Larry is a Senior Staff Engineer with the Java Software Division of Sun Microsystems. He was a co-architect of JavaBeans and very involved with Enterprise Java Beans. He is one of the people actually writing the CORBA Component Model which will presumably become part of CORBA 3.0.
Before I give you the scoop (from Larry) I need to explain how the OMG works.
The OMG works slowly. The entire process of finalizing an OMG specification usually takes over a year. The process of finalizing the OMG CORBA Component Model specification started back in June of 1997 (Ref. 5), when the OMG issued what is called a Request For Proposals (RFP). This RFP (Ref. 6) described what the OMG thought it was looking for in terms of a Component Model. It is noteworthy that no place in this RFP does the OMG hint that it is looking for a middle tier component model, ala EJB. The RFP focused only on client side issues, ala JavaBeans.
Companies were given until November 10, 1997 to prepare initial submissions for a CORBA Component Model. Seven submissions were sent in. The most noteworthy submission was a joint submission from nine companies: BEA, ICL, IONA, IBM, Netscape, Oracle, SunSoft, Unisys, and Visigenic. Compared to this nine company submission, the other submissions were insignificant. The nine companies included all the true power brokers of the OMG.
According to OMG rules, once the initial submissions are received, a second phase starts. In this phase, companies that submitted in phase one are allowed to review each other's submissions and prepare new submissions based on their new enlightenment. The OMG deadline for phase II submissions was set as July 6, 1998, but was then extended to October 19, 1998. This gives you a sense of how fast things move at the OMG - almost a full year was allowed to elapse between the phase I and the phase II submissions (not to mention from when the initial RFP was released).
The nine company phase I submission (Ref. 7) was a 78 page specification that completely focused on client side issues, as did the original RFP. So as early as ten months ago, the OMG powers that be were continuing to ignore the middle tier.
Now we can pick up the story from Larry.
Larry tells me that the original nine company submission was based on the RFP requirements, which indeed dealt only with client side issues, such as the use of component composition tools to put together client side systems. Soon after that November submission was completed the EJB specification was finalized. EJB started generating excitement in most of the non-Microsoft world. The nine companies realized that even though the original RFP did not deal with EJB issues, these issues had to be dealt with.
So soon after the phase I submission was completed, the nine companies started working on phase II, which is being completed as we speak. At this point, there are no public OMG documents that we can look to for guidance as to how close the final specification will be to EJB. Larry says that the final specification will be very close, and he should know, if anybody should.
According to Larry, the goal of the final submission will be to allow Enterprise Java Beans components to be treated as CORBA components. The focus is on interoperability, allowing a client to work with components running in different CORBA middle tier environments without regard to which vendor's environment is hosting that component. The specification will not really deal with the issue of portability, and therefore will not allow a component to move from one environment to another without source code changes. Larry agrees with my assessment that portability is probably even beyond the scope of EJB. Ultimately, he feels we need to look to Java to provide real portability.
Of course, my own view on Java portability is closer to Microsoft's, that portability is ultimately less important than functionality. I discussed this perspective in a previous newsletter(Ref. 8).
Although Larry is very close to the work being done on the CORBA Component Model, even he can't predict what will eventually happen with the OMG. Several other companies still have the right to submit final proposals when the October deadline occurs. Eventually a subset of the OMG will vote on what the actual specification will be. They could, in theory, decide to accept one of the other submissions or even no submission at all. However, given the big names associated with this particular submission, it is a good bet that the nine company submission will be accepted. In fact, of the 50 odd companies allowed to vote on this particular specification, almost 20% are co-submittors of this proposal (Ref. 5).
Under the most optimistic scenario, we still have a ways to go before we see any products based on CORBA 3.0. I estimate the following time line:
Of course, Microsoft is not standing still while the OMG tries to reach consensus on the middle tier. By the time its competitors come out with CORBA 3.0 products, Microsoft will have a 4 year head start with COM and MTS. NT 5.0 will probably have had a full year to solidify its hold on the middle tier market. CORBA 3.0 at that point will mainly be of interest to companies running on non-NT boxes; companies committed to NT will buy into the middle tier architecture based on COM and MTS, which, by then, will have evolved into COM+ and will be tightly woven into NT 5.0.
In the meantime, CORBA 3.0 may actually hurt companies trying to compete against Microsoft. All of Microsoft's major competitors will be committed to supporting CORBA 3.0. Many customers will avoid buying into a product knowing that that product faces significant retooling over the next 14 months. Most customers would prefer to either buy into an architecture that is already stable and tested, like Microsoft's, or wait until the retooled product is released and stabilized, which will probably not occur until the year 2001.
There is one point that works in favor of the CORBA 3.0 based products. Although it is true that Microsoft has a four year technological head start over CORBA 3.0, from a marketing perspective Microsoft has failed to exploit this lead. So while Microsoft has very competitive middle tier technology, Microsoft has done a lackluster job of explaining its middle tier architectural vision to its customers. If Microsoft doesn't take steps to correct this problem it will lose much of its potential advantage.
We can be sure that companies like BEA will be singing their songs to anybody that will listen. Microsoft needs to do the same. It is not enough to have great technology; one must also be able to explain what that technology is and what value it brings to the world.
There is some evidence that Microsoft is getting its marketing act together. Microsoft has a new group dedicated to marketing its distributed architecture. Microsoft recently introduced a new semi-official term ("Microsoft Component Services") to describe its component-based technologies, such as COM, MTS, and MSMQ (Ref . 10). This term seems similar in concept to the term MDCA (Microsoft Distributed Component Architecture) that I have been using. However Microsoft has still not articulated how its "Component Services" relate to the middle tier and why the middle tier is important to corporate computing. But it may, at least, have taken a first step.
It will be very interesting to see how Microsoft positions things at the upcoming Professional Developer's Conference. I, for one, would like to see some major hoopla around its middle tier architectural vision. This is one place where we already have the fire, and we could use a little more smoke. Microsoft can definitely learn some lessons in the smoke department from OMG, while the OMG can learn a lot about fire from Microsoft.
- Roger Sessions
Austin Texas
September 26, 1998
*I am indebted to the Object News website (http://www.objectnews.com) for discovering this white paper, yet more proof of the great job done by this indispensable site.
We are offering public classes in the Microsoft Distributed Component Architecture between October 1st and October 5th in Stockholm, Copenhagen, and Oslo. Roger Sessions will be teaching all of these classes. This is a wonderful opportunity to learn about Microsoft's middle tier architecture and how that architecture stacks up against CORBA and Enterprise Java Beans directly from the world's leading expert in this field. The cost of attending this one day class is $550 (USD, or equivalent in local currency) per seat.
If you are interested in attending one of these classes, give Janet Van Sickler a call. She can be reached at the following hotels on the following dates:
If she is not there, leave a message, and she will call you back.
Our class schedule is as follows:
When you arrive, ask for the ObjectWatch Class.
If you decide to come at the last minute, just come to the appropriate hotel. We can't promise to take you (our class size is limited to 25), but if we can, we will.
Mr. Bunny's Guide to ActiveX
By Carlton Egremont III
Published by Addison Wesley
ISBN 9 780201 485363
Mr. Bunny's Guide to ActiveX is the most original computer book to come out in the last five years. Whatever prompted Carlton Egremont III to write this book, or Addison Wesley to publish it, or me to review it is beyond me. But this book is destined for cult classic status.
If you have ever thought of working with ActiveX, or have ever heard of ActiveX, or have never heard of ActiveX, or wish you had never heard of ActiveX, this book is a must read. Frankly, you will learn almost nothing about ActiveX. But you will find out exactly what happened to Mr. Bunny when Farmer Jake decides to modify The Registry, and in the overall scheme of the universe, that is much more important. This is the gift of the year for your favorite computer geek. Do not miss it.
- Roger Sessions
Roger Sessions will be signing copies of his book COM and DCOM; Microsoft's Vision for Distributed Objects at the Microsoft Professional Developers Conference (PDC) in Denver on Monday, November 12 at 7:00 PM at the PDC Book Store. Please drop by and say hello. If you already have a copy of his book, bring it along. He would be happy to sign it. If you don't have one, they will be for sale at the PDC.
ObjectWatch has seats available for Roger Sessions' Introductory Class to Microsoft's Middle Tier Architecture. This one day class is a must attend event for anybody building products or evaluating using COM, DCOM, MTS, and the other Microsoft middle tier products.
If your company is trying to decide between a Microsoft, CORBA, or Enterprise Java Beans strategy, then this class is made from you. This is your opportunity to hear about Microsoft's middle tier technologies and how they compare to these other industry standards in an intimate classroom setting directly from the world's leading expert in this key area. No hype, no industry propaganda, just real information and unbiased appraisals.
We are currently accepting registrations for the following cities:
Seats are $550 each. Please sign up as soon as possible. Seating is limited to 25 seats, and often sell out.. For more details on these classes, write janet@objectwatch.com.
* For the Nordic classes, see the related article in this newsletter.
It's always a special treat to have ObjectWatch Newsletter readers join me at public speaking engagements. Some of my upcoming engagements include the following:
ObjectWatch specializes in middle tier architectures, with emphasis in the architecture we call MDCA, (Microsoft's Distributed Component Architecture). This architecture includes six key technologies: COM; DCOM; MTS, DTC, MSMQ, and MSCS. These technologies are the basic building blocks needed to create highly scalable, component oriented applications using NT as the middle tier.
Roger Sessions teaches all of the ObjectWatch classes. He is the author of COM and DCOM; Microsoft's Vision for Distributed Objects, published by John Wiley & Sons. This book is a gentle introduction to the three tier scalable distributed object architecture of Microsoft. This book can be purchased at your local computer book store, or on-line at http://www.objectwatch.com/bookstor.htm.
Roger Sessions is the President of ObjectWatch, Inc. He is the author of four books and dozens of articles. He has extensive experience with both CORBA and Microsoft middle tier architectures. He is a frequent conference speaker on middle tier architecture. Visit the ObjectWatch web page at http://www.objectwatch.com or contact Roger Sessions at roger@objectwatch.com for details on how
Recent issues of this newsletter are available at http://www.objectwatch.com
The ObjectWatch Newsletter does not accept advertising or rent out its subscription list.
This newsletter is Copyright (c) 1998 by ObjectWatch, Inc., Austin, Texas. It may be freely redistributed provided that it is redistributed in its entirety, and that absolutely no changes are made in any way, including the removal of these legal notices.
ObjectWatch is a registered trademark (r) of ObjectWatch, Inc., Austin, Texas. JavaBeans and Enterprise JavaBeans are trademarks of Sun Microsystems. CORBA is a trademark of OMG. All other trademarks are owned by their respective companies.
Return to the ObjectWatch home page.