Deployed transparently between server and storage
Learns applications’ data access patterns
Caches performance-critical data in RAM and FLASH using the data model with Bayesian analytics
Application read I/Os are serviced with microsecond latencies as though they were in a RAM cache. All writes go to the storage subsystem to ensure availability
Custom Silicon – For mapping of accesses and gathering of real-time statistics
Real Time Data Access Model – Updated as the actual program execution changes
Flash Management – The performance and read/write ability of RAM with the capacity of FLASH
