Yesterday we had a meeting to assess “where we are” with some of the cutting edge “technology” our customer has heard of and now demands.

What I could gather:

x`

1. We can install Linux on Z-series mainframes. I guess this is cool because we can install thousands of copies of Linux on these systems, and all of the networking is done at “memory speed.” Exactly how does this benefit our client, who doesn’t run Linux at all in the enterprise and is trying to REDUCE complexity? If you have thousands of Linux systems, you still have thousands of systems to manage—even if they are physically “one” box.
2. In our efforts to gain utility computing (delivering computing power on demand like electricity or water), we have figured out a way to bill somebody for 10% of a system, when we are, in fact, paying for 100 % of this system. This means we’ll have to mark up service so much that it would probably be cheaper for the customer to build a data center and buy the servers and applications themselves.

I guess we’re looking good. Once we can figure out how to get our customers to pay us “by the minute” for using their own servers, we’ll be golden.

In short, I don’t see anything on the market that looks good right now in terms of solving the current problem facing the vast majority of IT shops whether self-managed or managed by a sourcing vendor. The fact that most companies only utilize less than 20% of their computing resources cannot really be solved with anything currently on the market, but only via careful management and lots of hard work.

IBM, Sun and HP are all pushing their solutions to this problem, but they require basically trashing the existing infrastructure and purchasing new hardware and software. How one can save money doing this is beyond me, regardless of the reduction in “complexity” or the increase in efficiency.

The golden egg is to be awarded to the company that comes out with a product and service combination that allows for the repurposing, on the fly, of existing infrastrucure and software resources—much like an after-the-fact LoudCloud. It’s probably still two or more years out.