Showing posts with the label network

QUIC Notes: Stream multiplexing and congestion control

In a previous post I talked about some of the ideas that drove QUIC, but in this one I want to got a little more into how it actually works to get data from host to host. Again this is all from the notes I made looking through the source and spec - so always go to the code for the latest!All networking is constructed as a series of layers - each layer defining its own mechanisms for reliability, name resolution and other fundamentals. Though QUIC is a single protocol, the layering is still present within it - an individual piece of app data (say a HTML page) will be encapsulated with several different layers of framing: the stream frame, the QUIC session, and a cryptographic channel.This means any given QUIC packet can be peeled back. For example, a regular packet containing part of a web connection might look like: [[CRYPO HEADER] [ENCRYPTED DATA: [QUIC HEADER [STREAM FRAME HEADER [THE STUFF YOU ACTUALLY WANT]]]]]. While this looks like a lot of overhead, there are a few tricks that …

QUIC notes: Rationale, FEC and Head of Line blocking

Having been involved with ZeroMQ for a few years now, and having taken a deeper look at messaging last year, I enjoy occasionally dipping into various network protocols. One of the most interesting recent efforts is QUIC, from the Chrome team (who are just a short bridge away from my Googley home of the Google+ team), which is aimed to provide a post-SPDY protocol to be used instead of HTTP over TCP for high performance web sites.Things at Google tend to start with a design doc that lays out the rational and the plan for a given project, and happily the QUIC doc is openly available. It's a great read, and its worth highlighting some of the more interesting problems the team are addressing. To start with, the document lays out the (12) goals of the project and the motivations for them. These roughly break down into two camps for me: Usable NowThis is a protocol the team intends to deploy, and it is one that has to do the job of existing setups. That means it's got to be efficie…