Backup Deduplication 2.0 - Density

As we continue our requirements for the next era of backup deduplication, the next important area for improvement is in the denseness of systems. This means more raw capacity in less physical space. While getting sufficient power to the data center is a problem for some data centers, the lack of available data center floor space is becoming a problem for even more of them. Backup deduplication systems need to help address that pain by increasing density.Running out or low on data center floorspace is often more painful than running out of power. As long as you can afford more power, in many areas of the world you can keep getting it and the cost for more is incremental. Space on the other hand is harder to come by. Even if you can get and afford new land, it takes time to build a new data center and the cost to build that data center is anything but incremental. The solution is to increase the density of all the technology in a data center. In storage this means more capacity and performance in less space. Backup deduplication vendors must be prepared to participate in this discussion.

The current answer to increasing density from backup deduplication system vendors is that they provide compression and deduplication so they are already space efficient. They may also be counting on ever increasing drive sizes to improve density. While these are fair claims, they are lacking as density becomes a bigger problem. As we discussed in our article The Storage Efficiency Triangle, building a dense storage system is more than just adding deduplication and compression. It requires the tighter packaging of the hard disks themselves. The standard 12 - 20 drive slot shelf is not going to meet the density demand. There is a lot of open space in the typical storage shelf and that space needs to be used.

Of course greater drive density is going to mean a build up in heat as well as vibration, both of which can be deadly to drives. Heat dispensation has to be handled by more than just bigger, faster spinning fans, which would just consume more power and increase noise levels. You don't want to loose power efficiency to gain space efficiency. Vendors will have to spend more engineering time working on proper airflow designs so that power efficient fans can be used. The way that drives are mounted in the storage system can help this. Also the drive mounting techniques can lower vibration issues as well. This means potentially not using off the shelf hardware but instead vendors are designing their own hardware that meets these goals.

Density could also be increased by using 2.5" form factor hard drives instead of the 3.5" form factor drives that are commonplace in backup deduplication. While you do give up some capacity per drive and the cost per GB increases slightly, 2.5" drives give you much better density and in most cases much better power efficiency. This also may give you something that we will discuss in detail in an upcoming entry; better rebuild times. Rebuild times on backup deduplication devices tend to be long and in some cases can dramatically impact performance. By using more drives at a smaller capacity, backup deduplication systems built on 2.5" drives should deliver more effective rebuild times and be less susceptible to a major performance loss during the rebuild process.

Data center floor space is at a premium and deduplication systems have to help address this challenge by moving beyond the denseness through technology that deduplication and compression provide. The hardware now needs to be looked at so that further improvements in GB per tile can be made.

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

Subscribe to our RSS feed.

George Crump is lead analyst of Storage Switzerland, an IT analyst firm focused on the storage and virtualization segments. Find Storage Switzerland's disclosure statement here.