Most of the containers being employed today can be found in web applications that are, for the most part, stateless. But that’s only the beginning. A new survey of 648 IT professionals conducted by Portworx, a provider of data storage management software for containers, finds that more than half (55 percent) are planning to deploy a database inside a container as well.
Of course, databases require that containers be stateful. The survey finds that the primary reason IT organizations have not already deployed a database in a container is that they lack sufficient tools to manage them (56 percent).
In fact, 37 percent identified a storage solution could provide is the direct provisioning of storage for containers as being the most important tool they need, followed by converged infrastructure (23 percent), hybrid capabilities (20 percent) and DevOps integrations (19 percent). None of those capabilities are available using legacy storage solutions, the report claims.
The report also finds IT leaders have concerns about data loss (42 percent), scale (34 percent) and speed (30 percent). Put it all together and it’s clear that there’s desire to run databases in a container—it’s just that IT professionals currently have some doubts about their ability to successful.
Driving all that interest, the survey finds, is a need for increased agility (75 percent), reduced costs (53 percent) and performance (40 percent).
On one level, this interest in running stateful applications follows a natural trajectory. The survey results indicate batch processing jobs (44 percent) and big data workloads (32 percent) are next for containers. There are, of course, a lot of vendors with a vested interest in slowing the rate of this transition. The faster developers make the shift to containers, the freer they become to host their applications anywhere they see fit.
Containers may not be a mature technology yet, but they soon will be. In fact, a decade ago or so many of the same comments being made about the immaturity of containers were also being made about virtual machines.
Of course, the difference is that containers are more flexible. Today, a container can be deployed on top of a virtual machine, on a physical server or in a platform-as-a-service (PaaS) environment. Most containers deployed in a production environment today run on top of a virtual machine or in a PaaS. That’s because IT teams have yet to master the tooling needed manage containers on a physical server. But as that issue is addressed over the coming year, there will be a lot more containers running on a physical server. That approach should drive up server utilization rates dramatically compared to what can be achieved relying on virtual machines.
Of course, managing the contention issues that might arise from all those containers sharing the same physical resources will be the next big challenge. But as those and other issues get addressed, containers in their multiple forms will become mainstream elements of the data center environment before very long. In fact, the survey found that respondents with less than 20 percent container adoption are already trying databases, batch and, to some degree, big data workloads.
Eric Han, senior director of product management for Portworx, says the pressing need for agility on the part of both developer and IT operations teams will trump all other issues. In fact, the survey finds that as adoption of containers increases to 21 percent to 60 percent, databases constitute the largest share of workloads being containerized. Container resistance clearly is becoming increasingly futile.