Here Comes Automated Storage TieringAt Storage Networking World, at least one new category in storage is coming to the forefront; Automated Storage Tiering. These are typically devices that can sit in front of your existing storage platform and allow some of it to leverage a high speed solid state front end without you manually having to move data to a Solid State Disk (SSD).
At Storage Networking World, at least one new category in storage is coming to the forefront; Automated Storage Tiering. These are typically devices that can sit in front of your existing storage platform and allow some of it to leverage a high speed solid state front end without you manually having to move data to a Solid State Disk (SSD).The companies in this space include Avere, Dataram, Gear6 and Storspeed. There are other forms of these products from Compellent and NetApp that either can do dynamic data placement or can provide a large SSD as a front end cache. These solutions however are limited to a single storage system and only work on the respective manufacturers products.
Integrating stand alone SSD or even SSD as a drive type in a storage system is achievable as we state in our recent article "Integrating SSD and Maintaining Disaster Recovery", but what if you don't have the time or what if your environment simply is too difficult to apply these procedures?
Automated Storage Tiering attempts to address the challenge of integration and provide increased optimization of tier 0 storage. After all if you are paying 15X the cost for your tier 0 storage, you don't want even 10% to be empty if you can avoid it.
The basic architecture behind automated tiering is to place the system inline between the storage and the servers. It then will act as a large, in some cases, giant, cache or I/O accelerator. Most of the systems today are DRAM based and then are adding flash SSD as a second tier. Typically the solutions will automatically cache reads and writes to the device. There are options for the storage manager to hard set or hard exclude certain data sets from the cache.
One of the differentiations between the automated tiering systems is going to be on protocol. Some will be block I/O focused and others network file system focused. The network file system storage devices may be able to provide greater analytics. There will also be debate around how the device is really used. Is the device a big cache that simply accelerates the current storage or is the device really a storage platform that leverages the legacy storage as the slowest tier in its available storage tiers?
Automated Storage Tiering is now a legitimate category. How we use this category to help IT to meet continuing storage I/O performance demands vs. some of the other options available to us is the next step.
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.