HP HAS POISONED THE WELL for its non-X86 high end machines. Anything that does not have direct roots in a desktop x86 PC is now dead, and the money it would bring in is lost. As is the case with so much of HP recently, it is due to the same thing myopic mismanagement. By the time it becomes apparent to onlookers, HP will already have declared 'victory', fired most of the remaining people, and given management huge bonuses. Tough calls deserve many zeros on a check.
Now, let's begin by looking at the numbers. The only one I can easily find for the group are
here, but due to the re-org of the week, I am not sure how much of this is TSG and how much is ESS. Either way, the number adds up to billions of dollars, and it is going away. More bonuses all around, unless you are an employee, then you get
.6%.
So, why is this going away, and what do I mean by mismanagement, or mrs-management? If you look at the markets the beasts like Superdomes and, heh heh, Integrity servers sell to, they are not actually as concerned with things you would expect them to care about like speed and price, but more with reliability and long term strategies.
These computers are frighteningly expensive, very powerful, and usually come with a fleet of engineers to install, service and dust them. Service contracts are almost always worth several times the cost of the hardware, and uptimes are measured in the number of nines after the decimal point. They are expected not to break, not to crash, and to be fixed
now if there is a problem. In practice, they don't break, don't crash, and are fixed now or sooner. If you are running a stock exchange on this kind of system, the fact that they cost the GDP of several small nations is just fine.
HP had some of the, if not the best people, technology and machines in the industry. They had servers, OSes and technology that were the class acts of the field. Some were internal, some traced their lineage to Compaq, and more still was from DEC. This was a frighteningly lucrative mish-mash of parts, but it got the job done, and done well. Much of this was the reason that HP bought Compaq. It is the stated way of the future for HP, just look at IBM for your roadmap.
So, where does the mismanagement come in? When you buy a Superdome, you don't tend to buy a box, you buy a solution. This type of project usually covers the next three to four generations, and a consistent upgrade path. You don't want your 24 hours by seven days a week operation to go down for the next 10 years, and you simply expect there to be something from a vendor to make sure that's the case.
HP and others in the same field will give you a roadmap if you desire, and tell you their plans, because your plans depend on their plans. They may not follow it exactly. If they tell you that there will be a new generation of chips in 2009, and they will be 2.3x faster than their predecessors from 2007, you are expected to take it with a grain of salt. It may come in 2010 and be 1.9x as fast, or maybe in 2007 after all and be 3x as fast. Numbers don't really matter as much as the fact that they will be there, and you won't have to rip up and replace your infrastructure. You have an upgrade path and smooth transitions. Stockholders like smooth transitions, as do customers and employees.
So it is less about what you actually do, and more about you trying hard and making life as easy as you can for your customers. Smooth, good. Pain, bad. An updated roadmap with your customers in the loop wins repeat business. Deviations are occasionally permitted with good reason, coupled with a lot of ass kissing, and most critically, a backup plan.
This is where HP has failed its customers and it's not just once or twice, but three times. To make matters worse it made a fatal misstep in not admitting its mistakes, and pretending there was nothing going on. The problem is that anyone working with the affected technologies is more than smart enough to see through this. This is classic mismanagement, with incompetent executives inflicting a penny-pinching agenda on a technical area that is not price sensitive.
If you are selling widgets worth $1.5 million an hour on your web site, and your back end databases are reaching capacity, do you care if the next model up costs $7 million, or $7.2 million? Nope, not in the least. Do you care that a promised upgrade is not going to happen, but it will have a replacement for you in a year? Yup, you care, and care a lot. It may take years to recode, test, and qualify your new software and infrastructure, but it is do-able if you have enough warning and plan accordingly. HP actively screwed its customers three times. It started with the Alpha,, then went to storage, and now it's started hacking about with the OSes. The only saving grace for HP is that it does take years to move to bigger and better systems, but several customers we have talked to have already started the move, with many going to IBM, and according to Tru64.org polls, Sun is also slated to be a big beneficiary.
Source:
http://www.theinquirer.net/?article=20208
Linear Mode
