The rise of containers in enterprise IT
FYI, this story is more than a year old
The rise of application containers on enterprise roadmaps has been hard to miss, and it illustrates the rapid and deep impact that the combined power of cloud computing and open source can have on the IT industry.
As with other emerging technologies, application containers started with rip-and-replace ambitions before settling for an “embrace and extend” approach to legacy investments, creating an environment that is increasingly multimodal, rather than bimodal, at both application and platform levels.
The more IT operations (IT ops) people get involved, the faster the shift to embrace and extend needs to be. Application containers have started to get IT ops’ attention as the market moves toward container management platform–based production systems, although developers are the ones who continue to drive container and container management platform (CMP) adoption in a cloud-native application context.
In this context, IT ops have a strong role to play, if only to make sure enterprises remain in control of the various, increasingly container-centric, infrastructure-as-a-service (IaaS) and platform-as-a-service (PaaS) platforms they use.
Rather than moving everything from monolithic to microservices architecture (MSA) applications and every hybrid in between to a single platform, enterprises will mix and match increasingly container-centric platforms.
Start-ups usually adopt containers in one fell swoop to underpin a mix of use cases. These include the deployment of fairly standard large-scale environments (100 copies of the same workload, for example), cloud implementations with a particular focus on hybrid clouds, MSA/API-centric application design, and continuous integration/delivery/deployment (CI/CD) processes underpinned by a DevOps culture.
Enterprises cannot be so ambitious. They need to pick their battles carefully in the context of a clear business case. Because their starting point is development and test rather than production, CI processes would appear to be good first candidates.
Enterprises do not need to do it all in one go, and should simply start by using containers in order to reuse existing IT resources for test purposes, then move on to overhaul the build processes. Whichever starting point they choose, IT operations need to be involved from the start to help define and standardise container images and related management processes, for example.
Container/CMP interest and confusion are rising in equal measure. As a result, many will make easily avoidable mistakes as they struggle to understand how and to what extent containers and CMPs relate to existing IaaS, PaaS, and CM investments.
The way they do so is often complex, but successful adoption requires proper explanation rather than rip-and-replace shortcuts. Some expect IT ops’ influence over container adoption to grow as production system–related container technology and functionality matures.
Others see containers as yet another medium by which developers are increasing their influence over IT infrastructure decisions. Both points of view are valid in the context of application containers.
Unfortunately, most vendors are so busy wooing developers that they have hardly anything to say to IT ops. IT ops are here to stay. Vendors need to not only help them understand container technology but also explain how containers will impact their roles, tools, and processes.
Article by Laurent Lachal, Ovum Analyst