Re: [AVT] FW: New Version Notification fordraft-begen-avt-rtp-cnames-01

["Ali C. Begen (abegen)" <[e-mail]>]

> -----Original Message-----
> From: Roni Even [mailto:[e-mail]]
> Sent: Thursday, May 20, 2010 3:06 AM
> To: Dan Wing (dwing); Ali C. Begen (abegen); [e-mail]
> Cc: [e-mail]
> Subject: RE: [AVT] FW: New Version Notification fordraft-begen-avt-rtp-cnames-01
> 
> Hi,
> 
> >
> > * Question: Are there any ideas on the table for generating a unique
> > identifier that wouldn't be useful for traffic analysis?  That is, a
> > unique
> > identifier that is generated on each call?  For example, a BASE64-
> > encoded
> > SHA1-HMAC of one of the three pieces of information currently suggested
> > in the
> > memo (IPv6 address, unique FQDN, UUID), where the 'secret' is changed
> > for
> > every new RTP session?  This seems it could eliminate traffic analysis
> > concerns.
> >
> 
> RFC 3550 says that one of the usage of RTCP is for monitoring and "To
> facilitate third-party monitoring, the CNAME SHOULD be suitable for either a
> program or a person to locate the source."

What does "locate the source" imply?

We also saw this text the other day and actually I wanna understand this a bit more. The concern Dan is raising is valid in my opinion. The current 3 methods we have in the draft are sticky. If one uses the same CNAME in totally irrelevant RTP sessions, a monitoring agent can still infer that those users are the same person. For several reasons, this is not a good thing. That's why I am inclined to include Dan's suggestion (e.g., hashing) as the 4th method for those who want to take that option. We will also mention this risk in the sec. section.
 
> 
> Other comments
> 
> 1. For the UUID I assume the proposal is to use the string representation of
> the UUID maybe as a URN.

This is the second issue and let's have WG's input on this. Do we need ascii CNAMEs or not? Does it everything need to be encoded in base64?
 
> 2. As for the structure of CNAME, section 6.5.1 of RFC 3550 has the cname
> composed of user and domain name. it also says that
> "The user name SHOULD be in a form that a program such as "finger" or "talk"
> could use, i.e., it typically is the login name rather than the personal
> name.  The host name is not necessarily identical to the one in the
> participant's electronic mail address. "
> Maybe this document should recommend how to create a unique host part of the
> CNAME and not the whole CNAME.

If the host part is unique, why would we still need user@host?
 
Thanks for bringing these up. Let's reach an agreement and we will submit a revision accordingly.

-acbegen

> Regards
> Roni Even
> 
>