New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add Envoy project proposal #43
Conversation
Also-by: Matt Klein <mklein@lyft.com> Signed-off-by: Chris Aniszczyk <caniszczyk@gmail.com>
The Envoy project proposal is ready for RFC @cncf/toc |
cc @louiscryan |
+1 would be a great addition to the CNCF IMO |
This is fabulous! I view Envoy as an essential component of a modern cloud-native stack. Having the backing of the CNCF community will be mutually beneficial to all parties. |
Clearly I'm in favor. In addition to Envoy being a great proxy
implementation with the right community ethos they've also done a bunch of
work around configuration standardization for proxies that I think will
benefits other proxies in the ecosystem over time.
…On Fri, Aug 25, 2017 at 10:13 AM, Ross Kukulinski ***@***.***> wrote:
This is fabulous! I view Envoy as an essential component of a modern
cloud-native stack. Having the backing of the CNCF community will be
mutually beneficial to all parties.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#43 (comment)>, or mute the
thread
<https://github.com/notifications/unsubscribe-auth/AIoKPJIt_828Mchl8r2SavPakWflsa7dks5sbwDVgaJpZM4O9rqR>
.
|
+1 to adding Envoy to CNCF |
+1! Finally! |
The Cloud Foundry community is looking forward to collaborating on enhancements to Envoy for our use cases. I expect a governance model as provided by CNCF should help facilitate this. |
+1 yay! Happy to see Envoy being considered for the CNCF. Not only is this technology valuable for next-generation microservices architectures, it fits nicely with existing CNCF technology. |
So this is a great writeup that motivates the value this project provides. One suggested improvement would be for the writeup to clarify the relationship between envoy and istio. See https://news.ycombinator.com/item?id=14540900 that highlights the confusion that exists on the relationship of envoy and istio: "I remember reading about the release of Envoy by Lyft maybe a month or two ago. Then I saw the release of Istio maybe a month ago. I'm having trouble reconciling the two. I read that Istio is based on Envoy. Are they mutually exclusive? Is envoy obsolete now?" |
+1! |
@bradtopol I will defer to @caniszczyk on whether to add to proposal, but this is how I would describe Envoy vs. Istio: The service mesh architecture is composed of two pieces: the data plane and the control plane. The data plane is responsible for touching every packet that flows through the network (e.g., load balancing, buffering, rate limiting, timeouts, protocol translation, security enforcement, etc.). The control plane is responsible for configuring the data plane so that it can go about its job (e.g., creating route tables, service discovery, traffic shifting, authn/authz policy, etc.). Envoy is a "universal data plane." A rich set of configuration APIs allow it to be used in a large variety of deployments that may range from bare metal to FaaS. This inherent flexibility comes with the cost of complexity; Envoy by design is meant to to be useful in a lot of scenarios. In order to do anything useful, Envoy must be coupled with a control plane. Control plane complexity can range from human built static configs, to a config generator using salt/python/jinja, to an extremely complex system using all of the Envoy management APIs. This again depends on the needs of the deployment. In summary, without some control plane, human or otherwise, Envoy is not useful. Istio is an advanced control plane for service mesh architectures that aims to create a more seamless developer and operator experience. Istio requires a data plane to function. The project has chosen Envoy as its reference proxy, however other vendors have already demonstrated the ability to plug their proxy in (e.g., Bouyant and NGINX). By focusing entirely on the control plane, Istio achieves the benefit of a clear separation of concerns, and can focus on creating an end-user experience that is more magical, at the expense of not being as flexible (Istio has initially focused on k8s, though ultimately it will support other deployment options as well). Ultimately, Envoy and Istio are related, but not the same thing. For various reasons, Istio has chosen Envoy as its reference proxy. However, Envoy will see wide deployment outside of Istio, and other data planes may see deployment with Istio without Envoy. |
Non-binding +1 as show of community support. |
Non-binding +1. I'm excited to watch (and help) the Envoy project continue its march of success for years to come! |
Non-binding +1 |
+1! (non-binding). |
Thanks much for the clarification @mattklein123. |
+1 (non-binding) |
1 similar comment
+1 (non-binding) |
The vote has been formally called on the CNCF TOC mailing list: |
+1 to adding Envoy to CNCF! |
+1 (non-binding) |
2 similar comments
+1 (non-binding) |
+1 (non-binding) |
Welcome Envoy as our 11th project in CNCF, the @cncf/toc has accepted the project :) https://lists.cncf.io/pipermail/cncf-toc/2017-September/001202.html |
https://github.com/lyft/envoy
Also-by: Matt Klein mklein@lyft.com
Signed-off-by: Chris Aniszczyk caniszczyk@gmail.com