Re: Introducing a Session header...

In message <CABP7RbfSHZ494pmFSYFyX_g7M3YQyuG0GM4RZ2NJJ22nVMTvig@mail.gmail.com>
, James M Snell writes:

>A. We need a protocol-level concept of a "Routing Token" to address #2. The
>use of this mechanism is strictly to provide intermediaries with a method
>of establishing a generally-persistent routing path for all requests
>originating from the same client.

And here it might get tricky:  What do we mean by "client", if every
public PC in the library has been routed through the same HTTP-filter ?

I suspect we actually mean "user" and not "client".

(This is btw. where my "anon/identified-user" bit idea comes from)

>This token is NOT intended to be used to
>identify or access shared application state. In fact, it is questionable as
>to whether this routing token should be visible to the application layer at
>all.

Well, if the client-end is entirely in charge of producing the
session-id, it is in essence just a random number, and there are
only so many bad things you can do with a random number, so if the
application can beneficially use it (to look up server-side state ?)
then it should be exposed.

The "anon/idendtified" bit should always be exposed if we do that.

>The token could, theoretically, be transmitted entirely in the clear
>without risk of inappropriate disclosure of sensitive information. The one
>caveat, however, is that there must be a mechanism to protect the integrity
>of the token.

It is at this point, not clear to me if the token would be hop-by-hop
or end-to-end.  There are significant advantages to both, hop-to-hop
in particular if the "chain" gets broken by HTTP/1.1 hops underway,
but "end-to-end in HTTP/2.0 protocol" solves that too.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.

Received on Friday, 20 July 2012 20:16:31 UTC