Multi-cloud is now the dominant cloud strategy for enterprise IT, according to a RightScale 2015 survey finding that 82% of organizations are investing in multiple clouds for their application workloads. How should IT operations teams plan to support a multi-cloud environment, when each cloud provider comes with its own API, deployment methods, and toolchain?
Operations teams generally manage different processes and tools for every cloud. For example, they may use CodePipeline for deployments on Amazon Web Services, and Puppet for deploying on OpenStack in the datacenter. This cloud segregation can lead to vendor lock-in, complex migration projects, and duplicate processes for common tasks like continuous integration, credentials management, service provisioning, scaling, logging, and routing.
Let’s see how the operations story changes when cloud providers are standardized with a single overarching platform and API. Pivotal product manager Matt Reider shows how using the shared operations toolchain offered by Pivotal Cloud Foundry makes it possible to deploy and scale one application between four clouds with no code changes needed. This 10 minute video shows how quickly any cloud-native application can be migrated across VMware vSphere, vCloud Air, OpenStack, and Amazon Web Services:
This demonstration uses a simple, publicly available Python app called 9fellas, which connects to a platform-managed Redis datastore to visualize the number of instances running on each cloud. Since each public and private cloud is standardized under the same API and backing services, the same level of portability applies to any 12 Factor application running in a Cloud Foundry environment.
For companies preparing to manage multiple cloud infrastructures, consolidating cloud providers under a single platform offers more than easy migration of production applications on demand. Investing in a shared operations stack also helps to address application sprawl, reduce an organization’s attack surface, and enable developers to focus on building user-facing functionality rather than back-end infrastructure.
This level of ecosystem portability is complemented by the recent Cloud Foundry Foundation decision to certify interoperability between different distributions of the platform, providing IT teams with confirmation that their applications are compatible with any official Cloud Foundry implementation. IT operations teams now have an open, standards-based path to supporting multiple clouds today, with the flexibility to take advantage of other providers and clouds in future.
In a recent tweet, core contributor to Spring Framework Phil Webb positioned the cloud as the basic infrastructure for powering your applications – just like electricity. But if each cloud is a different electrical grid, then Cloud Foundry is the travel adapter.
Attribution: The RightScale 2015 State of the Cloud Report is licensed under the Creative Commons Attribution 4.0 International License.
About the Author