[V3] RealTime Internet Peering for Telephony (RIPT) BoF Proposal

We are discussing a BOF at IETF 107 on the RIPT (used to be RIPP) on the v3@ietf.org mailing list. 

Please join the list at https://www.ietf.org/mailman/listinfo/v3

Thanks, Cullen 

Description:

Though extremely successful and deployed on massive scales, the
Session Initiation Protocol (SIP) - as a 20 year old protocol - is
showing its age. It is difficult to deploy SIP technologies into
modern public cloud platforms. Its usage of raw IPs and ports, its
load balancing techniques, and its state model, make it incompatible
with services like global load balancers, autoscaling, service meshes,
and so on. 

As such, the goal of this BoF is to create a new protocol that
enables signaling and real-time interactive voice and video to operate natively
ontop of modern web platforms. This means that:

1. It needs to use HTTP URLs for addressing, and media and signaling
can be sent in this way. 

2. It follows the HTTP connectionless model, in that, clients can refresh
their state and manipulate the state, and can do so from any different
number of clients.

3. It aligns signaling and media state into a single piece of state,
allowing clients to manipulate it with HTTP rpc operations

4. It utilizes modern http cloud security techniques, including OAuth,
server-side TLS, and so on.

5. It utilizes modern techniques for load balancing and high availability
used in scaled web cloud environments

6. It makes use of modern http rpc techniques and data formatting

7. It follows modern web practice and makes security features mandatory,
not optional (including e2e media encryption, though the media flows HbH)

The recent efforts to standardize HTTP3 using QUIC, make it possible to
achieve the above without unacceptable latency and HOL blocking, inherent
in TCP-based protocols.

To remedy this, a design team put together a draft specification
called RIPP, which takes advantage of HTTP3 and QUIC to enable
real-time communicatins over those protocols, without extending or
changing them.

The purpose of the BoF is to gain consensus on the formation of a
working group to work on this problem. A strong goal is to 
build code and specifications simultaneously. 

Relevant Drafts:

https://datatracker.ietf.org/doc/draft-rosenbergjennings-dispatch-ript/
https://datatracker.ietf.org/doc/draft-rosenberg-dispatch-ript-inbound/
https://datatracker.ietf.org/doc/draft-rosenberg-dispatch-ript-sipdiffs/
https://datatracker.ietf.org/doc/draft-rosenberg-dispatch-ript-webrtc/