Building microservices is easy, operating a microservice architecture is hard. Many companies are successfully using tools like Kubernetes for deploys, but they still face runtime challenges. This is where the service mesh comes in. It greatly simplifies the managing of containerized applications and makes it easier to monitor and secure microservice-based applications. So what are the top 3 reasons to use a supported service mesh? Here’s my take.
Since service mesh operates on a data plane, it’s possible to apply common security across the mesh which provides much greater security than multilayer environments like Kubernetes. A service mesh secures inter-service communications so you can know what a service is talking to and if that communication can be trusted.
Most failures in the microservices space occur during the interactions between services, so a view into those transactions helps teams better manage architectures to avoid failures. A service mesh provides a view into what is happening when your services interact with each other. The mesh also greatly improves tracing capabilities and provides the ability to add tracing without touching all of your applications.
A service mesh is not a new technology, rather a bundling together of several existing technologies in a package that makes managing the infrastructure layer much simpler. There are existing solutions that cover some of what a mesh does, take for example DNS. It’s a good way to do service discovery when you don’t care about the source trying to discover the service. If all you need in service discovery is to find the service and connect to it, DNS is sufficient, but it doesn’t give you fast retries or health monitoring. When you want to ask more advanced questions, you need a service mesh. You can cobble things together to address much of what a service mesh addresses, but why would you want to if you could just interact with a service mesh that provides a one-time, reusable packaging?
There are certainly many more advantages to managing microservices with a service mesh, but I think the above 3 are major selling points where organizations that are looking to scale their microservice architecture would find the greatest benefit. No doubt there will also be expanded capabilities in the future such as analytics dashboards that provide easy to consume insights from the huge amount of data in a service mesh. I’d love to hear other ideas you might have on top reasons to use service mesh, hit me up @zjory.