<   Previous Post  SSD's are becoming a...
SVC certified with...  Next Post:   >

Comments (4)
  • Add a Comment
  • Edit
  • More Actions v
  • Quarantine this Entry

1 localhost commented Trackback

I saw that Snickers ad too. Amazingly it had significantly higher production values than the Hitachi ads all of which looked like they were filmed in Claus's back garden. <div>&nbsp;</div> I always preferred Murdock myself. Even as a kid I never got how BA kept on drinking that sedative laced milk just before the A-Team had to get on a plane.

2 localhost commented Trackback

Well, my own _personal_ perspective is here (along with many, many insightful comments): http://thestorageanarchist.typepad.com/weblog/2007/10/0039-ibm-and-sp.html<div>&nbsp;</div> But it really seems odd that the SPC has folks duking it out over who can build the fastest benchmark system without regard to price. The "winners" all clock in at more than $100/GB, while customers are asking for solutions that approach more like maybe $2-5/GB.<div>&nbsp;</div> Maybe we need a benchmark that measures how fast you can go with only $100,000...

3 localhost commented Permalink

True, that pure disk costs of $2-5/GB, maybe for SATA or big FC disks, but the SPC cost per GB includes the entire tested configuration, not just the disks.<div>&nbsp;</div>

4 localhost commented Permalink

Yes as per comments on you blog, I stand corrected about the host price being included. <div>&nbsp;</div> One thing though is your companies statement about the negation of cache. Yes SPC-1 does move the re-addressed area around during the test, however this is an attempt to prove the effectiveness of the cache algorithms, not an attempt to negate caching altogether. Infact, this should be to the benefit of devices like USP-V, DMX and DS8000 more than SVC with their much greater cache capacities.<div>&nbsp;</div> Anyway, like back in July we are in a cyclic discussion, the overriding feeling I get from the comments out there is that while some feel SPC is not relevant, and EMC refuse to contribute, most big vendors now do, and again while it is not representative of every customers workload IT IS A CONSISTENT YARD-STICK by which all products can be measured. The test being the same no matter what product is used.