Application programming interface management company Kong Inc. said today its open-source service mesh Kuma has been accepted as a sandbox-level project by the Cloud Native Computing Foundation.

Kong also announced an update to Kuma for monolithic and microservices-based applications, making it easier to deploy multitenant instances of the service mesh across different networking environments.

Service meshes create an abstraction layer across networking environments that helps to make it easier to deploy distributed applications within them. They work by automatically routing application traffic across multiple network underlays, which does away with the need to configure each application service for a specific network.

Kong’s Kuma service mesh is based on an open-source proxy server called Envoy that’s also hosted by the CNCF. In addition, the CNCF hosts another service mesh called Linkerd, but Kong said Kuma is lighter-weight and therefore easier to deploy in distributed computing environments.

Perhaps the best-known service mesh is Istio, which was created by Google LLC, IBM Corp. and Lyft Inc. specifically for use with Kubernetes, which is used to manage clusters of containerized application. Istio is open-source too, but it isn’t hosted by any organization and is notoriously complex to deploy and manage.

Kong said the popularity of Envoy means it makes sense for many organizations to work with a service mesh that’s based on it. The company argues that Kuma is not only more lightweight, but also scales up better and is more extensible. For example, the latest release of Kuma has added support for global and remote plane replication across hundreds of thousands of data plane proxies.

Also new is an ingress data plane mode that helps to automate cross-cluster and cross-platform service mesh communications, and a native universal Domain Name System service discovery API that makes it appear as if multiple application services are using the same shared cluster.

To date, very few companies have actually deployed any kind of service mesh inside a production environment. But with the number of microservices-based applications in deployment steadily increasing, Kong said, organizations will need to find an easier way for those apps to invoke network resources. The challenge now is to get service mesh platforms to the point where enterprise information technology teams are comfortable with both deploying and managing them.

That’s what today’s move is all about. Now that Kuma has been adopted by the CNCF, Kong said, it hopes more organizations will be enticed to contribute to its development and eventually build their applications on it.

“From a technology standpoint, it makes no sense for individual companies to create their own control plane but rather build their own unique applications on proven technologies like Envoy and Kuma,” said Marco Palladino, co-founder and chief technology officer of Kong. “We welcome the broader community to join Kuma on Slack and on our bi-weekly community calls to contribute to the project and continue the incredible momentum we have achieved so far.”

Constellation Research Inc. analyst Holger Mueller told SiliconANGLE that service meshes are likely to be a key element of next-generation applications, since they enable those apps to use a large number of microservices more easily.

“The competition to control that new software application layer from a services perspective is incredibly fierce,” Mueller said. “But that competition is what executives want, as it gives them more options to choose the service mesh that’s most suitable for powering their specific applications.”

Image: Kong