Today's post is a brief summary and highlights of this article written by Sudhir Srinivasan, SVP and Storage CTO at Dell EMC and member of Forbes Technology Council.
Petabyte scale storage environments are quickly becoming the norm, and the technologies which support such massive storage demands have been successful in evolving to provide adequate performance and scalability. The one area, however, which has not seen as rapid of development is how these growing infrastructures are operated and managed. As exascale storage environments become a reality, there is a growing need to address the massive amount of human resources that will be needed to effectively manage these environments. The answer, according to Srinivasan, is extreme automation.
“We have solidly embarked on the journey to the "self-driving" storage system: a fully autonomous system capable of adapting its behavior in response to changing workloads.” says Srinivasan. There are four levels of "self-driving" autonomy of storage systems that CIOs and CTOs should be actively looking for:
Level 1: Application-Centric
“The interaction with the storage system at this level is in terms of the application the user wishes to run rather than what the storage system needs to do to run that application.”
Level 2: Policy Driven
“Policy-driven automation is achieved by setting service-level objectives for the application that a storage system was directed to run.”
Level 3: Self-Aware
“At this level, the system performs the complex functions of an enterprise storage system, including the allocation of central processing units (CPU), memory and IO bandwidth across thousands of active IO requests every instant. In other words, the system understands what is going on at the macro level and adjusts across workloads.”
Level 4: Self-Optimizing
“We may be a long way from general-purpose AI, but as the explosive growth of data continues, autonomous operation will become the new norm, up-leveling the capabilities of the data storage systems, both in the cloud and in enterprise data centers.”
We may be a long way from general-purpose AI, but as the explosive growth of data continues, autonomous operation will become the new norm, up-leveling the capabilities of the data storage systems, both in the cloud and in enterprise data centers.