In a recent survey, the results prove beyond doubt that the move towards containerisation is nearly unstoppable. Sushil Kumar, Chief Marketing Officer, Robin Systems, explains why.
“You can observe a lot by just watching” – Yogi Berra, Hall of Fame Baseball catcher, manager and coach. Well known for his malapropisms.
If you can learn a lot by watching, you can also learn a lot by asking. So when we wanted to find out the current thinking about use of containers with databases and other applications, we asked. And 200 IT professionals told us what they think.
Increase in spend on container-based technology
For one thing, the survey results prove beyond doubt that the move towards containerisation is nearly unstoppable. More than 4 out of 5 respondents expected their companies to increase their spend in container-based technology.
What may be a bit surprising to some is the use of containers for stateful or data-centric applications. Containers have gained an incredible level of mindshare among developers, and there is a near unanimity that technology such as Docker’s will soon become the standard way to pack and deploy stateless applications. Opinions, however, can be divided on how useful containers are for performance-sensitive data applications, such as databases. As someone who has spent his lifetime working with databases, I can vouch for how transformative a container-based platform can be for these applications, but we wanted to hear what data professionals and IT managers thought about it.
And this is where the survey results may hold a surprise or two. Three out of 4 respondents told us they are actively looking to run data applications within containers. The drivers cited for this trend were not surprising. Respondents pointed to the ability to consolidate workload (without losing performance or predictability) and to reduce performance ahead (as compared to hypervisors-based virtualisation) as leading factors prompting greater adoption of containers.
In fact container use is growing for all things data. Containers are emerging as the preferred platform for running databases, with approximately half the respondents doing so. About 40% of respondents indicated they have deployed Big Data applications such as Hadoop and Spark within containers.
Vibrant container technology landscape
The other surprising survey finding is the vibrant container technology landscape. While Docker is certainly getting its share of interest and adoption, system container technologies such as LXC and LXD remain as the preferred containerisation technology for running data-centric applications, with 60% citing their use. That is understandable because, unlike application containers technology such as Docker that are designed to run a single process or service, System Containers are essentially lightweight VMs that can run multiple services and applications, have their own host name and IP addresses that you can SSH into, and pretty much manage like VMs.
The use of System Containers with data applications therefore provides many of the benefits of a hypervisor-based virtualisation but with bare metal performance and much lower management overhead because of shared OS and binaries.
We tried to get a feeling for how far along the respondents’ organisations are in their adoption of containers, and a majority are using them to some degree. Containers are already used in production at 35% of respondents’ companies, and another 26% are experimenting with them.
So what do we make of all this? For one thing the findings are consistent with the approach we have taken with Robin’s containerisation platform. Performance-sensitive applications such as databases should demonstrate superior performance on a containers-based platform thanks to the ability to consolidate without compromising performance or predictability. Robin’s vision is to provide enterprises a high-performance and elastic platform for stateful and mission-critical applications. We think a containerised approach will prove the right one, and it would seem our survey respondents agree.
Edited for web by Cecilia Rehn.