The IBM FUD thrower
I just found this piece of FUD in a tweet. IBM tries to move Sun customers to their pSeries. I don’t want to talk about the point, that they have to give massive discounts to do so, so the story of “an uncertain future” for Sun in the light of the proposed Oracle merger doesn’t seem to generate the intended attraction. But this is a different story. At the end it’s just the current fashion in FUD. But it was pretty obvious that they would roll out their FUD thrower out the the garage.
As usual the most interesting point is the “Legal Footnote” section you can reach via a link at the bottom of the page:
- There is an error, the first claim doesn't fit to the legal substantiation, but well … I'm sure this "50 V890 to one rack of p550" is based on their standard tactics, and when I look into the legal footnote, i'm correct. At first all this claims about performance and savings at all are based with the IBM patented marketing factor 3 for virtualization. So IBM gives itself a nice, but unjustified performance advantage. Especially when they take SPECjbb as an example, as you could do virtualization for Java processes just by running them side by side on the same instance of your operating system. When you want some administrative domaining, just use containers (Virtualization isn't a consolidation feature, a Unix can do this on its own, it's a security feature, but that's a different story).
- In the footnote not matching to the example there are other strange things, like assume 99,9% for IBM and 99,8% availability for Sun without substantiating such claims. This is a little bit strange given the fact that there are talking about the E6900 which was one of the machines where you could change everything while running (okay almost everything, removing the rack in operation is hard, as this system doesn't have auto levitation device). As far as I know when you want to extent a p570 from 2 to 3 or from 3 to 4 drawers you need new cabling, thus you need a downtime maintainance. You have to factor in planned downtimes as well as unplanned downtimes. But this first legal foot note is just strange thus I wont dig deeper until into this claim, until IBM fixed the error of a footnote not matching to the number.
- The second claim “20 Suns on 2 IBM server” can't be answered, as they didn't provide the information, what servers were used before. But I'm sure it's old servers vs. new servers again and 3times better utilization trick again. No need to write more about that.
- The third claim “180 V490 subsituted by 56 Blade Servers” collapes in the light of this utilization claim. Just by converting thus V490 into containers and using the same ratio I would get to 60 V490 and i'm sure we could do better than 60% as containers doesn't know the overhead of LPARs for example. And it's old server vs. new server again. Let's calculate it with new Sun servers, so you don't have to migrate to new operating system. You have two ways to do this: You could use Sun Fire T5120, it has 97 SPECint2006 rate (result). So when we assume the same utilization with those server as on the IBM server we would end with (58.2 SPECint2006rate per Server (60% of 97 SPECint2006rate) then I need 48 of this servers. Or one 6048 chassis, a single rack. That's even less than the 56 IBM need in their example . Or 12 M5000 (3-4 racks) giving you the expansion capabilities a blade can't offer ,but a V490 have. The example of IBM completely omit this point.
Okay, when you want to read this marketing bullshit on yourself, you can look at it on IBMs webpage. And don’t forget: After such a migration you have IBM Global Services in your datacenter. Do you really want this? ;)