solo.io KubeCon + ServiceMeshCon EU 2020 - Virtual, 8 Sep 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Welcome & Opening Remarks - Christian Posta, solo.io

Description

Don’t miss out! Join us at our upcoming events: EnvoyCon Virtual on October 15 and KubeCon + CloudNativeCon North America 2020 Virtual from November 17-20. Learn more at https://kubecon.io. The conferences feature presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects.

Welcome & Opening Remarks - Christian Posta, solo.io

https://sched.co/ZWiY

A

Well, hello and good morning or good afternoon, whatever time you may have welcome to the second service mesh con presentation track, this one is a little bit different.

A

Unfortunately, we can't all be in person with you, but we do hope that you are safe and healthy and that you found somewhere comfortable to watch this track about the different service, mesh technologies and the presentations and speakers that we'll have for you.

A

So my name is christian posta.

A

I am a field cto at a company called solo.I o where we build service mesh technology and I've been built, help helping to build architectures and application architectures based on microservices, and even before that things like service oriented architecture at large enterprises for the bulk of my career and over the last three three and a half years, I've been interested in technology that has now emerged into what we're calling service mesh I've written some books on this topic, including uh the first book on a technology called istio a couple years ago, and I'm currently writing the istio in action book for manning.

A

So please reach out. If you have any questions, I will be giving a talk in this track as well, and I just I just want to point you to the initial service mesh con, which was a large success. Great turnout, great uh speakers and and sessions and the uh so the whole thing went really well. It was recorded.

A

So if I can point you to the program from last year and a link to the youtube playlist where you can watch the um the various talks from from the speakers previously now that, as I said, that, that event was successful and the intention was to run that at the kubecon eu in amsterdam in march this year. But as you can see, we are in a remote and video format, and you know we. We still have a very, very compelling lineup of speakers and presentations.

A

So before we get going.

A

Although we've been discussing service mesh for the last few years, some in the capacity of it being kind of hyped some more recently, I would say in the capacity of adoption, in my experience, working in the field working with customers working hands-on to solve real challenges around deploying and managing applications in a cloud platform or cloud environment potentially built around things like microservices and so forth. To help the organizations move faster.

A

You know, we've we've come up with solutions. In the past we've tried out, I would say, service mesh when it was a little bit more in its infancy, but now as it's starting to mature across a bunch of different projects, we're seeing legitimate adoption people going to production with it some successful stories.

A

Some you know, they've hit some bumps along the way, which is good because that induces learning, um but there are still people coming to the these types of events that are just learning about service mesh. So if there's an entire track devoted to this, let's at least put a definition down. What is it and my definition is fairly straightforward when you deploy applications, they need to talk with each other, all right when you deploy applications and they talk to each other. They talk over the network and this network is not reliable.

A

That's the foundational piece all right, so this is not new. This position isn't new, but when you deploy into a cloud environment where things are highly dynamic service, a could be scaling up could be failing. Service b could be on the other end of a network partition. These cloud based networks. Aren't not all that reliable you don't own them. Typically, there might be layers of virtualization and software-defined networking and so forth is important and the services cannot ignore it to solve those challenges around the network and those challenges, typically being connectivity.

A

How can you do things like timeouts, retries circuit, breaking deadlines, retry budgets, that kind of stuff around security? How can we secure the transport and the application level uh messages that are flowing over the network and things like observability?

A

How do we understand what's happening on the network so that, when things do go wrong, we can, we can have a way to uh very quickly, diagnose and understand what is what is happening.

A

Otherwise, as applications talk over the network they're, you know black boxes, they don't they don't consistently expose the metrics that you might need, and so the one way to solve this, not the only way, but one way to solve this- is to put a little agent or interceptor co-located with your application and that when the application talks over the network, this little helper agent or sidecar service proxy takes on the role of establishing consistent collection of telemetry, potentially originating things like a tls connection or establishing helping to establish a mutual tls connection.

A

Without the application knowing or having to know the application thinks it's talking to the network, and so when you start to put these little service proxies next to each of the application workloads.

A

Now you have a application architecture that is more resilient to network problems.

A

That is more secure because it can do things like encryption and authorizations and so forth between the services can establish identity of the applications without relying on you know previous forms of identity and trust around firewalling and ips, and all that stuff we can assign identity to the applications and have that as first class and uh and also have traffic control and resilience, and some of this operational, um you know when you build these applications. You need to solve these problems.

A

Service mesh encompasses these these patterns. This is a one way of solving this type of problem. You typically see these proxies deployed with the application and some way of managing it. So that's where the control plane comes into play.

A

Service. Mesh adoption, as I've been uh trying to say, has been um first of all real. uh We work with customers at solo. I know others uh are seeing. You know the buoyant folks are seeing a a good upswing in adoption, and you know the istio community and and so forth. Some of the capabilities driving this adoption can be broken down into, uh like I said, security, observability and control over the traffic routing in and between the services for doing things like canarying and so forth.

A

Now these topics we're going to see presenters through the rest of the program, talk about and and touch on these on these topics, um but the meshes the particular meshes that we'll see and hear about in the lineup will vary. The ecosystem is very vibrant right now, new releases of functionality.

A

um All of these all of these projects that you see here, for example, linker d uh back, I think in june and uh early july, they they released a version of linker d that now supports multiple clusters, so we're starting to see not just adoption but real life, multi-cluster, non-trivial deployments and workloads and so forth. We saw kuma that was introduced into cncf, which is a service mesh based on the envoy proxy, and you see some of these others based on envoy proxy.

A

You have the smi which joined the cncf to drive a spec for an api around surface mesh, and you know, as we go even you know, at solo, we're working on technology to help run a service mesh across multiple clusters and so forth. So I check out the different meshes see some of the talks that we're going to part of this track that will cover these different technologies and, more importantly, the use cases that and the problems that uh these different companies like allegro and equinix and sherbank autodesk.

A

um You know that they had in in their real production environments that use the mesh to to solve that. So I'm super excited to see these talks. uh You can see the schedule here. um I have a talk later in the program on multi-cluster deployments and management and so forth, uh check out these talks.

A

Representatives from, like, I said, uh uh kuma uh from um linker d, from maish from istio and uh and then we we get to the toward the end of the program where we have uh my good friend, nick jackson, from hashicorp representative of the consul service mesh there um he'll be he'll, be closing things down at the end.

A

I want to say a big shout out and big thanks to the cncf for helping to persevere in these in these times and then keep this going and sharing you know being able to put together the the program committee- big, thank you to the program committee for organizing this and and um getting to a point where we can continue to have these uh these different collocated events with kubecon and especially for service mesh con.

A

Thank you very much and with that, let's get to the program again, thank you for joining us remotely and we hope you enjoy the rest of the talks. Thanks.