Across every enterprise, there is unprecedented investment in modern applications and the cloud and cloud native environments that support them. A recent KPMG 2020 US Survey found that 74% of CEOs are accelerating digitization of operations and creation of a next-generation operating model.
On a recent Turbonomic webinar, Linley Ali, Enterprise Architect and IT Strategist at Florida Blue, provided insights on application modernization at Florida Blue. Linley has been at Florida Blue for over eight years and was previously with Royal Bank of Canada (RBC) for 20 years.
Florida Blue is a member company of the Blue Cross Blue Shield Association focused on health insurance in Florida. With over 12,000 employees and $19B in revenue, Florida Blue is in the middle of an application modernization and cloud native journey. The following blog provides excerpts from our recent webinar, “DevOps Workflow: Better Collaboration Between Application Development and IT Operations”.
Can you tell us about your applications and your motivations to go cloud native and leverage microservices?
An organization like Florida Blue has a wide variety of platforms and technologies. Florida Blue is a large health insurance company with about 5 million plus members, and we do not have a “one size fits all” mentality. We are always looking to exploit technology to the greater needs and greater means. We started looking at cloud native and microservices architectures about four years ago.
How did you choose your platform for your cloud native initiative?
We were looking at different platforms. We considered Pivotal Cloud Foundry, Bluemix from IBM, and Red Hat OpenShift container platform. We also considered Kubernetes and Kubernetes very young at the time. We decided on OpenShift container platform, because it fit our profile, it fit our cost model, and it fit our investment framework. OpenShift allowed us to start small and expand and grow over time.
What is your motivation for moving to cloud native?
The original notion was that if I have my applications in cloud native environments, I can migrate workloads between on-premises and cloud environments.
What is the reason for Florida Blue running workloads on-premises?
We run an on-premises private cloud model with the same sort of capabilities as public cloud. The reason we went this route is because we have Health Insurance Portability and Accountability (HIPAA), Protected Health Information (PHI), HITRUST and SOC 2 compliance requirements.
We are running a cloud-like environment on-premises. At the end of the day, applications need network, storage and compute and they should not necessarily need to know, nor should they care about where the resources are being provided from.
How are you exploiting cloud native for your advantage?
When we started this journey four years ago, we originally thought of building out a Platform-as-a-Service (PaaS) and lowering costs. We do not think of it as a PaaS, cloud native is about the effective use of resources and it enables us to be more effective and meet the business needs.
How did you decide which applications need to go to cloud native?
There is no “one size fits all” model. Optimally, there are applications already on the right platform and you do not want to move these to cloud native, despite what some of the cloud providers would like to tell you.
What does multitenant mean for Florida Blue?
Each application team has built out a set of services and they share a cluster. Each application group has a cost model associated with the resources consumed and that allows us to charge back appropriately.
What are your operational challenges working with your DevOps and application teams?
It is not easy, it is not as simple as saying “I have an OpenShift container platform and have a platform of choice, just start deploying things”. You run out of resources or you overprovision resources. Especially for on-premises environments, we need to optimally pack the resources so that we can use them correctly. It takes 12 to 16 weeks to provision additional physical infrastructure, so you want to make sure you are using the resources correctly. Technologies like Turbonomic can help us plan capacity, give us insights into how resources are being used and if we are optimizing the resource usage, and give us an idea of what percentage of resources are available (25% to 30%) that gives us the runway time.
What is the role of DevOps in Containerized environments?
The original concept of DevOps is that I am going to teach my development people to learn operations and my operations people will understand development. The application teams did not really understand the operations world. They did not understand when in production, what impact the application has on the existing environment. Vice, versa, the operations team did not understand what the application was doing or the facets of the application context. In the current situation, we have a small DevOps team that helps the application groups onboard and understand what their applications are going to do. Plus, we help operational teams understand the context of the application, when an application spins up or alerts occur, we help operations track, trace, and monitor to effectively utilize those resources. DevOps is a critical role in making all this happen in this modern world.
How are you bridging the gap between application owners and IT operations?
Originally, we had small, medium, and large environments for deploying applications. But the application owners did not know, and they would always ask for a large environment, maximum CPU, and memory resource; everything was the maximum size, it was a new world for them, and they were cautious.
We are using Turbonomic to understand and contextualize the application, i.e., we looked at the application profile and it is better suited for a medium size environment. With Turbonomic, we built the trust with the development teams and they allow us to determine what they need. If the developers have written a truly cloud native application, we can start the application on a small or medium size environment and as the demand increases, we can scale up the resources as needed.
For the operations team, it is important that they have a full stack view of the environment, that is a requirement to maximize the operational efficiency. We use these insights to work with the application teams to help them understand how their applications are using resources.