It's been well over a decade since the broad adoption of virtualization technologies transformed the data center. In a relatively short period of time, we have transitioned the management of under-utilized server resources to running multiple machine instances that manage resources down to a single core.
However, the evolution of the data center didn't stop with virtualization.
Companies like Google needed even more resource control to develop and deploy applications faster and cheaper so Google developed cgroups, a Linux kernel feature that limits and isolates resource utilization. Google has used this technology for years to run all of their applications. Docker extended this functionality by providing open source, lightweight containers that run processes (file systems, network interfaces, etc.) in isolation.
The biggest and most immediate benefit of containers for the enterprise is faster (and better) application development and deployment. By leveraging lightweight and isolated containers for applications and services, developers write and share code from their desktop to their development stack and throughout the entire development workflow.
Management of containers
While most enterprises are not developing and deploying applications at the scale of Google, enterprise developers still need the ability to manage hundreds (if not thousands) of containers that need to be highly available, easily patched, updated and monitored. Google handled this challenge years ago by developing a container cluster management system called Borg. This technology was ultimately spun out publicly as an open source project called Kubernetes. With Kubernetes, enterprises now have the opportunity to develop and deploy applications and services faster (and better) by leveraging the knowledge of Google and the open source community. For enterprise production, this method easily enables the continuous delivery of applications and services.
Continuous delivery is the critical element for enterprise applications.
The question for enterprises is how to best leverage container development for real business value. Specifically for CIO's and CTO's, the challenge is really about enabling developers with the resources and tools to develop resilient applications and services that provide continuous delivery.
Hitachi has addressed this by focusing on the application development platform that extends from test/dev through production whether in a traditional data center or deployed within a cloud environment. Through close collaboration with Google, HDS is the first enterprise hardware vendor to provide an infrastructure platform for Kubernetes. The Hitachi Unified Compute Platform (UCP) is a converged platform that integrates compute, storage and network resources for the deployment of enterprise class bare metal, virtual machine, and now, Kuberenetes workloads.
With Kubernetes, Hitachi UCP can be immediately leveraged for the accelerated development and deployment of containerized applications and services throughout the development lifecycle. Enterprises can also maximize the value of their environment by running elastically scalable container clusters alongside existing workloads without disruption.
There are three critical areas where Hitachi UCP drives significant value:
First, Hitachi UCP automates and simplifies the deployment of infrastructure resources to support application development from test/dev to production. This includes existing bare metal and virtualized workloads as well as Kubernetes TODAY without any additional costs, licenses or upgrades. Hitachi UCP customers are already deploying Kubernetes along side their traditional workloads without having to invest in additional systems.
Secondly, Hitachi UCP extends enterprise application services via UCP Director and Kubernetes APIs. Hitachi UCP has 100% feature parity between UI's, CLI's and API's providing developers with powerful abilities that are native to UCP.
Finally, and for some most importantly, Hitachi UCP provides automated site-to-site replication, failover and fail back for disaster recovery. With complete site replication, it is possible to have a Kubernetes ecosystem fully redundant with UCP’s DR feature. UCP Director DR supports both synchronous (TrueCopy) and asynchronous (Universal Replication) replication modes, as well as adjustable MOTs to define recovery points.
Ultimately, this is about the value IT brings to the enterprise. The most consistent challenge for every enterprise is the ability to control (and manage) costs and to provide an agile development environment that maximizes the business value that IT can dependably deliver. As a pre-validated, integrated and cloud ready platform, Hitachi UCP is uniquely suited to efficiently manage applications on bare metal, virtual machine and container workloads.
If you are a Hitachi UCP customer today and looking at how you might leverage containers for application development and deployment, I’d encourage you to learn more here.