Idea: Using the SPDY protocol to improve Tor performance

Adam Langley agl at
Fri Feb 5 16:08:35 UTC 2010

On Fri, Feb 5, 2010 at 11:03 AM, Nick Mathewson <nickm at> wrote:
> I think there's a more generic mechanism lurking in the wings here.
> SPDY is only one of several possible exit-side optimizing techniques
> that might exist, and it seems overkill to add a new relay cell type
> for each one we want to play with.  I'd suggest that instead, the
> begin cell should tell the exit what kind of traffic transformation it
> ought to perform, if any.  This could either be a new field in the
> standard begin cell format, or it could be a new
> RELAY_BEGIN_somethingorother cell type.

One thing to keep in mind: SPDY assumes that all clients support gzip
encoding so you will probably want to munge the Accept-Encoding header

Also, we haven't developed any SPDY proxies yet and haven't yet added
per-stream flow control to the protocol.

If you have any SPDY questions, please feel free to ask myself (I'm on
the SPDY team here at Google) or spdy-dev.



Adam Langley agl at

More information about the tor-dev mailing list