Industry executives and experts share their predictions for 2019.  Read them in this 11th annual VMblog.com series exclusive.

Contributed by Haoyuan Li, CEO and co-founder, Alluxio

2019: Smart Enterprises Decouple Storage and Compute

Decoupling storage and compute will be a top priority for all enterprise CIOs in 2019. It's simply crazy, and very expensive, not to. The costs of not acting go beyond wasting precious budget on over-provisioning and data duplication. Enterprises that fail to invest in decoupling data and storage will also lose competitive advantage.

Every Global 2000 enterprise supports multiple datacenters, typically in dispersed geographies. Public clouds are appealing for many good reasons but no CIO is moving all of her data into a multi-tenant public cloud. While a key advantage of big data technology is the ability to collect and store large volumes of structured, unstructured and raw data in a data lake, most organizations only end up processing a small percentage of the data they gather in the first place. According to recent research from Forrester, as much as three-quarters of data that businesses store ends up not being processed. So most data sits in isolated silos and is never put to work to help the business.

Decoupling storage and compute can help enterprises use more data better.

First of all, no CIO can really predict the future analytical needs of her business. The right storage needs today might not be right for next year or even next week. With exploding data volumes, workload requirements might jump 10-fold in a week, month or a year. This is a problem you want to have! But provisioning storage co-located with  compute resources for worst-case scenarios is costly and rigid. Smart enterprises in 2019 will ensure that their technology stack for data-driven workloads provides freedom from underlying data infrastructure and significantly more flexibility for unpredictable and bursty compute workloads. This new stack mush support the business imperative to pursue cloud economics across all flavors of deployment - on premise, multi-cloud, hybrid cloud.

Optimizing analytics for the cloud involves placing data on reliable and durable shared storage, separating storage from compute, while matching or improving the performance requirements of existing workloads without replatforming the workload. The flexibility and cost savings can be game changers when you can tie costs directly to business needs by provisioning just the right amount of compute resources to just the right amount of storage for any workload.

Also, big data workloads are often "bursty", requiring data teams to provision their resources to peak capacity at all times, resulting once again in resource underutilization during longer-lasting off-peak usage hours.

The ability to scale up and down a cluster elastically to accommodate data-driven analytical and machine learning workload peaks and troughs results in lowered infrastructure cost and maximum business value. When separation of compute and storage is architected with cloud economics in mind, enterprises gain greater operational simplicity and workload isolation to better meet SLA and business requirements.

##

About the Author