vRealize Automation vCloud Automation Center vRealize Suite

SDDC Meets CMBU – Makes Great Cava

Engineers are expensive, and we need to maximize their value.  

It takes hours when engineers need to stand up an instance of vRealize Automation for bug reproduction, testing, or development work. Those are hours that they can’t spend writing new features, testing features, and developing whitepapers or resolving customer issues. The list of valuable things those engineers could be doing goes on and on.

Adding to this, is the complexity of any modern IT infrastructure with multiple vCenter servers, which could be across the globe. The workload needs to be balanced for optimal workload performance.  NSX helps to accomplish by automating network provisioning, adding more options for engineers as consumers to choose from when they provision a VM.  We need a tool that positions the workload in the right place to maximize the value of virtualization by applying policies to the multitude of choices available.

The Cloud Management Business Unit (CMBU) needed automation!

Using our own products has been fundamental to automating our own development processes and improving how we work, in addition to eating our own cooking. “Project Cava” is the internal name for this project. With Project Cava, we’re able to manage the largest of the CMBU’s global labs from a single place, automate the placement and provisioning of complex workloads, and provide monitoring and showback of those resources to the CMBU.

Watch about Cava on YouTube


With vRealize Automation as the main piece of the equation, we’re able to automate the provisioning of vRealize Automation builds from our build farm all the way through loading them into vCenter, deploying the machines, performing initial configuration, and even loading test data sets. This frees up about four hours of engineer working time for every time an instance of vRealize Automation needs to be deployed. Wait time is even further reduced down to an hour for the process to complete.

To achieve this, we use custom workflows in vRealize Orchestrator to move the build from the build farm to the correct vCenter server based on our pre-built policies. This workflow also updates properties on the requests. Once the build is uploaded, vRealize Automation clones the virtual machines and uses software components to perform the initial configuration and load test data based on the scenario that has been chosen at the request time.

SDDC Meets CMBU - Makes Great Cava - Cloud Automation Engineering - VMware Core Confluence 2017-03-13 19-13-37
Interaction flow throughout pipeline

Automation Delivers Results.

Within the CMBU, Cava is managing five data centers: two in the USA, one in India, one in Bulgaria, and one in Ireland.  This lets our users choose from a single portal where their workload gets deployed geographically. The vCenter server, network, and datastore are all abstracted from that choice.  vRealize Automation allows these decisions to be made based on policy so that the workload always ends up on the right hardware. This lets developers focus on adding value in the way they should be, rather than trying to figure out which vCenter server they want to be in or which datastore and/or network they can use. Now Engineers only have to know what they need and where they need it. Workload optimization is automatically taken care of for them.

Bringing Cava to the CMBU

We are developing on our products all the time. When we have an internal builds complete for vRealize Operations, vRealize Business for Cloud, and vRealize Automation we are able to bring these latest capabilities to the platform in a rapid and repeatable fashion. We’re upgrading our Software-defined Data Center (SDDC) every month with the latest and greatest that VMware has to offer. This means that we’re developing end-to-end use cases at the same time the features are undergoing heavy development. This has resulted in feature improvements and a rich backlog of features and functions based on real world use of vRealize Suite in the CMBU.

To date, we manage 3000 VMs with a weekly churn of 200-400 VMs to serve the Engineering, Product Management, Professional Services, and Global Services teams connected with vRealize Automation. This year, we’re expanding this to the full CMBU, so the teams building vRealize Operations and vRealize Business for Cloud will be able to reduce developer friction with their infrastructure and get more value out of their engineering teams.

Customers love Cava too.

When we meet with customers with larger deployments, they’re thrilled about what they can do with vRealize Automation. Additionally customers are excited to see how VMware has managed to streamline their operations by using the vRealize stack on top of vCenter and NSX. Customers also want to know how to automate the deployment of vRealize Automation so they can adopt a DevOps philosophy around their own use of vRealize Automation for workflow and blueprint development as an alternative to having a fixed “dev” deployment of vRealize Automation that must be shared.  We’re in the middle of a beta program for the automated deployment blueprint with one of those customers and receiving great feedback. Stay tuned and we should have a packaged blueprint available for download soon!


Leave a Reply

Your email address will not be published.