The need for bi-plane architecture in API management

webMethods API Control Plane was recently released to provide a single pane of glass for distributed runtimes and APIs to manage and monitor them. It introduces the concept of a control plane that is separated from the data plane. This bi-plane architecture is not a new concept in the networking context, but it is new for API Management.

Let’s define our planes

Data plane: This is where your APIs day-to-day operations take place. It can include API gateways, other proxies, or developer portals; what we call generically call runtimes. You can have multiple runtime clusters in the data plane that are distributed across geographies and data centers in private or public clouds. These runtimes serve the API traffic routing calls, transforming requests and responses, shaping the traffic, or publishing your APIs with your consumers. It is best that they stay close to where the action is, i.e. where the backend applications are hosted or where the API calls are coming from, in order to minimize the latency. This ensures that there is minimal impact of network issues on the daily operations of the business. 

Control plane: Control plane sits on top of the data plane to separate the control and monitoring aspects from the run time operations. It abstracts configurations to a central location which manages multiple runtimes at once. It also shows aggregate metrics for different API personas to help them understand the overall performance and KPIs of their investment, making sure that everything is optimally operational.

Who cares?

Different API personas involved in the API lifecycle have different concerns and live in their areas of interest. The API developer is concerned about their own APIs, delivering them in a timely manner and making sure that those APIs work and perform as intended, therefore their day-to-day concerns are limited to the data planes their APIs are located. Similarly, the API consumer cares about finding and subscribing the right API for their application and live in the data plane that lists the APIs that are available to them. On the other hand, the platform owner’s area of responsibility is the entire API landscape, however distributed that is. Therefore, they live in the control plane, monitoring the performance and analytics of the entire platform, making the necessary moves to ensure the efficiency of it. They care a lot about being informed quickly about any issues on any runtime and being able to take action on these issues. Similarly, the API Product Manager, key in utilizing the API Control Plane for Product Managers, may be responsible for APIs that span multiple runtimes. They want to make sure that all their APIs, wherever they may be proxied or published, have the right reach, the right configurations, and the right packaging, also they want to monitor the business KPIs to ensure the right return on API investments. 

Having a dedicated control plane spanning all data planes benefit the API platform owner and API product manager tremendously. The control plane provides a single tool to stay on top of their areas of responsibility, and relieves them of the burden of having to traverse multiple runtimes, trying to make sure the configurations of these multiple runtimes and many APIs hosted on them are consistent and are as intended, combine and analyze metrics from all of them, and just be aware of the entire landscape. Now, let’s take a deeper dive into the functional benefits that this separation brings.

What are the benefits of separating these planes?

Having the control plane separate from the data plane enables a single pane of glass to manage multiple runtimes that may or may not be aware of each other. You can govern individual runtimes or APIs on these runtimes and compare the configurations of multiple runtimes or APIs from this central place to make sure that your API investments are running optimally. It allows you to discover issues without having to traverse multiple runtimes and make corrections quickly. 

The control plane gives central platform and product teams views into the operations of multiple business units while giving these BUs independence to select and use the technology that makes the most sense for them. Control Plane can talk to all these different runtimes in the data plane from different vendors to provide a unified view to the platform owner. 

The ability for the control plane to configure and gather metrics from these multiple technologies allow federated API governance. This means that BUs can enforce their own rules and policies while also conforming to organizational best practices. All these runtimes can be visualized from the control plane to ensure that the API landscape is in compliance. 

Speaking of API landscapes, one of the biggest advantages of the control plane is to provide a single view of the entire organizational landscape. In the data plane, each runtime’s visibility is limited to the APIs on it. However, the control plane catalogues all runtimes, APIs, their versions and applied policies, letting you get ahead of API sprawl. This API catalogue enables reuse by listing all APIs on all runtimes, prevents creating duplicate APIs, and lets you terminate zombie APIs. 

And one of the greatest advantages of the control plane is that it shows individual or aggregate metrics from the data plane. Through its monitoring and analytics dashboards, you can control operations and discover problematic runtimes and APIs in a single tool. Control plane also allows grouping of runtimes, for example in certain geographies or cloud providers and shows you insights to help with your investment decisions.  

Control plane presents you with these benefits while the data plane is hard at work serving API calls and enforcing runtime policies. The communication between the control plane and the data plane is asynchronous so that it doesn’t negatively impact the processing of APIs calls. 

webMethods API Control Plane to rule them all

We released webMethods API Control Plane to provide you with this single pane of glass to give you all these benefits. It sits on top of your data plane to visualize your API landscape all together, or individual runtimes or APIs. It gives you piece of mind to make sure that everything is running optimally and saves you time from traversing multiple runtimes to configure and observe them. All of this happens while your data plane processes API calls and enables your business operations.