This paper takes a look at some of the pertinent vendor strategies of the day (Early 1996) and is extracted from the ATSI Vendor Strategy seminar which of course has more up-to date information.
A new paper will be published shortly which outlines how the vendors have altered their strategies substantially since this paper was published.
This paper is provided with the complements of ATSI for clients and colleagues. If you require information or assistance regarding large scale distributed systems design and products contact ATSI at (408) 476-4300 or use the form provided at the end of this paper.
While many computer vendors had set their eye on capturing market share from IBM and other large legacy organizations, a new rapidly growing market emerged that caught the vendor strategists by surprise.
The market was in fact already there, however, all that was needed was a killer application that would catch the eye of the public to trigger exponential growth. This is typically the case in many markets as the birth of a new industry grows at the expense of another through innovative new ways of doing business. In Internet's case, the Web Browsers brought new meaning to the Web in an easy to use interface. Surprisingly, this rapid change in computer systems direction caught many vendors off guard. Including the now not so new upstart in Redmond Washington, that had so effectively captured market share from megaliths like IBM. In a move that could be paralleled with IBM's missing the boat on PC and UNIX products, Microsoft the megalith of the day now like IBM had found themselves in a similar position.
What is it that causes large organizations with hundreds of millions of dollars at their disposal to fail with their strategy?
1995 could almost be viewed as the year of the Web. Finally after many years Internet was recognized by general industry and the public at large as being a key component to doing business. The interesting fact is which computer vendors were taken by surprise. Neither Microsoft nor Lotus had any Internet related products. IBM was not really interested in Internet, still trying to push Open Blue Print et al. SAA (SNA diluted). DEC was still trying to find itself again having struck a close relationship with Microsoft thorough the David Cutler "loss" (David a key architect of VMS, moved to Microsoft to develop NT). HP hot on IBM's heals with legacy emulation and investing heavily in DCE, like most of the rest of the pack, appeared to consider Internet a haven for geeks, hackers and other undesirables. Sun, probably more by natural progression rather than calculated strategy, was there in the midst of the Internet market with more than 50% of the Internet server market. Whether Sun can capitalize on this position is however another question.
What we now observe is every major vendor scrambling to develop Internet based technology, lest they be left behind in the Cyberspace dust.
All to often we tend to focus on the immediate market around us an ignore trends in parallel markets. Trends that all to often overtake an consume existing markets. These trends often leave individuals without work and wondering what happened to their career's. It is imperative that one keeps track of the industry at large, and not only in ones own segment. Not meant as a plug for our ATSI seminars, but these types of activities can be invaluable in taking staff out of the cycle of their busy lives and throwing them in front of an external expert that will show them the world through new eyes. Simply walking the isles of trade shows can completely alter product development, or visiting your local computer store, can often save projects that would be doomed to fail due to their outdated technological foundations and competing products (What strategy vendors have employed from a historical perspective, and what they can do ensure that they stay on top of things is covered in the ATSI Vendor Strategy ).
The rapid rate of change in technology is exemplified in the rate of change in microprocessor transistor densities:
Number of Transistors Found in Intel Microprocessors
It is this phenomenal rate of change that is changing the way corporations need to develop future corporate and product strategy.
With the incredible slew of products emerging for the Internet, it is unlikely that the sudden momentum behind the web will slow down. We will however certainly see attrition in segments of the web field. Visions of high speed interactive TV quality multimedia and 3D graphics are still a long way off due to the relatively slow performance provided by the massive Internet WAN, even private networks still lack much of the bandwidth, or at least the cost effective bandwidth needed to support these bandwidth guzzlers. The fact remains that there is no more effective way to send a written message than the Internet, while the marketing features of being able to provide up to date brochures at a fraction of the cost of paper products, all without cluttering your customers office, will ensure the continued rapid growth of the web even putting aside the other potentially hot Internet markets.
The next interesting area which is starting to fuel the Internet's growth is in the transaction orientated business and the so called EDI market. This market brings businesses wholeheartedly into the clutches of the Web. While many businesses have for some time depended on the Internet for most of their electronic mail, the Web Transaction Environment brings the core business functions of organizations into the Web, thus making the Web an integral part of the business.
When mentioning transactions, anyone who has been around the industry for the last five to thirty years will probably start thinking of Transaction Monitors like CICS, TPF, TopEND and Tuxedo. This is not surprising since IBM claims that CICS alone is installed in More than 20,000 enterprises, in 90 countries with about 50,000 server licenses in force. And that 90 plus percent of the Fortune 500 companies use CICS with and estimated billion lines of code. 85 % of all IBM mainframes run CICS while there are and estimated 300,000 CICS application developers! (If you want to find out more on this area and how to integrate TP monitors with the Web, attend the ATSI Distributed Systems Technologies Seminar). Things change, however, when one digs deeper into the new transaction orientated business and learns about the typical business requirements of these next generation transaction systems.
Many of these new transaction systems are required to support millions of users and many hundreds of different applications. Applications in these new systems may only last months and be required to be brought online days or a few months after they are conceived. Like their predecessor applications, these new transaction systems still remain mission critical and must meet the ACID test. In other words they still have to be built with just as much or even more resilience to the earlier transaction systems. One of the key difference between earlier transaction systems and the new ones being that we no longer have months or even years to develop the systems.
Transaction loads on these new systems can buildup instantaneously, where as in the past the transaction systems typically experienced a gradual or more steady increase in load. When one is faced with these types of problems, it becomes apparent that the typical Transaction Monitor design, which relies typically on one transaction manager routing transactions to resource managers, cannot sustain the loads and dynamic application requirements of new business systems. It is here that advanced systems architects, have had to come up with new more dynamic designs. These new designs (the topic of our next white paper) typically employ standard object oriented concepts such as those embodied in the CORBA OTS (Object Transaction Service) specification, holding objects as transaction and employing advanced techniques to create dynamic super-scalar architectures.
It is with these objects that the next religious war is being waged. Since without standards one object cannot effectively interact with another. There have been numerous different object systems found in many of the OO languages such as Small Talk, C++ etc, though proprietary CASE tools and Middleware products. There is now at last only two key object standards: CORBA objects and OLE objects.
Every major vendor except for Microsoft, fully endorses and wishes to promote the CORBA standard. However, Microsoft who has managed to capture the desktop, not unlike IBM in the early days, wants to force their own standard so as to maintain control of the market. This rightfully annoys most, however Microsoft like most other vendors, is not in business for charity. Microsoft, like any other typical vendor who has the opportunity to control the market, is going to do everything they can to accomplish their aims.
Microsoft had been effectively building their corporate future completely around OLE technology. "We are driving everything to this," said Jim Allchin, Vice President of Advanced Systems at Microsoft, referring to the company's OLE strategy. Microsoft strongly encourages developers to implement OLE objects, and promotes OLE as the correct way to develop applications even over the WIN 16 & 32 API. Even the window manager in Cairo, Microsoft's next OS implementation, is based on OLE objects. All had been going well for Microsoft, even though they do not yet have an commercial distributed version of OLE (referred to as Network OLE, Distributed OLE, Distributed COM (DCOM), and Enterprise OLE). There are far more applications supporting OLE than any other object technology, probably somewhere around the 1000 mark, far more than any off the shelf CORBA applications. CORBA, which is really an offspring of the UNIX vendors, has tended to target large scale corporate custom applications. Like many of the UNIX applications, the CORBA technology is complex, essentially requiring advanced C++ programming before you can create any useful application. There are also essentially no high level cut and paste interfaces to CORBA that are typical with OLE based applications. OpenDOC, the first real high-level interface to CORBA (part of CORBAfacilities), while representing a fine architectural concept and a well thought out object model for the desktop, is pretty much useless when it is not found in any applications.
Microsoft has the objects out there, they just do not have an efficient means of implementing them in a distributed object architecture, jet. CORBA implementations on the other hand, have excellent core technology for implementing distributed object systems that have been shipping for some time, but with basically no higher level environments that customers expect in modern systems both from an applications development and an interactive users perspective. Ironically, you can use some of the better CORBA ORB's to implement a distributed OLE implementation, but this is not the direction of OMG's CORBA, and it can't be since Microsoft owns OLE and CORBA is an open standard.
For Microsoft things were going well, sure they were not there yet there from a distributed OLE perspective, but the applications were already out there waiting for Distributed OLE (DCOM) to enable them. A smart strategy, inversely similar to Microsoft's strategy with Windows: Ship the product, the users already have DOS, when they load Windows they will need applications. Windows afforded Microsoft the opportunity to capture the applications market with Word and Excel. In addition to core OLE, Microsoft has been developing their implementations of OLE Transactions and OLE DB which would further bolster their control of the transaction in an enterprise and eventually hopefully capture the entire enterprise's systems.
This is where we get back to our Internet subject. From all the evidence, Microsoft had not accounted for the Web, this massive explosion came broadside, without Microsoft's expectation. All of a sudden upstarts like Netscape were being compared to the Microsoft giant! The Web is where the media attention is, an area Microsoft has been spending fortunes on to brand Microsoft into the minds of consumers. With Internet, Microsoft was certainly not the center of attraction, on the contrary they were heading off in their own direction with the proprietary Microsoft Services Network.
To make matters worse for Microsoft, Sun came out with Java which also captured the media's attention. While there is a tremendous amount of hype surrounding Java (Java being primarily a new implementation of an OO language with imbedded support for Web interfaces and a sounder architecture than many other OO languages including C++), Java does offer a key function that could seriously undermine all of Microsoft's efforts to build their future on OLE. Java does objects, and coupled with CORBA ORB's allows you to implement advanced distributed object implementations that are devoid of Microsoft technology. To make matters worse, unlike the CORBA geeks, the Java folk were developing real applications that were easy to use (a trademark the that made Apple successful and which is now fueling Microsoft's success - see Clive's Law).
The Microsoft monopoly was all of a sudden challenged, their core strategy could potentially fail. If users continued to whole heatedly adopt the is new World Wide Web Internet world and Browsers that did not even have a mention of Microsoft, with applications implemented in Java applets (objects) not OLE, Microsoft, like IBM could potentially loose the market!
So what is it that caused Microsoft, and IBM for that matter, to not anticipate the market movements more effectively? There are probably many reasons, but the most reasonable conclusion is what we mentioned earlier in this article. All to often we tend to focus on the immediate market around us an ignore trends in parallel markets. A curious feature of upstarts is that personnel tend to receive positions not because they are the best person for the job, but because they were just there first. The problem usually is not that these people are incompetent, but rather that the nature of the job prohibits them from having the time to look up and see where things are going (Yes… another plug for ATSI Seminars..)
To their credit, Microsoft is doing allot to get back on track with mainstream market direction (Internet enabled MS applications, IIS, ActiveX - a lighter weight OLE implementation for Internet, etc.). Whether Microsoft can hold out in the face of the rest of the industry and make OLE the defacto standard for Internet, or whether the market will adopt open CORBA based object technology still remains to be seen. What we will see for now is the high-end back-end systems supporting core CORBA technology and integrating OLE objects (ActiveX) along with Java applets and other objects on the desktop. The opportunity for vendors to develop strategies and products that capture this new market is where the fun begins.
ATSI Home - Seminars = Service = Strategy
Send mail to webmaster@InfoTelesys.com with questions or comments about this web
site.
Copyright © 1997 InfoTelesys, Inc.