• E-mail this page
• |  Print this page
• | 
• | 
• | 

SSD Can Mean Hard Cost Savings

In our last entry we talked about the time savings and potential increase in productivity and revenue that deploying SSD can enable. This entry we will focus on the hard cost savings associated with SSD. In the right situation, SSD can actually be less expensive than mechanical drives.

The first misnomer when comparing costs between SSD and mechanical drives is often those comparisons are to the cheapest SSD capacity vs. the cheapest drive capacity. This simply is not fair. SSD essentially has two classes of storage; Flash SSD and DRAM SSD and both deliver significantly better performance than tier one mechanical storage.

To get increased performance out of mechanical drives workarounds have been developed that are costly from both a power perspective and a physical assets perspective. Replacing these workarounds with SSD can further increase performance while reducing both power and cooling costs. In reality the price comparison should be to the very high end of tier one mechanical storage and should factor in all of these workarounds that people choose when trying to get the maximum performance out of tier one mechanical storage.

Pulling performance from mechanical drives often follows a path that while it increases performance also increases cost. Typically the first workaround for poor mechanical drive performance is to use standard array sets but with 15k RPM fiber channel drives. If this does not deliver the performance required the next step is to greatly extend the drive count, still using 15k RPM drives, but sometimes tripling or more the original set. These array groups are almost always front ended by fast storage controllers. If the system isn't a virtualized storage system that can perform wide striping, the last resort is to short stroke the drives in the array group, which formats the drives so that only the faster edge of the platter will be written to. While this will increase performance, it also severely reduces the addressable capacity of those drives.

In combination with the above techniques, performance-hungry applications will use a very high server count accessing the storage to increase parallelism. These drive configurations are now well suited to a small number of threads.

The result of all the workarounds is high acquisition, power, and cooling costs and of course a high degree of complexity, especially when compared with SSD. Depending on the environment, a choice needs to be made between Flash-based SSD (read heavy) and DRAM-based SSD (write heavy). Especially in write-heavy environments, only a fraction of the data needs to be mounted on SSD. The result is a significantly smaller disk allocation that requires less power and delivers better performance. All of this then reduces complexity.

When compared with a 50+ 15k RPM drive array that is short stroked, front-ended by multiple storage controllers, and requires a high server count for streaming, SSDs can reduce both CapEx and OpEx now. There is no need to wait for further price drop in memory.

Join us for our upcoming Webcast, SSD: Flash vs. DRAM...and the winner is?

Track us on Twitter: http://twitter.com/storageswiss.

Subscribe to our RSS feed.

George Crump is founder of Storage Switzerland, an analyst firm focused on the virtualization and storage marketplaces. It provides strategic consulting and analysis to storage users, suppliers, and integrators. An industry veteran of more than 25 years, Crump has held engineering and sales positions at various IT industry manufacturers and integrators. Prior to Storage Switzerland, he was CTO at one of the nation's largest integrators.

« Yieldex Wins Amazon Startup Contest | Main | Fedora 10: Building A Community, Not Just Code »