Research Efforts

There are a number of research efforts on how to improve real-time communications across the Internet. A small subset is listed below. This section needs lots of pointers to on-going work.
Multicast address allocation:
Eventually, many thousands of multicast sessions may exist concurrently. Currently, the IPv4 multicast address space is very limited and thus requires careful global allocation to avoid collisions. The IPv6 multicast address space is very much larger, supports administrative scoping and may allow random allocation.
Scalable multicast routing:
Multicast routing needs to work for both a very large number of small groups and a smaller number of large groups, without routers not on the multicast tree having to know about groups.
Compensating for packet loss:
For the foreseeable future, the Internet will have areas and times of high packet loss (1% to 10% and higher).
Playout delay compensation:
End systems need to compensate for network delay variations.
Synchronization of different media:
Several audio and video streams coming from one or, less commonly, several sources need to be synchronized (lip sync).
(Semi)reliable multicast for (near) real-time services:
In some circumstances, such as near-real-time video and audio on demand, it may be possible to improve quality by retransmission.
Internet integrated service models:
It is not clear yet how many different services beyond the current best effort model are needed to support a wide variety of real-time and near real-time services.
Making use of ATM quality-of-service features:
It is desirable if Internet services can make use of the quality-of-service guarantees offered by subnetworks.
New conference control mechanisms:
Different modes of establishing one-on-one and group communication are being explored.
Interoperation with POTS, ISDN, T.120, ...:
For at least the next year, the Internet will not replace the telephone network for voice calls. Thus, interoperability with plain old telephony, ISDN and some of the ITU-defined circuit-switched video conferencing equipment is needed.
Composition of multimedia applications:
Rather than having a single application that handles all media, it may be preferable to compose conferencing and other multimedia applications from reusable building blocks.
Integration of real-time services with WWW:
Both delivery and interactive multimedia services need to be more closely woven into the World Wide Web.

Last modified: August 8, 1996 by Henning Schulzrinne