It’s been almost a year since Mesosphere’s Jörg Schad told us that the next big topics in the container world are service meshes and machine learning. Suffice it to say, he was right.
This prediction certainly came true so let’s try it again and see if we can predict what’s next for service meshes. Three months ago, we talked to Brian Harrington, Product Manager for Istio at Red Hat about Istio’s latest features but we also discussed the future of service meshes. Here’s what he said:
I believe the future of the service mesh is tied part and parcel to serverless computing. We are converging on a state where developers are successfully breaking down code bases to atomic components.
This trend is even reflected in the question around the modularity of Istio. I feel that the future of the service mesh is one which operates in symbiosis with technologies like Knative and Apache Whisk, enabling developers to re-adopt the “UNIX ideology” of “do one thing and do it well” for the purpose of building the future of applications.
Read the entire interview here.
The problem is that when you first enter the service mesh world, it can get pretty overwhelming so in order to make things easier for you, Solo.io has created a service mesh orchestration platform called SuperGloo. This open-source project aims to manage and orchestrate service meshes at scale. In short, it aims to “simplify the installation, management, and operation of your service mesh, whether you use (or plan to use) a single mesh or multiple mesh technologies, on-site, in the cloud, or on any topology that best fits you.”
Check out their GitHub repo, docs, and join their Slack channel.
The rise of service mesh architecture
One cannot ignore the growing popularity of service meshes, especially when Istio is on everyone’s lips these days, including ThoughtWorks’. There are at least three leading service meshes though: Istio, Linkerd and HashiCorp Consul’s Connect feature. Right now, SuperGloo supports all three.
Idit Levine, founder and CEO at Solo.io explained in a recent blog post that SuperGloo is “designed to address current and future challenges, from early stages of usage to complex super-meshes involving many instances on multiple clouds.”
Here are some of SuperGloo’s features:
Dead Simple API – SuperGloo boils down service mesh config to the basics, while being expressive about the features that are most important to users.
Installation – Automate complicated workflows such as supporting multi-tenant (multi-namespace) installations using the one-click SuperGloo API.
Routing Traffic control – Apply sophisticated HTTP/2 features to any/all traffic in your mesh, such as traffic shifting, fault injection, header manipulation, and more.
Routing Resilience – Outlier detection, connection pooling, timeouts and retries.
Observability Metrics – SuperGloo automatically configures existing Prometheus, Grafana, and Jaeger installations to scrape and visualize data from your mesh.
Plugability – Highly extensible open-source Go architecture with development SDKs making it easy to extend SuperGloo with new features and meshes.
SEE ALSO: If you want to check in on the microservices in your Istio service mesh, Kiali’s got you covered
Next stop: Multi-mesh?
What does the future hold for service meshes? If you ask Solo, they’ll say that the future is definitely multi-mesh, which is why they prepared SuperGloo for such a scenario. So in order to make sure you’re not on the wrong side of history, you should consider exploring this option.
This service mesh orchestration platform promises to make it easy to explore different meshes and migrate between them, as well as to allow using the same tools with different meshes. What’s more, SuperGloo will ‘glue’ meshes of any type into a multi-mesh.
Getting started with SuperGloo
- After cloning, run
make init
to set up pre-commit githook to enforce Go formatting and imports - If using IntelliJ/IDEA/GoLand, mark directory
api/external
as Resource Root
The post Service mesh orchestration platform SuperGloo makes working with service meshes a breeze appeared first on JAXenter.
Source : JAXenter