<   Previous Post  More Thoughtput
So I guess that only...  Next Post:   >

Comments (6)
  • Add a Comment
  • Edit
  • More Actions v
  • Quarantine this Entry

1 localhost commented Trackback

Hi Barry,<div>&nbsp;</div> I too have been excited about this kind of shift in technology for quite a while. I always wanted to get my hands on a Texas Memroy Systems RAM-SAN but never got the chance :-( Guess there arent that many of them around, may be they are a little ahead of their time?<div>&nbsp;</div> I see from my day to day exposure that many high end storage arrays have a mixture of hosts/applications that need little if any performance and some that need exceptional performance. So I agree that most apps wont need this kind of performance but that some would benefit greatly from it.<div>&nbsp;</div> Almost inevitably the apps that demand the performance push the write pending rate of existing arrays so high that cache is flooded, while a seemingly increasing number the other apps could quite happily get by with fat slow drives.<div>&nbsp;</div> The way things are going at the moment I see all high end storage arrays supporting and shipping with a lot of SATA disks - bit like the DMX-4. And who knows, may be a bunch of SSD as well some time in the hopefully not too far distant future?<div>&nbsp;</div> Nigel

2 localhost commented Permalink

Chris, Nigel,<div>&nbsp;</div> Yeah, I'd thought about the TMS products to solve some of my performance needs, but they to aint cheap either! <div>&nbsp;</div> I do hope and think that there will be some presence in the high end before too long, but when you think even with today's HDS announcement of 200K IOPs SPC-1 for USP-V, that would only need about 8 of these SDD's maybe 12 tops to flatten the box. Thats a lot of heavy metal for just 12 drives. The same goes of a DS8000, and for that matter SVC. Thats why I see it as a potentially disruptive technology when it comes to today's model of high end and modular storage controllers.<div>&nbsp;</div> PS. I had to send my evaluation SSD back last week :(

3 localhost commented Trackback

Barry,<div>&nbsp;</div> I am happy to loan you a Texas Memory Systems RamSan DDR system for testing. If you would like to take a look, just let me know. Prices have dropped substantially in the last year. We would especially like to publish an SPC result with your IBM SVC. Together, I think we could max out the test!<div>&nbsp;</div> Our flash based system should start shipping in late Q4.<div>&nbsp;</div> Woody Hutsell713-266-3200

4 localhost commented Trackback

Barry,<div>&nbsp;</div> That sounds like a score with Woody's offer!<div>&nbsp;</div> If you take him up on it please be sure to give your thoughts and the results of any tests you do. I've blooged a little about the RamSan in the past and love the idea of it, but never had the chance to use one or even speak with someone who has. So, your thoughts and insight would be very much appreciated.<div>&nbsp;</div> Nigel

5 localhost commented Permalink

Hi Barry,Have you looked at the power consumption of the zeus iops though? I was wondering how energy efficient they are since SSD proponents boasts of low power consumption of flash drives. Would really appreciate it if you provide data.Thanks!Zet

6 localhost commented Permalink

I'm curious if you've tested similar SSD devices previously. My company recently attempted to roll out a gang of 2.5" 32gb SSD drives in Sun Coolthreads servers and we achieved a 30% failure rate within the first 7 weeks of deployment. I'm wondering if the big numbers quoted by manufacturers in their white papers is in any way related to real world performance. <div>&nbsp;</div> Just for the sake of completeness the devices we had problems with were supposed to have 1,000,000 hours MTBF and 140 years write life at 50GB daily write.