What you need to know about Kubernetes enterprise platforms
Without a doubt, Kubernetes has
become one of the best tools for container orchestration. It's an open-source,
extensible, and modular platform that offers many functionalities. More info at
Considerations of Kubernetes
enterprise platform
There are two types of shared
clusters--single-tenant and multi-tenant.
The former is dedicated for use
by a single team and features a simpler model. This is because all API access
and codes come from one team.
On the other hand, a multi-tenant
cluster is shared by multiple tenants or application teams. They access the
same cluster and are only designated a namespace that defines one team from
another.
Whether you use a single-tenant
cluster or a multi-tenant cluster depends on the size of your organization.
A small company would benefit
from each team building, managing, and owning its own cluster. Medium to large
enterprises, however, would be more efficient with a single platform team
handling multi-tenant clusters.
Between the two types of clusters
in a Kubernetes enterprise platform, multi-tenant clusters are more advantages
over the other kind.
This is because…
Costs are shared among application teams,
effectively reducing maintenance overhead.
Cluster nodes are easier to fill up with
pods, allowing for better resource utilization and efficiency.
The proliferation of versions is reduced
because the version and upgrade schedule is consistent.
Add-on features and security policies are
contained in a single consistent platform
Integration with existing enterprise tools
is easier.
Moreover, there is dedicated
support staff available to assist organizations in various functions, including
Kubernetes deployment.
However, there are concerns about
security, access control, and the use of resources with multi-tenancy.
By following Kubernetes best
practices in building, owning, and managing multi-tenant clusters, such
concerns are addressed.
Some of the best measures to take
are:
Regulate access to resources by
implementing role-based access control.
Ensure that pods are running with
appropriate privileges by implementing appropriate security policies.
Limit each team's access to resources and
the ability to auto-scale clusters by enabling resource quotas.
Ensure pods communicate with each other and
with other network endpoints by implementing network policies that enforce the
process.
Provide pods that access cloud resources
with IAM credentials.
By taking these measures, you can
use multi-tenant clusters to your advantage.
Kubernetes enterprise platform:
Managed vs DIY
Managed Kubernetes is delivered
as a service by either:
Service providers: Docker Enterprise,
Platform9, Redhat Openshift, etc.
Cloud providers: Amazon EKS, Azure AKS,
Google GKE
DIY Kubernetes, on the other
hand, means you manually do the following:
Download Kubernetes
Set up the open-source orchestration tool
Run the platform or use an open-source
tooling
While it is difficult and
time-consuming to run DIY Kubernetes, the managed option comes with its share
of issues.
For instance, authentication
can't be customized or set as a single sign-on. You must first authenticate
with the cloud provider.
Also, you still need to own and
manage the worker nodes, instead of relying on the orchestration tool to manage
clusters fully.
For any issues, however, there
are existing solutions.
Consult with Kublr, a provider of
a comprehensive container orchestration platform that is optimized for
enterprises, especially large ones.
Comments
Post a Comment