DRAFT TEXT v2 : All contributions most welcome

[Gorry Fairhurst <gorry@erg.abdn.ac.uk>]

DRAFT TEXT  4th Oct 2001


This is the second draft of the document being prepared to come
before DVB to present the case for the existence of
this mail group and the work we intend to do.  The future
of the mailing list depends on finding a suitable relationship
with DVB technical module and other such organizations. 

Contributions of text  and/or comments from all on the list are welcome.
Please send to:

ip-dvb@erg.abdn.ac.uk

---------

Transporting IP over Satellite

Contribution to DVB-?? Meeting of ???? presented by Keith Fenwick
(Hughes Network Systems) and prepared by Keith Fenwick, Gorry Fairhurst
(University of Aberdeen), Adrian Jelffs (Skystream),

<<< add more contributors as text received>>> 

Summary

This note addresses the problems of carrying IP data over DVB, reports
on work being proposed for study in the Internet Engineering Task Force
(IETF) to improve transport efficiency and proposes that DVB should be
party to this work.

Background

The need for DVB systems to carry data has already been recognised. It
has also been shown that problems exist with the current MPE
specification, which led to an RFI being issued by DVB-IPI AHG.  This
RFI requested information on current deficiencies (with any solutions)
for the current MPE specification.  The responses to the RFI are now
being handled by the DVB-SI AHG.  <<INSERT REFERENCE TO RFI>> The
majority of the RFI responses have been concerned with the signalling of
information to resolve the DCV service from the IP address.
In addition, the expertise of DVB members tends to be in delivering
traditional broadcast services and not data delivery using IP. 
In contrast, the Internet Engineering Task Force (IETF) has a
long-standing tradition of developing  high quality, open protocols
which have a long life-time of use. Examples of work in this area
include IP over Ethernet, IP over ATM, IP over Firewire, IP over Cable
Data Networks, TCP over Satellite. The IETF also provides a peer-review
of all documents issued, which ensures that the document is examined by
experts from applications, operations, routing, enetworking, security
etc. However, the scope of IETF is normally limited only to the services
provided to the IP network layer and the issues concerning IP
networking. 

Using DVB Systems to Carry IP Traffic

DVB systems need to carry IP traffic to provide true multimedia and
interactive services. In addition, some DVB links may support only IP
services.  Multi-Protocol Encapsulation (MPE) has the advantage that it
is an existing ETSI standard, which has been widely implemented.  It is
therefore likely to exist as such for the foreseeable future.
MPE however has some disadvantages when  used for services supporting
IPv4 and IPv6 traffic which include:
1. It wastes up to 15% capacity in unnecessary overheads (MPE Tax). A
more efficient scheme may eliminate may of the headers present in MPE,
since they are not required for IP services.
2. It is too complex, incurring significant per-packet processing
overhead. A simpler more efficient scheme may significantly reduce
receiver processing costs.
3. It contains many options, which in practice are not required for many
IP services. The extra options reduce the inter-operability between
vendor implementations.
4. Some desirable functionality is missing, including: address
resolution of IP addresses to MPEG2-Transport Stream (TS), Packet ID
(PIDs), support for IPv4 and IPv6 multicast. The use of header
compression may also be desirable in some circumstances.
5. Some additional features are desirable, e.g. SNMP support.


Potential Solutions

At present, only outline solutions have been proposed, including a
leaner encapsulation. A leaner encapsulation, offering faster
processing, may also permit a more flexible use of MPEG-2 PIDs.  The
position of the encapulsation form the IP network viewpoint, and its
relationship with the DVB systems stream needs also to be considered to
ensure good and efficient interoperability.

The following principles are suggested for this work:

(i) Minimal overhead.  The proposed encapsulation should minimize
protocol overhead, in terms of the number of additional bytes to be sent
in addition to the IP datagram, and the processing load which is likely
to be envisaged when implementing such schemes. In some cases, the use
of header compression may represent a useful trade-off in increased
processing overhead, but reduced packet header size.

(ii) Robust operation. The encapsulation should rely on the End-to-End
principle when proving reliability. As such, it should minimise the
impact of link failures on the end-to-end transmission. The need for and
appropriate use of ICMP messages needs to be defined.

(iii) Fate Sharing. The encapsulation should not assume it is the only
path from the sender to the receiver.  It should permit alternate paths.
Transparent link fragmentation should be used.

(iv) Minimal set of required options.

Integration with IP Quality of Service (QoS) mechanisms may also need to
be considered.

 >>>> Details to be added. ? CONTRIBUTIONS WELCOME <<<
 
 
Recommendations

DVB supports and contributes to the work being proposed within IETF to
take advantage of the knowledge within that group and to align with
other transport systems, e.g. IP over Cable Data Networks.

References

EN 301 192 : Digital Video Broadcasting (DVB); 
DVB specification for data broadcasting 
European Standard (Telecommunications series)

S. Bradner, IETF Working Group Guidelines and Procedures, 
RFC2418 (BCP25 ), September 1998.

Bradner, S., Editor, The Internet Standards Process -- Revision 3, 
RFC 2026 (BCP 9), October 1996.

Hovey, R., and S. Bradner, The Organizations involved in the IETF 
Standards Process, RFC 2028 (BCP 11), October 1996.

<< Add more references received via mailing list >>