Pivotal: Using Google Cloud to accelerate R&D for its cloud-native platform

Pivotal Software set itself a grand goal: to transform the way the world builds software. The company aims to achieve this by giving its customers tools to more easily run cloud-native apps. To this end, Pivotal offers Pivotal Cloud Foundry, an integrated platform that simplifies how big companies build, deploy, and operate applications. The result — improved operational efficiency, higher developer productivity, and reduced operating costs. Most importantly, Pivotal Cloud Foundry gives companies a platform to deliver better products faster. Many of the biggest companies in the world use Pivotal Cloud Foundry in concert with multiple cloud platforms for storage and other services.

From the outset, Pivotal built Cloud Foundry to run across multiple cloud providers. And of course, it has long used public cloud services as the foundation for its own software development. After several years, Pivotal saw its cloud expenses soaring and needed a lower-cost and more flexible solution. The company switched to Google Cloud Platform because it was less expensive, offered better performance, and scaled more quickly.

“We tell our customers that they can port their development platform on Pivotal Cloud Foundry to a low-cost, high-performing cloud service. It’s important for us to follow the same advice we give our customers, and so for our development work we switched to the best provider for us: Google Cloud Platform.” — Mike Dalessio, Vice President, Engineering, Pivotal

Testing Pivotal Cloud Foundry at massive scale

The company’s internal Pivotal Cloud Foundry environments run on Google Compute Engine. So does the firm’s development and testing infrastructure, and its ongoing research and development. With Google Compute Engine, Pivotal can quickly scale to meet the elastic demands of its R&D iterations because of the speed with which virtual machines (VMs) can be provisioned. Pivotal also saves on costs, because as usage increases, the cost per VM goes down, thanks to Google Cloud Platform’s sustained use discount pricing model.

Pivotal rolls out new versions of Pivotal Cloud Foundry by using a well-designed continuous integration pipeline that runs quality checks. This ensures that only rigorously-tested software makes its way to customers. Concourse, an open-source CI framework running on Compute Engine, drives this process. The company also uses Compute Engine to stress-test Pivotal Cloud Foundry, at one point even using GCP to deploy and maintain 250,000 production application containers to ensure Pivotal Cloud Foundry could scale to meet any load.

“Google Cloud Platform was the ideal choice for our intensive benchmark testing of 250,000 application containers. It scaled up without any pre-planning required on the part of Google Cloud Platform because of its extremely fast VM provisioning times. Because of Google Cloud Platform’s high performance, it stood up to the massive load without any problems,” says Dalessio.

Big benefits for Pivotal’s engineers — and its customers

Pivotal engineers say they are more productive with GCP. Cloud infrastructure costs have been reduced by 40 percent since Pivotal moved to GCP, Dalessio says. Speed has increased as well. With the previous cloud provider, it took several minutes to create a new VM. With GCP, it takes seconds.

GCP’s benefits go beyond helping Pivotal engineers create, maintain and test its platform; they also accrue to Pivotal customers. Google engineers worked with Pivotal to bring GCP’s key data and analytics services to Pivotal customers by writing a Cloud Foundry service broker for Google Cloud Platform. With this service broker, customers can use a full range of GCP services on Pivotal Cloud Foundry, including Google BigQuery, Google Cloud Storage, Google Cloud SQL, Google Cloud Pub/Sub, Google Spanner, Google Cloud Vision API, Google Cloud Speech API, Google Cloud Natural Language and Google Cloud Translate API. All those services can be added to new apps easily.

The relationship has been so successful that Pivotal was recently named the Google Cloud Technology Partner of The Year for 2016.

“With Google Cloud Platform, we’re giving our customers easy access to the full suite of sophisticated GCP services such as machine learning, something that would otherwise be difficult for them to do. It lets them quickly use powerful data service seamlessly on the platform,” says Dalessio.