Write back caching (SSD) – write operations are directed to SSD for immediate acknowledgement back to the server, and the data is written from SSD to hard disk at a later time. This accelerates write operations, but also allows faster read operations of data in the SSD cache.

Predictive read ahead caching (SSD and memory) – Identifies sequential read I/O patterns, pre-fetching subsequent related data from disk into memory. It also addresses the virtual server “I/O blender effect” by using sophisticated algorithms that can identify patterns that are interleaved I/O requests from multiple different VMs.

Data pinning – Frequently repeated operations, such as booting virtual machines, or end of month processing can be mapped and loaded into memory from disk ahead of time, allowing the memory cache to be ‘pre-warmed’.

StorMagic’s Predictive Storage Caching features allow organizations to build a solution which uses multiple storage types as a cache and delivers exceptional performance in a cost-effective manner. Compared to a system with no storage caching, Predictive Storage Caching can help organizations achieve significant performance gains and considerable reductions in CAPEX and OPEX. At a basic level, in many StorMagic use cases, organizations have discovered that simply adding a small amount of SSD as cache on top of their existing hard disks has dramatically improved performance, increased the life of their hardware and saved money, all by leveraging the caching features that StorMagic has to offer.