Re: Recognising one's own messages on an anonymous broadcast channel?

<bergstrom.henrik@xxxxxxxxx> wrote:
A user sends operations to our server on a private channel. The result
is broadcast to all users (cannot be sent on the private channel).

The client includes a nonce (chosen uniformly at random) in the message
sent to the server over the private channel. The server includes this
nonce in the result.

1. The result must be anonymous, i.e. no other client must be able to
identify which client that sent the operation to the system or that
two operations belongs to the same client.

Everything sent over the private channel stays secret. This is ok.

2. The client that sent the operation must be able to identify the
result as his own.

That's ok.

3. No collisions! I.e. no client should mistakenly identify a result
as his own.

Ok as long as the nonce is sufficiently long - 128 bits should do it.

Prerequisites:

o It's ok to add "user id:s", "correlation id:s" etc. to the input
messages.
o It's ok for the server and client to share secrets, if necessary.
The input channel is a persistent connection.
o The algorithms have to be message-based rather than stream-based.
o The algorithms used have to be "fast" since a quite high number
(100k) of messages are processed per second by the common server.

These all seem to be satisfied.

--
Kristian Gjøsteen
.

Relevant Pages