Keys to the CASL for Nimble worth considering

In our last post, “Flash for Cache, just not for Writes” we talked about Nimble Storage’s approach to accelerating read performance using a true SSD Flash cache. Simple, elegant, FAST. But what about increasing write performance?

Write BIG, Write FAST

Nimble Storage solves the disk I/O write bottleneck by once again identifying the problem and applying new, inexpensive technologies that we unavailable 10-15 years ago. The solution is based on what they call Cache Accelerated Sequential Layout, or CASL.

At the core of CASL write performance story is a design structured around Write Optimized Data Layout. When throughput is required for spinning media like SATA drives, the best way to push performance is to write in long, continuous streams. Small, random writes spawn head seeks and drastically reduce throughput.

To solve this problem, Nimble Storage’s CASL “collects or coalesces random writes to the array, compresses and writes them sequentially to disks. This results in write operations that are as much as 100x faster than traditional disk-based storage.” The key to CASL is that the system uses DRAM to facilitate the collection and consolidation of data which is extremely quick.

All this adds up to a relatively inexpensive array that competes head-to-head with gear twice the cost.

I wish there were more complexity to it, but sometimes the smartest ideas are simple!

cs200 button_nimblepricing

  • Share/Bookmark