One of the big trends that’s emerged when I’ve talked to IT leaders of midsize companies over the past couple of years is that in today’s digital-first world they’re moving away from trying to individually build and optimize many different systems, each suited for a particular workload. Instead, they want a scalable infrastructure that can fit together seamlessly and be much more easily integrated and managed.
This is really important. With the relentless march of technology, we have never had it so good when it comes to the new capabilities, performance, reliability, and cost of modern midrange IT infrastructure. There is just so much potential to be had, but accessing this to the full becomes a major challenge in all but the simplest of solutions.
As we move away from traditional client-server architectures to distributed cloud-native approaches, we’re moving out of an age where we can really come to grips with managing systems or devices individually. Services are becoming composed of chains of interconnected applications and systems, and they all need to play well together. No matter how good a device may be when considered on its own, if it doesn’t fit in well, it will cause bigger problems than it solves.
IDC research shows that too many companies still rely on the default tools that come with their application or appliance, and this leads to expensive and error-prone configuration and change management. Having a solid strategy on integrated management and automation is essential. Anything on your infrastructure short-list, and especially your data storage systems, should place this high on the selection agenda.
Integrated and comprehensive support for common open management APIs is critical. Having proprietary features that can only be accessed or managed via custom consoles will typically mean these features are never used in reality or will cost significantly more over the lifetime of the device. Devices also need to be an active participant in infrastructure optimization, sharing operational data seamlessly so that they can be part of an intelligent AIOps approach, enabling the infrastructure to adapt easily as demands and workload needs change.
Where are you on your approach to programmable infrastructure? Are you mainly looking at a single glass of pain, or a single pane of glass? To learn more, read the IDC Market Spotlight
Andrew Buss is Research Director at IDC.