Cisco Blogs
Share

CloudCenter 4.9 Now Supports Kubernetes and Azure Stack


May 1, 2018 - 0 Comments

The Cisco team is focused on continuing to deliver multicloud solutions to customers, which is why we are proud to announce the release of CloudCenter 4.9. This release comes with several key updates including, but not limited to, native support for Kubernetes and Azure Stack, as well as new features provided by the recently acquired Cmpute.io team.

Let’s dive into what else is new with the CloudCenter 4.9 release.

Support for Kubernetes

A key part of Cisco’s multicloud strategy is to provide broad support for on-premises and cloud-based Kubernetes as a container orchestration platform. Kubernetes is a production-grade solution that changes the way customers build and run applications. But, it also acts as a common point of integration across a broad portfolio of Cisco software solutions such as CloudCenter, AppDynamics, and recently announced Cisco Container Platform, as well as infrastructure solutions such as Hyperflex and ACI. And broad product support for Kubernetes delivers the foundation for Cisco’s Open Hybrid Cloud Solution with Google.

At Cisco, we know that supporting container-based workloads and container orchestration in production is a big change for our customers that rely on VM-based IT tools and operations workflows. Now, CloudCenter support for Kubernetes helps IT operations address some of the key challenges when using production-grade container orchestration in production, and at scale across multiple business and IT groups.

With the release of CloudCenter 4.9, users can now use the interface and application lifecycle management features to deploy and manage containerized applications to Kubernetes clusters. The clusters can be those deployed in a private data center created by the new Cisco Container Platform, or those hosted in a Container as a Service (CaaS) environment in the cloud such as the Google Kubernetes Engine.

Deploying containerized workloads on Kubernetes is not unique. But what is unique is that CloudCenter can now deploy containerized workloads to Kubernetes and VM-based workloads to on-premises and cloud environments, both in a logically similar manner. In CloudCenter, container-based services are now elevated to the same logical level as VM-based application services. CloudCenter deploys AppDynamics agents for consistent monitoring of VM-based and container-based workloads.  (see info on new AppDynamics for Kubernetes).

Overall — CloudCenter extends the declarative orchestration that makes Kubernetes powerful and popular — to VM-based workloads as well.

Four reasons that is really important:

1- Streamline the IT Ops Container Transition

As of 4.9, you can model an Application Profile (a declarative and deployable blueprint) with container services and deploy that to a Kubernetes cluster. This provides a non-opinionated way to deploy and manage containerized workloads. That is table stakes for a container-based workload deployment and management solution.

But with CloudCenter, you can now also model an application profile that includes container-based service along with VM-based or cloud-based services. Then, you can deploy that single “hybrid” blueprint to Kubernetes and traditional VM or cloud environments. This works for both traditional enterprise as well as cloud-native application architectures. It also supports a flexible mix of on-premises and cloud target environments.

Hybrid topology – mix containers, VMs, cloud services in single deploy

CloudCenter’s unique “Either / Or” hybrid topology approach is a great way for IT organizations to immediately support developers and their new love and demand for Kubernetes in production. It also provides a bridge from the current state, where IT Ops use VM-oriented tools ops processes, to some future state where most workloads and IT Ops tools and processes are largely optimized for containers.

2 – Consume Kubernetes Without Deep Expertise

Successful adoption of Kubernetes in production must address the needs of multiple IT Ops roles. Cluster operator and application operator roles are obvious. But what about more general developers who may or may not have deep Kubernetes experience?  Call them “Kubernetes consumers.”

For those developers that aren’t proficient or interested in kubectl commands, CloudCenter now auto-generates the pod manifest file (.yaml format ) specific to each deployment at the moment a user clicks “deploy.” CloudCenter takes key parameters described in the familiar CloudCenter modeling tool, and auto-generates the YAML file for use in Kubernetes. Because nobody wants to write YAML!

The most retweeted tweet from KubeCon Austin was a quote from @kelseyhightower keynote

3 – Avoid Kubernetes Platform Lock-in

A less obvious, but perhaps more valuable aspect of  the auto-generated YAML file approach is that it exposes an important Kubernetes myth. The myth goes something like this: “a developer can take any application, package it in a Docker container, describe it with declarative YAML for all the things the developer cares about, and run it on any cloud with zero lock-in.”

That may be true in some cases. However, “all the things the developer cares about” may include environment-specific parameters like storage drivers, or network and security configuration. By having CloudCenter auto-generate the YAML file at deploy time AFTER the user picks the target Kubernetes environment (on-premises or GKE for example), the container and everything the developer cares about remains portable and cloud-agnostic to a reasonable degree. CloudCenter keeps containers portable across Kubernetes environments.

Additionally, this approach actually keeps you from getting locked-in to CloudCenter. CloudCenter works with Cisco Container Platform and cloud Kubernetes services. Other container management platforms that bring containers and Kubernetes to the enterprise wrap Kubernetes with proprietary tools and technology. They may no longer be based on opinionated services, but they still lock your containerized workloads into their platform. With CloudCenter and Cisco Container Platform, you get pure Kubernetes and no lock-in!

4 – Apply IT Governance and Controls in Production

Many IT organizations are still thinking through the basics of Kubernetes in production: how many clusters do we need? Should we apply resource quotas at the namespace level? How will we monitor and manage the network within and between clusters?

However, others are moving ahead and as they scale usage across multiple development and application operations teams and are now thinking about traditional IT governance and control concerns. Who gets to create or change a cluster? Who has access to a cluster? Who gets to execute kubeclt commands in production? Is the security team going to inspect container services before deploying in production? How will we limit usage and costs on-premises and in the cloud?

Governance wrapper spans multiple Kubernetes environments

The power of CloudCenter as a multicloud management platform also applies to Kubernetes. The CloudCenter service library can be curated to only include inspected container services. Usage controls can be applied to Kubernetes environments just like on-premises VMs or cloud-based environments. Roles and permissions can be set to determine who can deploy what, when, and where in different pre-production and production Kubernetes clusters.

It’s easy to download, install, and start using Kubernetes. But there is a lot that goes into making production-grade Kubernetes work in production. CloudCenter 4.9 does a lot of heavy lifting to apply critical governance and controls to make Kubernetes work at scale in a complex IT organization.

In addition to Kubernetes support, the release of CloudCenter 4.9 comes with a variety of important features.

Azure Stack Support

CloudCenter now supports Microsoft’s Azure Stack on-premises cloud as another standard deployment environment. It has also been validated to work with Cisco’s Azure Stack on Cisco UCS infrastructure solution. CloudCenter can now be added to automatically deploy and manage applications in Azure Stack environments with governance and control.

Cost Control Feature Additions

The Cmpute.io acquisition was completed in January and the acquired team and product are already adding features to the latest CloudCenter release. These include:

  • Cost planning with non-standard discounts: For customers that negotiate non-standard discounts with cloud vendors, a discount can be set and globally applied to all consumed services displayed in CloudCenter. Previously, if you have a custom discount rate, administrators had to calculate and manually enter the discounted price for each service. Now, when CloudCenter automatically syncs and updates prices, it applies the additional discount globally to all standard cloud list prices.
  • Multicurrency support: A user can choose to view costing information in multiple optional currencies including US Dollar, Euro, GBP, Chinese Yuan, and Japanese Yen. Cloud bills are usually transacted in USD. But now, platform admins can set the exchange rate and users can track their consumption in their local currency.
  • Cloud account usage report (AWS only): Administrators can view detailed account specific usage information directly in CloudCenter. visualizing an entire invoice – including storage and network costs – without logging into the AWS cloud portal.

Auto-scaling Enhancement

CloudCenter now supports auto-scaling based on multi-node usage metrics. In the past, scaling triggers were based on single node metrics such as CPU utilization. Now, scaling can be triggered based on multi-node metrics, such as when a certain percentage of nodes exceeds a threshold CPU utilization.

Cross Region DR

CloudCenter has seen some very large customer deployments. But the good news is, CloudCenter now supports high availability across geographically separated data centers. Customers can achieve high availability and distribute their architecture without having to co-locate CloudCenter Manager in the same data center.

Various Cloud Provider Enhancements

Our customers continue to offer ideas for new features that help them deploy and manage workloads in various cloud environments. So we made sure that CloudCenter 4.9 includes a wide range of cloud-specific enhancements for environments such as Amazon Web Services, Microsoft Azure Pack and Azure RM, IBM Bluemix, VMware vCenter, OpenStack (Pike), and Google Cloud Platform.

All in all, CloudCenter 4.9 is an exciting step forward for Cisco, as we aim to provide multicloud solutions for customers everywhere. But it doesn’t stop here, as we’ll be unveiling additional major feature and architecture enhancements as we move forward with CloudCenter releases later this year.

Special shout out and thanks to all the developers working tirelessly on CloudCenter!

See details in CloudCenter 4.9 release notes.

 



In an effort to keep conversations fresh, Cisco Blogs closes comments after 60 days. Please visit the Cisco Blogs hub page for the latest content.