AS/400 Strategies & Solutions Conference
by Rick Smith (May 25, 1999)
At this recent DCI conference, I attended the panel discussion about the AS/400 in a Date Warehouse and e-Business World.(Article follows the pictures).
Moderator: Wayne Madden, Publisher; NEWS/400
Panelist: Thomas Bittman, VP and Research Director, Server Strategies; Gartner Group
Panelist: Susan Whitney, VP Server Marketing; IBM
Panelist: Roger Pence, Editor; AS/400 Letter
The discussion about AS/400's future was rather heated as everyone (audience and other panel members) was asking Susan about the future of the AS/400 with respect to IBM's other server offerings (S/390, RS6000 and Netfinity). Susan started out with some historical background by saying that in the client/server world of the past, the people affected by server problems are company employees. While this causes lost time, money and productivity, the problem is internal. In the e-business world of the nineties, processing is now completely server centric and server problems not only impacts a company's employees, but also impacts their customers, suppliers and even their brand image. Some of her customers say that they are doing well when they are "not in the paper" (referencing the recent outages at several large Web sites, like Victoria's Secret and Macy's recently).
She also stated that the key strengths of the AS/400 product line were availability and scalability. Since servers are the "engines of e-business", availability is a critical issue since your business suffers if your supply chain can't access your server because it is down being repaired or rebooted. Scalability is also important since people can't accurately predict demand and nearly everyone underestimates. She said that some estimates were 10 to 100 times lower than what actually happened and that you need a product line that can easily scale to your needs.
The panel discussed many other topics. Here are some of the highlights:
Although the Gartner Group had not conducted a formal study, Tom subjectively felt that the TCO (Total Cost of Ownership) of the AS/400 versus NT systems were about the same. To me this was somewhat hard to believe, since PCs are so inexpensive to buy and deploy. After I thought about it for awhile, it made more sense and I rationalized it like this: Systems with comparable performance to AS/400s have 4 and 8 way processors and don't sell at the local retail store for $1,000. These SMP (symmetric multiprocessor) systems also cost more to maintain and when they break or go down in the middle of the night, someone has to fix them. This costs money, but the largest cost is the lost productivity and lost profits, during that downtime, which is extremely expensive. From what Susan says, the AS/400 keeps churning for very long periods, without intervention, so even if the AS/400s initially cost more, they have more availability and no productivity or profit losses -so the TCO balances out. Tom also noted that as Microsoft begins to change some of their licensing strategies for the enterprise versions of Windows 2000 software, costs will increase and the TCO of Wintel systems may be higher.
Many people felt that although AS/400 may be more "available" than NT servers, NT may be good enough for some users. Susan countered this by stating that the tolerance for outages are decreasing. She went on to say that UNIX and NT have expanded from their beginnings as departmental systems and are now expanding into Enterprise wide situations where it is a "new world" for them - they have not been tested for these stressful environments. She also felt that the "bar is being raised" for availability in her Fortune 1000 customers.
Susan also pointed out that IBM's Netfinity servers are an improvement over a standard NT server. IBM has added additional value by including "best of breed" features such as remote service path diagnostics. She also went on to say that it will be some two or three years down the line before UNIX is as reliable as the AS/400 operating system.
There was a general consensus from the attendees that the AS/400 is not considered a "sexy" piece of hardware, while NT servers are. Purchasing decisions are being made by functional executives, so cool image is important - and buyers want to use servers that are "with it". Susan responded that we need to watch IBM's marketing - "Watch this space". We shall see. Microsoft has marketed immature technology, told everyone that it is cool and millions followed (many knowing that there would be problems, in spite of better and more reliable technology being available - but the Microsoft products were "cool" to work with). This is a fight for "mindshare".
The audience wondered what IBM is doing to get young programmers interested in developing applications for the AS/400. To me it appeared that IBM had some plans to promote AS/400 in schools, but appeared somewhat limited. Wayne said hundreds of colleges and universities have AS/400 classes, but the audience wanted IBM to get the AS/400 into the hands of high school students. Personally, I can tell you what that does, since I was a student that HAD access to an IBM computer in high school. It wasn't anything close to an AS/400, it was an IBM 1620 with 16K of RAM and a 10 megabyte hard drive. Although considered trivial by today's or even the first AT standards, it ran the scheduling for the entire school and seemed (in my mind) to be quite powerful. That experience had a long term impact on my purchasing habits as I bought some of the first IBM PCs and even this article is being written on an IBM Thinkpad. First experiences are an unbelievable driving force over a lifetime of purchasing decisions.
Roger also felt that IBM needs to create some better development tools for the AS/400. He went so far to say that WebSphere is "an awful tool". From what everyone said, it seemed that IBM needs to come up with a development tool that has the pizzazz of Delphi when Philippe Kahn first showed it. (SeeInfostream: Software World March 1995 and read about it).
Susan's recommendation for people doing Data Warehousing was to perform this function on the same machine that currently holds the data. When the audience was surveyed (4 responded), 2 people kept the data on an AS/400 while the other 2 transferred it to an UNIX system.
Domino is a powerful application for the AS/400 and needs to be seen and ,as Roger puts it "touched", in order to fully understand what it can do, since it is difficult to describe.
Susan pointed out that from IBM's market research, they found out that many people consider IBM products to be "too expensive". She again alluded that IBM is going to be creating a dynamic marketing image.
This panel discussion was an engaging 80 minutes and was one of the liveliest I have been to in a long time. There was intense interaction and everyone (including the audience) was involved.
© 1999 Rick Smith All rights reserved.