Join Newsletter
Forgot
password?
Register
Trusted Business Advisors, Expert Technology Analysts

Taneja Blog

Taneja Blog / Data Center Systems

Storage Performance - Maybe It Never Was the Array’s Problem

In one of the biggest news announcements of 2012, QLogic has announced a major product initiative that looks poised to fundamentally change customer challenges around storage performance. QLogic’s Mt. Ranier is onboard storage-adapter-integrated hardware that can interact with the IO stream to redirect hot data to locally attached solid state cache. The cache may either be SAS-connected SSDs, or a NAND Flash PCIe daughter card that can be installed next to a Mt. Ranier and connected via specialized cable.  QLogic essentially turns your storage adapter – today an HBA, but tomorrow perhaps an Ethernet spewing CNA – into a storage performance accelerator. 

QLogic is not the first vendor to announce products that insert flash performance in the server I/O path as cache. We previously identified this type of approach to storage performance as “Server-based Storage Accelerators” and wrote about in InfoStor (http://bit.ly/GU6UjS). Existing products so far have been flash-based, in-the-server, storage acceleration technologies; primarily PCIe form factor cards with on-board flash memory that plugged into a server motherboard slot. Using drivers or other software inserted into the server OS’s software stack, the products would intercept I/Os, cache data from SAN-attached disk onto a NAND-flash cache, and then redirect future I/O requests to that cache.

These products altogether offer tremendous potential.  They can deliver performance without re-engineering the entire storage infrastructure; because the data is still stored on the array they do not break storage practices; and they are incredibly cost effective.  But these seemingly simple products are still a subtle hotbed of innovation – there is a small battle is on among vendors around how elegantly redirection software can be inserted into the IO path so it has minimal impact on host processing, can keep up with IO extremes, and keeps latency super low.  Marvell Dragonfly, Proximal Data, and SanDisk FlashSoft seem to be doing some serious pioneering here.

But this battle for offload efficiency is part of what makes the QLogic announcement big.  By sitting right in the IO path, QLogic can use their adapter to offload without much host involvement.  On a second front, the announcement is big because QLogic adapters are so ubiquitous in the marketplace.  By combining caching with the adapter, cache capabilities could rapidly become available nearly anywhere, and those products may make it into OEM configurations way sooner that would be the case with any startup vendor.  Storage accelerating adapters may become commonplace real fast, without customers having to purchase them separately from servers and go through the trouble of cracking the box to install them.  Meanwhile, for any host involvement (such as cache flushes when snapshots are taken), QLogic has a highly respected place to integrate that functionality – right in the QLogic unified driver stack that is deployed in most enterprises already.

Looks to me like the Server-based Storage Accelerator just went mainstream.  Time’s right to start thinking where they fit in your environment today, as some will probably show up in your next server acquisition.

  • Premiered: 09/07/12
  • Author: Taneja Group
Topic(s): Qlogic server Acceleration SSD IO Storage perfrmance

Comments

There are no comments to display. Scroll down to leave your own!

 

Leave a Comment

You must be logged in to comment. Click here to log in or register if you don't have an account.