John Brady's Weblog

Well, as I said in my previous post, Jonathan Schwartz would have to layoff a load of people at Sun to make it profitable again. The only question was would he strike quickly and take the penalty in Q4 of Sun's financial year, or wait it out a bit longer until the next financial year and blame things on the situation he inherited?

Well, he's decided to get on with it, and has
announced that between 4,000 and 5,000 people will be laid off.
This makes another layoff of over 10% of their employees, which I make to be at least the fourth such layoff of this size. There have of course been other smaller layoffs as well, such as the recent layoffs of 200 SPARC engineers.
So now Sun will have laid off at least 40% of its total workforce over the past 5 years. Way to go guys! I think I already blogged about the 'death by a thousand cuts' scenario playing out at Sun. And here it is again.

Bizarrely, Sun have announced this massive layoff as a growth plan. Another example of Sun's brilliant senior management reasoning abilities. Who else would try and position laying off 5,000 people as a growth in the company? Wierd!

In summary, I am not surprised at any of this, as I have seen it coming for some time now, and is exactly what I said would happen. I was just unsure of the precise timing.

And don't forget, this is not just 5,000 people from what was Sun in 2001 of 40,000. This is Sun which has since acquired a long list of other companies, including the likes of Cobalt, and more recently StorageTek, Seebeyond and Tarantella. The list is really quite long when you examine it. Sun may actually be laying off people from the very companies that it spent a lot of money to acquire in the first place. Talk about bad strategy and management!

PS. I've also had a chance to recently benchmark a T2000 system with an UltraSPARC T1 processor in it, and it is not fast! In reality a single 1.2 GHz T1 CPU is comparable to 32 * 250 MHz UltraSPARC-II CPUs. So, yes, a lot of total processing capacity, comparable to an E10000 system from 1997 or so. But in today's world when all other CPU cores work at 1 GHz or above, squeezing 4 threads onto one core just ends up with all 4 running quite slowly. An impressive engineering effort, but only really applicable to highly parallelised, scalable, multi-threaded applications, like web sites. But not really suited to back end database systems. The T2000 was out performed in all of my tests by a 2 * UltraSPARC-IV+ V490, which can be doubled in capacity to 4 CPUs. Something the T2000 cannot do. Don't believe everything Sun tell you!