Latency vs. Capacity Storage Projections 2012-2026

Wikibon believes latency storage vs. capacity storage is a key storage dimension, with different functional requirements and different cost profiles. Latency storage is found within the datacenter supporting more active applications, and in general has a high read bias. Latencies can vary from 1 millisecond down to a few microseconds; the lower (better) the latency the closer to the processor resources it is likely to be. It is also found for the metadata layer for capacity data. The boundary for latency storage will reduce down to 500 microseconds over the next three years.

Capacity storage is found in archive, log, time-series databases for the Internet of things and many other similar applications. In general it is write-heavy. Latencies are generally above 1 millisecond, do not have to so close to the processor, and are suitable for remote private, public and hybrid cloud storage. Some parts of the capacity market place will have latencies as low as 500 microseconds over the next three years.

Wikibon has added this dimension to the other storage dimensions projected, which include HDD vs. Flash, Hyperscale Server SAN vs. Enterprise Server SAN vs. Traditional SAN/NAS storage, Physical Capacity vs. Logical Capacity and SaaS Cloud vs. IaaS Cloud vs. PaaS Cloud. All these dimensions are projected for both Revenue and Terabytes. There is a strong correlation and interaction between the latency/capacity dimension and HDD/Flash dimension.

Wikibon provides detailed breakdown of the storage projections for Premium clients.