Archive for September, 2007

Vendor Roadmap Accuracy and Hardware Capability Lists (HCLs)

Know What Virtualization Is, But What Is Next? – Chapter 03

Vendor roadmap accuracy and hardware capability? Depending which side of experience curve on these issues you may have suffered with, you will have a debatable view on vendor roadmaps and hard capability lists (HCLs). This is a sticky topic. Some will say that negative comments are not appropriate, some will say more negative comments should be made to get vendors to see the light. Well, I choose to take a different route, I will point out a few issues with vendor roadmaps and HCLs, but I will also point out that the issues with these items we, the IT consumer have created, in fact, we are our own worst enemies.

Hardware vendors and software vendors are not effective at very high rates of change, a classic example of haste-makes-waste. Do hardware vendors test their designs? Yes. Do they attempt create quality products? Yes, at least most do. Do they rush solutions to consumers? Of course they do. Who suffers? We do. But we are the ones that set the expectations. But are we? That is the question. In fact, our expectations are based in part on vendor roadmaps. How many times have you been in a roadmap meeting and asked for details or clarification, only to have the vendor say… I don’t know, I need to get back to you on that… What the heck? The roadmap meeting is supposed to answer such questions!

Changing tracks from roadmaps, to HCLs, there are some key issues, the speed of computer technology refinement, yes, refinement, since computing technology has not see a fundamental leap in concept since the microprocessor. The industry is micro-scaling the same techniques, increasing speed, reducing power consumption, improving efficiency of the technology, but are still silicone based. Demand for capacity, 64bit on micro-architecture, was it really necessary? Yes and no. Yes, to offset costs, improve scale of virtualization, sure. But no, in that it did not gain us any real processing power, mainframes can still do more than any micro-architecture, and mainframes running 128bit will prove this point with emphasis.  So why have we not retired completely 32bit and simplified the HCLs again? Cost, 64bit is not cheap compared to 32bit. Thus HCLs are very complex, and that is a problem.

So, what do we have, roadmaps that are very loose, and HCLs that are complex. Combine these issues, with our unreasonable expectations, not a good situation. There is no doubt that things are rushed, yes, rushed to market. The pressure is to sell units no matter what anyone says.  Vendors are cutting back on the total research and development time, and vendors that use generic components often push issues back to OEMs, and OEMs push back to component providers, so what is the typical consumer to do? No one sees accountability and responsibility as important?  I can not tell you the number of times a given hardware vendor, and I mean all of them, have passed the buck, either implicitly or explicitly, for issues that should have been found in the vendor research and development or certification process.  But then again, we are setting the expectation?

As long as we believe we can save money we will always chase the latest and greatest.  So now we get to the real expectation, which is our fault and the vendors, we are chasing cost savings, and hardware vendors are slicing time of their design, development, testing and certification timeframes, so less stable platforms result, this true both for hardware and software design, but in hardware design this is quite ugly. This is one of the true negative sides of just-in-time computing models.

I once worked for a company that said, “We will not implement or release a solution before is stable, solid, and robust enough, so that our consumers have a consistent positive experience.” Not a mission statement, but burned into us routinely reinforced by quality assurance policy and procedure. When this statement was popular, it was representative of the entire computing industry, so I ask you, where has that industry mindset gone? We killed it, because it held us back, made us do good work, and made us accountable for real results, and we believed that was costing us money. But was it really?

2 comments September 24th, 2007


September 2007
« Aug   Oct »

Posts by Month

Posts by Category