Cisco continued its HCI momentum this week, announcing at Cisco Live its new Cisco Container Platform (CCP), which runs on its HyperFlex hyperconverged infrastructure (HCI) solution. This news follows last week’s HyperFlex 3.0 release and acquisition of Skyport.
One of the key themes of Cisco Live was the need for businesses to be more agile and move faster. Businesses want to modernize the infrastructure to meet the changing demands on IT, and they are adopting the DevOps model. Along with that, containers have grown in popularity, as the lightweight nature of them makes them ideal for organizations that have embraced the concepts of DevOps.
However, there is a dark side to containers. Allowing developers to fire up containers willy-nilly on whatever cloud they want can create an unmanageable environment. Also, workloads in containers can behave differently in a private cloud versus public. IT operations need to get a handle on containers before they run amok and slow down development instead speeding it up.
CCP is Cisco's solution to that management issue. It's a turnkey, open container platform that brings order to containers and enables applications to run the same way on premises as they do in public clouds. Cisco has been building solutions to enable its customers to move to hybrid, multi-cloud environments but do it in a way that’s scalable and seamless to the application developer. The solution masks the complexity that developers and IT pros face in configuring, deploying and managing container clusters based on 100 percent upstream Kubernetes.
While there are several container environments, Kubernetes is a sensible starting point for Cisco because Kubernetes has gained significant momentum over the past couple of years. Also, Cisco and Google have partnered on many solutions, so this extends the relationship between the two. Late last year, the two companies announced a joint hybrid, open cloud offering, and CCP is a key component of that.
From an architecture perspective, CCP is designed to be turnkey from the network to the Kubernetes cluster and everything in between. However, customers will have choices at many points in the stack as illustrated in the graphic below.
CiscoWhat's included in Cisco Container Platform
Key components of CCP include:
- CCP Control Plane, which runs on top of a Kubernetes Cluster
- Control plane services include:
- User interface
- Rest API endpoint
- APIs to execute workflows
- A systems manager that owns the state of everything in the Kubernetes Control Plane, such as the data store and scheme. It also handles authentication and authorization logic.
- Abstraction for an IaaS back end (Openstack and vSphere) and runtime/orchestration abstraction (k8s, Docker)
- Storage is HyperFlex
- Compute hardware is UCS or non-UCS
- Network can be an ACI-based SDN or a traditional network
CCP isn’t designed to be a panacea for all container problems, but it can greatly simplify things in the right situation.
Should you use Cisco Container Platform?
Businesses should consider CCP if they:
- Are working with native Kubernetes
- Don't require a full-functioned, integrated PaaS. Think of CCP as PaaS light.
- Are using containerized applications in production and want to benefit from Cisco-enabled extensions and services
- Want a turnkey solution
- Are considering the Cisco – Google open hybrid cloud
IT environments are becoming increasingly dynamic and distributed, enabling companies to work faster and accomplish more. However, things have also gotten increasingly complex, and IT teams are drowning in complexity as they try to cobble together multivendor solutions, hence the growing interest in HCI. Cisco has brought the benefits of hyperconverged to containers, enabling them to be deployed faster and simplifying management of them.
The initial release of CCP is slated for April 2018 and is optimized for HyperFlex 3.0. Current HyperFlex customers can license CCP separately; non-customers can purchase it as an integrated bundle. A second release is scheduled for summer, at which time CCP will be available as standalone software and will include support for VMs, bare metal and public clouds. Pricing will be subscription-based and depends on the number of cluster nodes deployed, with volume-based discounts.