New SSDs: Smaller, and Yet Larger

Arthur Cole

The prevailing theory is that IT spending will remain healthy as the economy slogs through the recession, particularly if that spending is aimed at streamlining data center infrastructure and lowering future operating costs.

In that light, it's no surprise that storage vendors are flooding the market with ramped-up solid state disk (SSD) technology, boasting both increased capacity and faster throughput over legacy network architectures.

Clearly, there is no comparing today's SSDs with traditional hard-disk drives (HDDs) in terms of throughput. SSDs and their Flash memory components were built for speed and are likely to provide the lion's share of storage for heavy transactional environments and high-speed database applications. It's in the area of capacity, and particularly the cost of that capacity vs. HDDs, that SSDs could use some improvement.

That's why many of the new models use multi-level cell (MLC) technology, which offers the ability to pack more data into each Flash cell, although with a slight throughput penalty. IBM is said to be aggressively pursuing MLC technology as it pushes the development process down to 25 nm and smaller. According to a leaked roadmap for the coming nine months, current 34 nm devices are being reworked for 25 nm with a resulting increase to up to 400 GB. There are even plans for a 600 GB version of the X-25M drive farther down the road.

SMART Modular Technologies is pushing forward in MLC technology as well, pairing it with a SAS interface in the XceedIOPS device. The combo delivers more than 30,000 P/E (program/erase) cycles compared to maybe 5,000 cycles for straight MLC solutions, which should give the drive a solid five-year lifespan in heavy operational environments. The drive is available in sizes ranging from 100 to 400 GB.

The MLC-SAS combination is also available from Unigen Corp. in the form of the Orion1500 device. Built on the SandForce SF-1565 processor and the Emulex BR-456 SAS-to-SATA bridge, the unit delivers 25,000 IOPS with sequential transfer speeds of 260 MBps over a 6G SAS interface and is available in sizes up to 400 GB.

And speaking of SandForce, the company recently announced that its SF-1000 processor has been clocked at 30,000 IOPs per watt, and since it requires no external DRAM cache, it can support up to six flash packages providing 384 GB on a 20 nm MLC process. The system is being targeted at everything from enterprise server, storage and RAID card applications to appliances and embedded systems.

Clearly, the trend in SSDs is following a smaller-process/greater-capacity paradigm that probably will take a few years to play out. That doesn't mean you should hold off on SSD deployments now in the hope that something better will come along later. Rather, it means that SSD infrastructure in place today should see a steady stream of improvements in speed, capacity, power consumption and a host of other factors as the technology matures.

And it's coming just in time for enterprises to retool their data environments in anticipation for the next upswing in the business cycle.

Add Comment      Leave a comment on this blog post

Post a comment





(Maximum characters: 1200). You have 1200 characters left.



Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.