Monday 31 December 2007

IBM – the future

Directors and shareholders of IBM must have tucked in to their Christmas dinners with a certain amount of satisfaction that the company was still very successful and that there was plenty of revenue coming in from the various parts of the company.

I don’t want to start the New Year being thought of as the blogging equivalent of Cassandra, but things may not be all good at Chateau IBM.
Cassandra, you’ll remember from Greek mythology, was the daughter of Priam, king of Troy. She was given the gift of prescience by Apollo, who later, because she didn’t return his love, cursed Cassandra so no one would ever believe her predictions.

There’s really three areas I want to mention in this blog: hardware, IMS, and little systems.

In terms of hardware, I’m really thinking about selling new computers. Like new cars, every time IBM brings out a new mainframe, someone is going to want it! It will seem like an appropriate time to upgrade – and IBM’s sales figures will look good. But, in fact, it seems that IBM is only selling to the faithful. How often do you hear of a company getting rid of a room full of servers and installing a mainframe? Probably not that often. Exploiting zLinux on a mainframe may make it look like the mainframe is an exciting place to be, but revenues must be small compared to that achieved from z/OS. What I’m really saying is that IBM needs to find a way to reduce drop-out, ie small VM/VSE sites getting rid of their mainframes, and, more importantly, start selling to new customers.

Secondly, I have been doing a lot of work recently with IMS. That’s Information Management System, not IP Multimedia Subsystem. IMS, as you probably know is a database/transaction management system. It’s like DB2 and CICS combined (sort of). It’s used at lots of large mainframe sites and is the core of those companies’ business.es Version 10 was announced a little while ago by IBM. So, IBM has this brilliant piece of software that so many major companies rely on, yet, when was the last time it sold a version of IMS to a new customer? I don’t actually know the answer to that question, but I’m let to believe that almost all IMS users have been users for a long time – ie there are no new customers. Come on IBM! If you have such a brilliant product, why aren’t you selling it?

How do you get people interested in mainframes? The answer is to have lots of them around and let people play with them. Now before you start sneering and saying that will never be possible, let me suggest a way. How about FLEX-ES from Fundamental Software? It provides a way for developers to test mainframe software on a laptop. Or there was UMX technologies. And, of course, Hercules – the Open Source mainframe emulator. Platform Solutions has a product called the Open Mainframe. There’s even Sim390. If IBM was to embrace these technologies and not try to smash them like the Hulk, people would be more familiar with mainframe systems because they would be more commonplace.

I have a final revolutionary thought. How about VM on Intel chips? Everyone says that Microsoft’s attempts at virtualization are a bit weak at the moment. VMware seems to have the lead in the market place. Why not adapt VM itself to run on Intel servers. This would be a good way of training new people in mainframe concepts and it becomes only a short step for them to become mainframers. Why not allow Hercules or FLEX-ES technology to plug in to a server running VM. That way z/OS would become a mid-range system and sales would grow hugely. Eventually those mid-range people would definitely consider buying a shiny new mainframe because it would not be a risky step for them to take.

This is the way to a successful future for IBM, and they might even sell IMS to a new site – but will they believe me?

Happy New Year to everyone.

3 comments:

Anonymous said...
This comment has been removed by a blog administrator.
Anonymous said...

Thanks for the nice blog
Regards,
Jack
http://db2examples.googlepages.com/

Anonymous said...

Amen. IBM, you make money off of software, and that was proven when Amdahl was a viable competitor.

What if Fujitsu and Hitachi allowed their MSP/VOS operating systems (and the VSE/VM equivalents) to run on Hercules? Or PSI? (Consider this an open call to whomever reads this...)

I'd also like to see if BS2000/OSD (from Fujitsu-Siemens) would run on Hercules.