JDN, >That really makes me bite my nails over Itanium. If the press is willing to blow Sun's mundane problem out of proportion, think how nasty those hounds will be if deployed Itanium servers run into a problem of the same magnitude.
That's true, they'd be all over it like white on rice. But, look at it this way, current Intel based (PIII or Xeon) servers have a great track record. I don't remember a single reliability problem with these servers. Anyone else? That results from a good overall system design involving the OEM, like a Compaq, working with Intel, plus board, DRAM, cards like NICs, SCSI, fibre channel control, etc. makers. It's the same process for Itanium based, just a few changes, like the CPU chip!!!
You know, I don't think people around here, in general, appreciate the groundwork, working relationships and precedents that have been laid with the Xeon and the PPro before that. That is, with the OEMs. You don't just willy nilly switch to a whole new CPU vendor. It would be like Sun all of a sudden canning TI as CPU chip source. Actually, it would be worse, because, whichever OEM decided to do that would be all alone vs. the other three major OEMs. AMD, as a company with an alternative CPU chip, has a huge row to hoe, IMO.
By the way, having to reboot an enterprise server like a Sun is a very big deal. Ask Ebay. And, it's not as though a customer has many servers doing the same thing. In the case of Ebay, they had no backup to their main auction servers. When that one key server went down, their whole business was dead in the water. SUNW has amazing "ridethrough", that's all I can say (not the servers, but the stock).
Tony |
