Where is the Sun Blade Server comparison?

Mentioning Sun squeezing in 4 Intel hex-core or SPARC hex-core processors into a single blade, squeezing in 12 blades into a shelf, and squeezing in 4 shelfs into a rack... to produce 1,152 cores per rack... or 27 cores/U would have been nice.

Mentioning Sun squeezing in 4 AMD quad-core or Intel quad-core processors into a single blade, squeezing in 12 blades into a shelf, and squeezing in 4 shelfs into a rack... to produce 768 cores per rack... or 18 cores/U would have been nice.

http://www.sun.com/servers/blades/x6275/gallery/index.xml?p=1&s=1

I guess mentioning Sun's product would have placed Sun and Fujitsu on the top of the blade food chain as far a density and computational power, though.

Sun Who???

Well done David, Did you mean Oracle?

Noted the icon, probably a good move as you seem to be so on the pulse............

re: Where is the Sun Blade Server comparison?

Yawn! Give it up Sun are dead as everyone who lives in reality has realised/been saying for the past 6 months!

Oh and if Sun's products were really that good overall (instead of having some nice technical one offs (mostly because other manufacturers concentrate on what actually matters!)) why have they conststenly underperformed compared to their R&D and engineering talent?

Oh and your Sun blades arent as revolutionary as you thought. 4 hex cores in a blade, yep HP can do that, I fail to see your point?

Reg wheres the sunshiner still living in the clouds icon?

As for the rest of the article I find it amusing that Fijitsu claim that their blades chassis consumes less power yet admit that each blade can't hold as much RAM, that should be a nice chunk of the "saved" energy consumption then! I would be interested to see if their fans are more efficient seeing as HP have some pretty clever Thermalogic tech that means they spool down fans to their minimum required level (whilst still keeping reasonable internal temps, unlike the IBM stuff) coupled with the pretty efficient fan design.

Adam

Sun Blade Comparison...

Chris Bone posts, "Did you mean Oracle?"

Oracle does not own Sun yet. Perhaps the Sun trademark may remain, if Oracle can purchase Sun, as Catalyst remained when Cisco purchased them.

Adam posts, "why have they conststenly underperformed compared to their R&D and engineering talent?"

Perhaps, because the people who cover the trade consistently ignore Sun's products in the comparisons, even when they compare very favorably, as in this case.

Adam posts, "HP can do that, I fail to see your point?"

You sound like Matt Bryant! LOL!

My point is that Fujitsu was compared against other vendors (including HP) in this market, but not Sun.

Fujitsu was cited with the highest number of cores per 1U rack. Sun offers the highest number of fast Intel 5500 cores per 1U: Sun - 18; Fujitsu - 14.4; Dell - 12.98; HP - 12.98; IBM - 11.2

I do not have a specific "beef" with HP - they are a competitive company with competitive products. Just observing, citing, and correcting a little unfair coverage in regard to Sun.

@ David

David, take your point, however thats not really the point!

Fujitsu, HP, and IBM for some time have realised that cores/performance is only part of the answer, the real question is how do you harness it, manage it effectively, keep it secure, power it and finally keep it cool.

It was not so long ago that Sun claimed the biggest/highest performance Unix box on the market which began to fail all over the place (www.forbes.com, search for sun screen, I believe the article is still there). Ok so they they bought that box from Cray but did they design the x86 architecture from the ground up or did they do this through aquisition?

I do have to agree with your last comment however, its about time the world woke up to the fact that there is more than one capable x86 vendor out there!

Interesting.

I always though the old FSC blades designs were often overlooked and far too often described as a clone of the IBM chassis. You can see now why Fujitsu didn't see the need to buy Sun, their own designs are better than the Sun ones. The power comparison with the hp blades is amusing - I'm guessing they left disks out of the Fujitsu kit during the comparison as the only picture I've seen of the new chassis had blades without hotswap disk slots, unlike the hp blades which do have hotswap disks. Unfortunately, the link to the info on the new blades on the Fujitsu.com website is broken so I can't confirm.

If their SAS or SATA disks are not hotswappable, where you have to shutdown and pull the whole blade out to replace a disk, then we wouldn't want them for a lot of our implementations. Despite using a lot of SAN attach, we still like having hotswappable local disks and have some blades without SAN attach which just use local disks and local storage blades.