Artificial Intelligence for IT Operations (AIOps) makes it possible that after offerings such as Software-as-a-Service (SaaS) or Infrastructure-as-a-Service (IaaS), there is now also Storage-as-a-Service (STaaS). A guest article explains Nevzat Bucioglu, Country Manager Germany at Infinidat.
Hardly any other technology is pinned on as much as artificial intelligence (AI). This should help recognize and cure illnesses, bring vehicles autonomously and without accidents from A to B, or check identities online and prevent fraud. Some of this is still a long way off. In IT, however, AI in the form of Artificial Intelligence for IT Operations (AIOps) finally makes it possible for offers such as Software-as-a-Service or Infrastructure-as-a-Service to now also include Storage-as-a-Service (STaaS) exist, even with capacities of many petabytes. AI is used on several levels.
In the box
To benefit from STaaS, users should ensure that it doesn’t just say “AI” on the box but that Deep Learning (DL) is used. Only DL enables a storage system to perform the self-optimization required for STaaS. DL is a sub-area of machine learning and goes beyond it in one essential point: A DL system is based on the functioning of the human brain and cannot only be trained but also learn on its own. The system constantly adapts what it has already known based on new inputs and thus continuously optimizes itself.
DL can optimize application environments and performance over time when deployed in a storage array, providing essentially maintenance-free operation after a one-time installation and configuration. It dynamically adapts to changing application, user, and performance requirements—without administrative overhead. For systems with multiple types of storage media, DL can also ensure that the system selects the optimal media for each job. This enables IT departments and service providers to meet even the most demanding SLAs.
Outside the box
However, a natural STaaS environment has not yet been achieved. To do this, AIOps should also be deployed outside of the array and leveraged to provide predictive analytics, early problem detection, and enhanced proactive support. In addition, the storage environment should not form an AIOps island in the infrastructure. Instead, it should be combined with corresponding offerings from other providers so that companies can use cloud automation, orchestration, and AIOps platforms, for example. The storage system must be equipped with the necessary interfaces (APIs). Companies looking for a STaaS solution should pay attention to the technical equipment of the array under consideration, but also APIs and a connected ecosystem of partners.
Flexible pricing models
In addition to the technical requirements described and the vital ecosystem, the pricing model for STaaS is still of central importance. After all, with a service like electricity from the socket, you only pay for the electricity you use, not for 200 KWh if you currently only need 50 KWh, just because you want to be prepared for the eventuality that the need arises is higher than planned. In addition to a pure OpEx model, the storage provider should offer a more elastic model that goes beyond the well-known “pay-as-you-grow model” and combines CapEx and OpEx pricing to provide a cost-effective and fast Enabling storage capacity expansion as well as reduction.
All good things
Using AIOps in the described combination with a connected ecosystem and an elastic price model enables Storage-As-a-Service. A system that optimizes itself and thus offers the best performance and use of the storage media; that recognizes capacity bottlenecks and works without failures with predictive analyzes and proactive support; and last but not least, with its price model, it offers the flexibility of the public cloud in a private cloud.