As IT organizations that have adopted Kubernetes head into 2020, their next big challenge will be deploying Kubernetes clusters at scale.
A survey of 1,300 attendees of the recent KubeCon CloudNativeCon 2019 conference conducted by Platform9, a provider of managed services, finds about 32% of respondents said they expect to be running 50 or more Kubernetes clusters in production within the next six months.
Nearly half of the respondents (49%) identified monitoring those clusters at scale is their biggest challenge, followed by upgrades (44%) and security patching (35%). In addition, 65% of the respondents said they would require 99.9% or higher uptime, with 56% noting their organization needs to provide a 24/7/365 follow-the-sun support model for highly distributed Kubernetes environments.
According to the survey, 80% of respondents said they have deployed Kubernetes on a public cloud, while 50% said they have deployed Kubernetes in an on-premises IT environment. Another 12% said they have deployed Kubernetes in an edge computing environment. Of those that have deployed Kubernetes at the edge, well over a third (39%) said they are running Kubernetes in 100 or more locations. Nearly half (47%) of those respondents said they are running 11 or more servers in each location. The top two applications being deployed at the edge are edge gateways/access control (43%) and surveillance and video analytics (32%).
Kamesh Pemmaraju, head of product marketing for Platform9, says the survey results not only confirm the fact that Kubernetes is being deployed in multi-cloud computing environments, but organizations also appear to prefer to deploy many smaller clusters versus choosing to rely on one or two clusters. The primary reason for the decision generally comes down to application owners preferring to retain control of an entire cluster. However, Pemmaraju notes, it’s also a lot easier to troubleshoot a small cluster than a larger cluster running multiple applications.
It’s still not clear to what degree IT organizations will prefer to rely on managed Kubernetes services. In the case of the cloud, IT organizations can rely on managed services provided by the cloud service provider or a third party such as Platform9. However, unlike a cloud service provider, a third-party provider will also manage an on-premises Kubernetes deployment regardless of the IT infrastructure it’s deployed on. Historically, managed services have accounted for less than 20% of the total IT market. But given the shortage of Kubernetes expertise, many organizations may have no alternative than to rely more on external expertise. The issue many organizations will need to decide is to what degree they want to continue to rely on managed Kubernetes services should they decide to invest in training their own IT staff.
Of course, any survey of attendees at a conference dedicated to cloud-native platforms is likely to find a high rate of adoption of Kubernetes. However, as more developers build cloud-native applications in 2020, it won’t be long before most IT organizations find themselves trying to figure out one way or another how to manage a Kubernetes cluster, especially Kubernetes at scale.