Blade Insanity

Sometimes there just can’t be unity…

I spend most of my time looking at technology and how it converges for computing and telecommunications applications. One long standing trend that is now very much in the public eye is: How can one compute architecture possibly be used for both communications AND computing?

While the two share a common need for computing power - the degree of that power, what supports it, what powers it and what cools it differ considerably. In a nutshell:

Computing Needs
- Ever increasing levels of High performance CPU’s
- Lots of memory (make that truck loads)
- Large amounts of 110v power delivered to run the systems
- Very large disk farms to support
- Cooling on the order of the Arctic

Telecomm Needs
- CPU power of the right amount (not too much, not too little)
- Must meet cooling requirements of NEBS (Network Equipment Building standard)
- Lots of weird I/O so that when you pick up your cell phone or home phone - you get dial tone
- Must meet power requirements as set by NEBS
- Must work with legacy stuff way back around the time of Edison

Blades *may* allow us to have one architecture for both. But the big word is may. There have been a number of compute blade appproaches. A history on that would be fun to write and I may post one here one day for fun, but briefly they include: Multi-bus, VME Bus, cPCI, MOMA (my own modular architecture), and now AdvancedTCA. These are all standards meaning that a body or organisation has received a specification from members who’ve peer reviewed it then gone ahead and voted on it. Then, they’ve gone ahead and started developing useful products around the standard so that target industries can use them, enjoy the benefits of their being a competitive ecosystem, and then drive innovation faster in the areas that the target industries specialize in.

This is great, and the way things should be (in technology, and maybe even in life). When alternatives show up to to the standard, that’s okay, they usually address the outlying pieces that are really nicheoriented, and that is a great business for cutting edge technologists (of either gender!) to go after. But proliferation of pseudo-standards to the fracturing of bona-fide standards only fractures markets and makes them less viable - and removes the customer’s confidence in new technology causing them to keep with the old. That means industries don’t grow - and jobs go overseas.
Innovation is great! And frankly we need that in this country more and more every day (especially social innovation & entrepreneurship, see the excellent book by Peter Drucker, “Managing in the Next Society” for more) - but innovation that destroys new standards geared to building and expanding markets may be more about enhancing profits for one rather than raising the tide for all.