Infinidat Promises Storage Bigger, Faster, Cheaper
Want large quantities of high-performance storage at a fraction of the usual cost? That’s what Infinidat is promising its customers.
As Infinidat tells it, the evolution of storage technology has been to develop smarter and smarter hardware, processes operations at the behest of “dumb” software. With this traditional approach, getting more high performance storage means throwing more money at the problem to buy larger and larger quantities of flash.
Infinidat developed a software solution to what’s traditionally been a hardware problem. Their storage technology uses an MVL trie index combined with machine learning software to create what they call a “neural cache,” which makes predictions about which pieces of data are likely to be accessed by your workloads.
The storage solution works by deploying a DRAM and NAND flash as a transparent cache in front of an NL-SAS disk. The machine learning algorithm adapts to your particular workload over time, increasing the efficiency of the cache until 99% of reads and 100% of writes can be served by DRAM.
The result is storage that performs like a petabyte of RAM, all the while running on commodity spinning disk.
Infinidat has described this technology as providing ten times the performance at one-tenth the cost of traditional solutions.
Our Take
If Infinidat’s storage solution works as well as claimed, at such high volumes, it will solve a key pain point for many enterprise IT infrastructure managers, as well as cloud service providers.
Storage is expensive, and the demand for storage is continually growing – driven by the unruly behavior of end users, developments such as the rise of big data, and the exponentially increasing quantity of data produced by today’s digital society.
Bigger, faster, cheaper storage should provide benefits to all of us. We’ll have to see if Infinidat’s technology developments remain a closely-held differentiator, or if others in the industry are able to build comparable solutions that will drive down costs across the board.