IBM has been pushing hard to promote the AS/400 as a Web server and the release of its WebSphere 3.0 has positioned the AS/400 as a serious contender in this market. Search400.com talked to Jim Mason, president of Cape Code Bay Systems about the impact this new version of WebSphere will have on the AS/400 and whether or not IBM can garner enough market share to compete in the e-business arena. Mason, who consults and teaches on AS/400 applications, client/server and Web technologies, says WebSphere is a good start, but wonders whether IBM can be aggressive enough to pull it off.
SEARCH400.COM: WebSphere 3.0 has received high marks from a number of independent reviewers. How significant are these upgrades to AS/400 users?
MASON: Very! The AS/400 now can be a serious contender as a Web application server for the first time. Java server applications based on servlets, JSP, XML and EJB are rapidly becoming the preferred method for delivering critical e-business applications over all the other technologies out there. With the features and performance of WebSphere Advanced Edition 3.0, the AS/400 has an opportunity to gain share in the most significant e-business projects that companies are looking to deliver.
SEARCH400.COM: Will WebSphere 3.0 encourage IT to use the AS/400 as a Web server? Does it make any difference?
MASON: Yes. It will support servlet api 2.2, EJB, JSP 1.0 and XML 1.0. All are key levels of function needed for serious server e-business applications.
SEARCH400.COM: The AS/400 always seems to be the last in line for new software leaving a gap between it and other platforms. Is IBM's e-business push going to change that?
MASON: Only on Web-specific software. Related issues like DB2 probably won't get the budget needed to match other IBM platforms in function. Compare DB2/400 to IBM DB2 UDB 6.1 that runs on most other platforms (Linux, Windows, AIX, OS/2 etc). Also, look at the Sun JDBC 2.0 specification. It's clear there is a lot of missing function on the AS/400 side now. This doesn't mean anything to RPG developers normally, but it's a show-stopper on the e-business side when people look at data warehousing, business intelligence, etc.
IBM has to make the AS/400 current in all areas related to e-business if AS/400 has a chance to increase share in that market space. Right now, they've done a great job with Java. DB2 isn't current. They are just catching up with WebSphere 3.0 on the Web server side.
IBM Rochester might do well to rethink their approach to AS/400 middleware. Rather than develop it native for the AS/400 they should consider porting more of it, similar to the way they've done with Lotus Domino. There would be performance penalties for sure, but the benefits would be lower costs and much faster time to market. This would eliminate the gap in Rochester's ability to deliver new middleware function on a timely basis compared to its efforts on other platforms. Overall, the performance penalties are less important to the success of the AS/400 than the lack of current versions of critical e-business middleware.
SEARCH400.COM: Is it vital to the success of the AS/400 that it does?
SEARCH400.COM: Do you expect the installed base of AS/400 to increase, decrease or stay about the same in the next few years?
MASON: Increase. Whether it's a small or large increase depends primarily on IBM rethinking its current software strategy. They need to focus on open systems frameworks and Java for e-business. They need to market, well ... I have some creative ideas that would make an impact, but IBM wouldn't want to be that creative.
Generally, using a common code base across IBM where possible would benefit the AS/400 division with less custom development and corresponding delays in getting frameworks current.
SEARCH400.COM: What will be the one thing that will make it or break the AS/400 as a server and/or Web server?
MASON: E-business support including Java, Java frameworks and Database. That includes both runtime, administration and development tools. They should use IBM software where possible and write as little as possible. AS/400 application vendors aren't delivering very technically advanced applications today in most cases. This is perhaps the largest hurdle to overcome. A strategy of providing better tools to customers to build their own component based solutions faster would reduce the heavy dependence on application packages that are older technology: expensive to build, implement and maintain.
IBM has a lot of room for improvement here. AS/400 division should attempt to break out of the pack in application productivity by going with more advanced object component support for applications. Their money is fragmented now in supporting Windows specific software, e-business and older AS/400 proprietary frameworks. The AS/400 division should attempt to run all IBM software and add a few key development technologies and frameworks on top of the IBM open system infrastructure.