Multicloud Kubernetes as a Service

Digital Transformation is still in everyone's mind. Not surprisingly, organizations seek to improve developer agility with the hope that they can react faster to market changes. This is specially important in the current geopolitical and economic environment. Kubernetes is gaining more popularity as it helps them shorten software’s time-to-market.

Organizations that adopt a multi-cloud strategy for Kubernetes have a comparative advantage as they can place workloads in the location that best meets their requirements. This strategy is also tames the fear of vendor lock-in. The concerns with Public Cloud don't end there. There are also performance limitations and challenges to meet ever increasing compliance. We see every day mew testimonials of public cloud customers that are repatriating their workloads due to massive overspending. A multi-cloud approach can help them keep costs down and predictable

Most of these challenges are easily addressable with private infrastructure, but end-users expect to consume it in a self-service manner and with the same agility they get from public cloud. The key technology that can make this happen is Infrastructure as Code. In the "IaC Avengers" team we have demonstrated before how to do this for many use cases and across all products in the Dell Technologies portfolio. You can find our demos in our YouTube channel

However, we have now taken this one step further. Our customers want a multi-cloud future where their end-users can choose the right location for their workloads by themselves and in an agile way. This is exactly what this demo shows.

Dell partnered with RedHat, Equinix and Intel to build an environment where we could show this and other use cases.

In this demo we use ServiceNow as a single self-service portal for multiple services (VMs, Storage, Kubernetes …) no matter where they run (public cloud, private cloud and co-lo). This provides a consistent experience for users and offers the opportunity to compare costs across clouds for the same service. This, in conjunction with a chargeback mechanism will no doubt drive sensible consumption of IT resources and keep costs down

The demo uses RedHat Ansible Automation Platform to do the automation. Ansible follows a simpler but powerful declarative syntax that is easy for IT engineers to learn. ansible allows them to achieve consistent automation that is agile and seamless to the end user.

When it comes to Kubernetes we have deployed RedHat OpenShift Cloud Platform in all clouds. RedHat is the market leader for Kubernetes platforms with strong presence in both public and private cloud environments. Having a single platform across the board allows customers to enjoy consistent container management

The project was implemented in an Equinix datacenter in Singapore and it included the Equinix Cloud Fabric solution that provides high-bandwidth low-latency connectivity to all 3 main public cloud providers. This is no small feat because it helps us to provide persistent storage to our Kubernetes workloads from a single storage system located in the co-lo datacenter. We often referred this to as "cloud adjacent storage". The benefits of this design are immense. We are no longer exposed to expensive egress charges because the storage is sitting outside the public cloud itself. This also makes it easier to meet data sovereignty requirements.

For this demo we have used a Dell PowerFlex storage system which provides unmatched performance and scalability. Many organizations are using this design to replicate to/from on-prem datacenters to implement robust and affordable disaster recovery. Having a single storage system for all workloads, regardless of their location, means we also get consistent storage management

Finally, one of the biggest challenges in a multi-cloud world is data protection as this can become very complex. Luckily, Dell PowerProtect Data Manager allows our customers to protect their workloads regardless of where they run. In the particular case of Kubernetes environments Dell PPDM can protect all namespace resources such as pods, secrets and even persistent volumes