When we talk about creating a highly available, distributed system, Kubernetes is pretty much the first platform that comes to mind. Born from the need to address the complexities involved with managing containers at scale, Kubernetes has gone from orchestration tool to the “control plane” of the enterprise in record time. While Kubernetes being open-sourced began as a nightmare for those already in the space, like Docker and VMware, the complexities involved with Kubernetes management have created an ecosystem full of opportunities for innovation. Here are five companies leading the way.
1. MayaData Kubera
There is quite a lot of competition in the field of Kubernetes management at the moment, leading to several vendors combining their offerings to increase value. MayaData Kubera is one such example that takes the weakest link in containers, storage, and builds its offerings around it. Kubera is wrapped around the leading open-source Container Attached Storage (CAS) OpenEBS. CAS allows Kubernetes to automate the management of storage, as well as provide granularity concerning storage policies. OpenEBS, like Kubernetes, is a CNCF project, and Kubera ships with a special Enterprise Edition that includes 24/7 support.
Another example of an organization being innovative and combining its offerings is PlanetScale, which is combining Kubernetes and database management into one single service. Similar to how Kubera is built around OpenEBS, PlanetScale is also built around a CNCF graduate project called Vitess. Originally invented to help scale YouTube, Vitess can be used to massively scale MySQL instances that on their own run out of steam at a little over a terabyte. Essentially a “sharding” middleware, Vitess allows you to grow your database horizontally without adding sharding logic to your application.
While PlanetScaleDB eliminates the need to deploy and manage both Vitess and Kubernetes, what’s really innovative about it is the BYOK8s, or bring your own Kubernetes, concept. Rather than impose a specific version of Kubernetes on consumers, PlanetScaleDB lets you bring your own version of Kubernetes and then manages it for you, alongside Vitess. This is especially useful for organizations already building MySQL applications and looking to scale up without a dedicated NoSQL database. PlanetScaleDB also provides a single interface or virtual control plane for organizations to manage databases across multiple clouds.
DevOps teams love automation, and that’s the basic idea behind our next offering on the list. Originally named Loodse, Kubermatic takes an automated and economical approach to manage Kubernetes. It does this by automating as much of it as possible while claiming to bring down operational costs by as much as 72 percent. Kubermatic is a multicloud and edge platform for Kubernetes that allows users to automate the management of Kubernetes clusters regardless of where they exist. This is especially useful for organizations with distributed architecture like hybrid/multicloud, on-premises, IoT, and edge environments.
In addition to being a truly cloud-native approach to managing Kubernetes, Kubermatic was also just open-sourced under the Apache 2.0 License, making it a lot more desirable in the process. Kubermatic simplifies Day 2 operations and also features automated lifecycle management via an API call, multi-tenancy with predefined environments, and infrastructure logging and monitoring with Prometheus and Grafana. Lastly, Kubermatic supports all major public cloud providers, including a recent update that added support for KubeVirt. KubeVirt is an open-source project that runs VMs on Kubernetes, allowing developers to bring their VMs into Kubernetes workflows.
4. Spectro Cloud
Yet another vendor taking the hybrid/multicloud approach to Kubernetes management is Spectro Cloud, which just announced the release of Spectro Cloud, a hybrid platform for Kubernetes infrastructure management. This is also another example of the “dual” approach we’ve been looking at with the other vendors since this effectively combines management and operations into one single offering across public clouds, on-prem, bare-metal, and hypervisors. Claiming to hit the sweet spot between a managed service and a DIY approach, Spectro Cloud looks to strike a balance between getting help with Kubernetes and giving up control over your infrastructure.
Similar to PlanetScale’s BYOK8s, Spectro Cloud prefers to let users deploy their cluster where they like, with their own version of Kubernetes. This allows developers a chance to play around with different options as opposed to being stuck with a single cloud vendor’s approach. Spectro Cloud achieves this by building a custom Kubernetes stack based on the organization’s chosen approach and business needs and then automating deployment and maintenance accordingly. While this is Spectro Cloud’s first product to hit general availability, what’s interesting is that on the same day as the release, it received a top-three ranking by Enterprise Management Associates (EMA) in the hybrid cloud management / managed hybrid Kubernetes category.
5. HPE Ezmeral
Everyone seems to be offering Kubernetes management as a by-product of some service or the other, so why not get it as a by-product of your hardware? Yes, we’re in the era of the public cloud, and no, hardware servers aren’t extinct yet. Taking a hardware approach to managing Kubernetes is HPE’s new Ezmeral Container Platform that’s being advertised as a “pay-as-you-go” approach to deploying Kubernetes anywhere. While the idea here is to supply ready-to-use building blocks in the form of HPE servers built for co-location and the edge, Ezmeral can be used to manage containerized applications across any infrastructure, including bare-metal, VMs, public clouds, and non-HPE hardware.
Ezmeral is a rebranding of MapR and BlueData, both HPE acquisitions, and is said to take aim at hardware competitor Dell’s Tanzu and IBM’s OpenShift. With Ezmeral, customers can manage multiple Kubernetes clusters with the help of MapR’s data fabric for Kubernetes. Another advantage is HPE’s claim that users can also run non-cloud-native applications without the need to rewrite code. Ezmeral ML Ops is the other half of this offering from HPE, which is based on BlueData and aimed at organizations running AI and ML workloads. HPE’s claim is that it can shorten AI deployment time from months to days.
Kubernetes management: Miles to go before we sleep
While the big players in the Kubernetes management field are still AWS, OpenShift, Rancher, and Pivotal, what’s been achieved so far in the field, according to many, is still just the tip of the iceberg. From managed Kubernetes to storage-based Kubernetes, to bring your own Kubernetes, and even pay as you go Kubernetes, the options are virtually endless. What’s consistent, however, is the fact that there are endless amounts of Niches available in the Kubernetes ecosystem, and all you have to do is find that unique angle. That being said, the writing is on the wall, people want quick and easy to use Kubernetes that runs anywhere, without any fuss and leaves them free to focus on the more important tasks on hand.
Featured image: UIhere