Create ROADMAP.md

First version.
This commit is contained in:
Wenbo Zhu 2018-02-28 18:29:35 -08:00 committed by GitHub
parent 316618c5e3
commit 457c14150c
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 93 additions and 0 deletions

93
ROADMAP.md Normal file
View File

@ -0,0 +1,93 @@
# Background
gRPC-Web has been developed internally at Google
as part of the future front-end stacks for Google's Web applications and cloud services.
Over time we plan to open-source and publish
most of the features and make them available to gRPC users.
Like everywhere, Web platforms and technologies are constantly evolving, often with many inter-dependent ecosystems.
As much as we like to open-source everything, we also need keep the balance bewteen creating a reusable and stable open-source solution
and meeting those requirements uqniue to Google's Web ecosystems or their applications (such as search).
The purpose of this document is to list all the features (as wish-lists) that we believe are useful for gRPC users.
We hope to use your [votes]() to prioritize the work to either publish some existing solutions or develope some of the features
directly in the open-source repo. For the latter case, please mention if you are interested in contributing to any of the road-map
features.
# More efficient message encoding
The binary protobuf encoding format is not most CPU efficient for browser clients.
Furthremore, the generated code size increases as the total protobuf definition increases.
For Google's Web applications (e.g. gmail), we use a JSON like format
which is much more efficient (comparable to JSON) and compact in code size.
Note this is also one of the reasons we need support the non-binary ("application/grpc-web-text") wire-transport format.
# Streaming-friendly transport implementation
Currently the gRPC-Web client library uses XHR to ensure cross-browser support and to support platforms such as Reactive-Native.
We do plan to add fetch/streams support at some point, which is more efficient for binary streams and incurs less memory overhead
on the client-side.
However, fetch still has certain gaps compared to XHR, most notably the lack of cancellation support.
Progressing events, I/O event throttling are other concerns.
# Bidi support
We are yet to finalize the bidi transport story. At the same time, please tell us your exact use case,
and maybe explain why server-streaming is insufficient.
# Security
We plan to publish a comprehensive guideline doc on how to create secure Web applications.
Native support such as XSRF, XSS prevention may also be added to the gRPC-Web protocol.
# Compression
Do you need request compression? Brotli?
# CORS
We plan to support CORS preflight as specified in [PROTOCOL-WEB.md](https://github.com/grpc/grpc-web/blob/master/PROTOCOL-WEB.md).
# Local proxies
In-process proxies will eliminate the need to deploy an extra proxy such as Nginx.
We have plans to add proxy support in Python, Java, Node, C++ etc. Let us know if you are interested in implementatin any language-specific
in-process gRPC-Web proxy.
To minimize maintenance overhead, we don't have any plan to add gRPC-Web support to any new HTTP reverse proxies other than Nginx and Envoy.
# Web framework integration
This is to provide first-class support for gRPC API and gRPC-Web in popular Web frameworks such as Angular.
Note Dart gRPC will be using gRPC-Web as the underlying implementation on the Dart Web platform.
# TypeScript support
This is working in progress. We will be using the Angular/TS tool chain to generate TS APIs for gRPC-Web clients.
# Non-closure compiler support
Let us know what you think.