Rightsizing Tiered Storage Systems

A multi-tiered storage system with automated data movement provides the best solution for managing the data explosion IT is experiencing. While tiered storage strategies can cut enterprise data storage costs and address storage capacity issues, rightsizing the storage tiers is a difficult exercise in many environments.

The purpose of this lecture is to go over the commonly used tiering estimating algorithms and methods (usually based on IO skew calculation) and explain their shortcomings in different workload contexts (cyclical data workloads specific to telecom industry, high performance workloads, etc.), as well as propose a new storage tiering estimation methods which attempts to solve these issues and provide more accurate estimates.

Unpredictable rate of storage growth and fluctuations in data rates often lead to performance issues. Automated storage tiering software can solve this problem and optimize storage allocation for performance if the sizing of the tiers is estimated right.

The impact of the following factors that are usually overlooked on the tiering mix will be discussed in this lecture: data movement speed and tiering overheads, storage based replication, snapshots and clones, IO size, Sequential vs. Random IO, etc.