Wednesday, December 08, 2004

ETS: Wednesday keynote

Mike Gilpin of Forrester talked about SOA as a "platform for organic business." Everyone's desired buy online, pick up off line is supposedly hard to do. He also talked about interaction platforms, which is front-end integration (instead of back-office) and software innovation networks. In short, most successful new companies will be part of the ecosystem of a major player.

The keynote speaker was Andy Mulholland of Capgemini. There is no killer app, but lots of near lethal new technologies. People now assume that everything is connected. People have discovered that standardization is more important than customization. This leads to SOA. We now have vertical industry standards, not just IT standards.

Compliance is an issue. Auditors care about process, not function. He then says the killer app is process, after starting off saying there was no killer app. The bust was about content. Companies that were about changing the process are thriving today.

Another interesting point he had was that the browser is a mobile client. This was part of great slide that is way too complicated to duplicate here, so you'll have to visualize. Concentric squares containing enabling standards: semantic stds, URI, WS-RF, and WS-Eventing, infrastructure: mobility, web services, grid computing, and the web itself, networks, and finally agents.

An IT project needs to either have a really big business impact, lower costs while providing a business impact, or lower costs with the same impact. Obviously, the second choice is the best. He also showed a progression from license cost savings to internal cost reductions by shared resources to internal value created through process redesign to, finally, externally created value through more efficient market interaction.

Miscellaneous points:
  • A goal should be to turn application data into content.
  • Data mining is not BI. BI is a look into what is happening now.
  • Most of the grid computing payoff he's seen has been is BI. Capgemini harvests the MIPS of 6500 PC's in their grid. It's difficult to get people to accept that they don't need a server; they simply need MIPS.
  • Most of the opportunities for startups are in verticals. There is a trend to simplicity, so take something that is big, complex, and hard and make it simple. My colleague Olin, who's attending with me, immediately thought of Asterisk, the open source PBX.


3 Comments:

Blogger Guy said...

See, I have a beef with many of Forrester's messages lately. For example, I believe the trends indicate:

most successful startups will come out of nowhere, from nothing

their standardization/customization is a flip-flop reminiscent of a politician. clearly, the people (in control, right?) want mass customization

press them on that "process is all-important" position, and i'll bet you either get a very wooden response, or a series of contradictory statements. the pendulum may already be in full back swing, right?

7:19 AM  
Blogger Eric said...

Remember that the standardization/customization remark was by Andy Mulholland of Capgemini, not by a Forrester person. The Forrester view may be different.

1:42 PM  
Blogger Guy said...

fair enough

yet, remembering this piece still gives me fits:

http://www.forrester.com/Research/Document/0,7211,35646,00.html

as if "bad data" carried no information, had no value

not to mention Heisenberg's Uncertainty Principle

3:13 PM  

Post a Comment

<< Home