IBM tests 4-terabyte solid-state drive tech
First it was Intel. Now, Big Blue is keen on solid-state drives.
IBM said Thursday it is testing a 4-terabyte, high-speed solid-state drive array targeted at the enterprise, as the technology giant gives its imprimatur to flash-memory-based storage.
For years, flash memory cards--the first mass-market SSDs--have been limited to digital cameras and music players like the iPod. But SSDs are now poised to hit technological critical mass in terms of storage capacity, speed, and availability as they find their way into everything ranging from tiny netbooks to massive enterprise storage arrays.
High-performance enterprise storage is where IBM comes in. Engineers and researchers at the IBM Hursley development lab in England and the Almaden Research Center in California have demonstrated performance results that outperform the world's fastest disk storage solution by more than 250 percent, according to IBM.
Under the rubric Project Quicksilver, IBM coupled solid-state drives with its storage virtualization technology to achieve a sustained data transfer rate of more than 1 million input/output per second (IOPS), with a response time of less than one millisecond in a 4.1-terabyte rack of SSD storage. SSDs are being supplied by Fusion-io.
By comparison, Intel is commercially shipping SSDs (X25-E Extreme) that individually achieve random data reads of 35,000 IOPS and random writes of 3,300 IOPS. In a 3.8-terabyte storage array using 120 SSDs, Intel claims 4.2 million IOPS.
IOPS is a crucial benchmark for large customers that process credit card information or run reservation systems, for example.
"It's feasible that we could get it commercialized within 12 months," said Charlie Andrews, director of product marketing for IBM systems storage. "Right now we have a screaming (fast) system, but there's more work to be done in terms of long-term reliability and integration with systems applications. We don't want to get distracted with 'push the hardware.' We want to focus on the solution piece first," he said.
Compared with the fastest industry benchmarked hard disk drive system, Quicksilver not only improved performance by 250 percent but did this in less than one-twentieth of the response time, one-fifth of the floor space, and with 55 percent of the power and cooling requirements, IBM said.
"Performance improvements of this magnitude can have profound implications for business, allowing two to three times the work to complete in a given time frame for classic workloads," the company said in a statement.
IBM's said its first implementation of solid-state drives was for select IBM BladeCenter servers in June of last year.
Brooke Crothers has been an editor at large at CNET News, an analyst at IDC Japan, and an editor at The Asian Wall Street Journal Weekly, among other endeavors, including co-manager of an after-school math-and-reading center. He writes for the CNET Blog Network and is not a current employee of CNET. Disclosure. 
Either way, I'd say somewhere in the upper 5th if not 6th figure area.
How would Fusion's ioDrives fit in IBM's storage? I thought they were SSDs made for servers?
The workload benchmarked was 70/30 4K all miss which is somewhat representative of a typical database workload.
If people have actually played with SSD devices, they will know that the claimed X0,000 reads and Y,000 writes are all very well when doing 100% read or write at say 512 byte transfers, but bump up the transfer size, and throw in a mixed workload and you will prune the men from the boys... i.e. IOPs will be no where near those claimed by manufactures...
While EMC has been shipping SSDs since the 2Q, the question I'd ask is how many DMX quadrants would be needed to provide this same level of IOPs... many many... which defeats the power and spaces savings that SSDs can provide.
Anyone interested in more details of what we have been doing see my recent blog post :
http://www.ibm.com/developerworks/blogs/page/storagevirtualization?entry=1m_iops_from_flash_actions
For write workloads, storage controllers have added cache and the like to hide such latency, but that becomes less improtant with flash. Also consider a (large) FC-AL loop of HDD, say 128 HDD, then the dual loop config needs to only support say 128x350 IOPs ~=45K iops (assuming all HDD are flat out - which is unlikely to be possible) so thats one flash device... There are likely to be bottlenecks further up the system too... So while it does help improve latency by using existing controller devices, it cannot max out the devices, not can it provide the best latency.
With a modular system like IBM SVC, then you have 50us of latency top to bottom of the stack, and with the flash devices typically giving 400us under heavy load, you can keep to around 700us for the entire system, host -> SAN -> SVC -> flash.
Monolithic controllers are fine until they reach their limit, then you need to bring in another monolith and scale out in islands, with a modular scale out box, you simply add more processing nodes (SVC) to the cluster, and more storage behind the cluster and still have a single image and control point.
For this config we used around 40 FusionIO cards.
Was this testing also done with some form of data protection?
- by bitbucketio September 10, 2008 10:16 AM PDT
- orbist : I assume you're using the IOdrive without the IOsan port. How many IOdrives were in your test?
- Reply to this comment
-
(14 Comments)Are there any plans to use the IOsan in your storage solutions independent of the SVC?
http://www.fusionio.com/PressDetails.aspx?id=44