When Internet Protocol was defined, communications links were slow and memory and hard-wired logic were expensive, so it was appropriate for all packet routing functions to be implemented in software and for routers to hold as little state information as possible, i.e. for routing to be connectionless. Since then, memory sizes have increased by a factor of a million or more, and transmission speeds by maybe 100,000, but CPU speeds have only increased by a factor of about a thousand, and memory interface speeds by not much more than that. Therefore, in today's world, routers can keep much more state but the amount of processing required to forward an individual packet should be minimised.
Also, the Internet was designed as an IT network, for tasks such as sending e-mails and downloading web pages. By their nature, these applications transfer unpredictable amounts of data at unpredictable intervals, and simply need it to arrive at the destination as soon as possible. Now, IP networks are increasingly used to carry live media such as video, audio, and telephony, in regular streams of packets that are all of the same (or similar) size and need to arrive at their destination at regular intervals. Research during the past decade has shown that there are other technologies that would provide a much better service for this traffic, as well as being simpler and more reliable, and would also be better for carrying the IT traffic. Moreover, they enable new business models, as well as being much more secure than IP.
International standards development organisations (SDOs) are now working on these new technologies. Participation is needed from experts in all the fields that will use the new infrastructure, particularly those that are not well served by the current technology. This is your chance to influence the specification, and make sure it includes the facilities your industry needs.
A trial implementation of one of the possible technologies has been built using a platform designed for the purpose. Follow this link for details.
ETSI's Industry Specification Group on Next Generation Protocols is looking at technologies that would be suitable for next-generation Radio Access Networks, connecting mobile devices to the Internet and other services.
Details of its scope and how to join, as well as copies of its output documents, are available here.
This project in IEC TC100 TA4 is developing a set of control protocols for the new switching technology. It will also support legacy technologies and networks consisting of part new and part legacy technologies, to aid migration to the new technology. Details of Part 5 (Transmission over networks) are here and of the other Parts here.
Join the project's LinkedIn discussion group here; alternatively send your comments to FNfirstname.lastname@example.org.
This group's work on Technical Report 29181 "Future Network - Problem Statement and Requirements" is nearing completion; Parts 1 to 7 have been published, and Part 1 (Overall aspects) is available for free, from here or via this direct link.
Standardisation is now beginning, with the first drafts of the "switching and routing" Parts of ISO 21558 (Architecture) here and ISO 21559 (Protocols and mechanisms) here.
Join the project's main e-mail discussion list (for discussion of all Future Network topics) here. Join the list for messages specifically related to switching and routing here.
We still need more input to inform the standardisation process, particularly hard information on requirements for particular industries or applications, for instance:
Please post to the e-mail discussion list or write to FNemail@example.com.
Below are some numerical values that have been identified for user requirements. For instance, many applications require low latency, but how low is "low"?
Above 150ms conversation becomes un-natural; it is difficult for speakers to talk against and over each other. This is the "round trip" time, including coding delays and two transits across the network, so the worst case network latency should be less than (150/2) = 75ms.
Above about 40ms musicians have difficulty in performing when hearing their own sound via headphones or loudspeakers. The exact figure is different for different instruments (for speech it is about 50ms), and even at lower latencies the delay is noticeable. Again, this is the total "round trip" time, so network latency needs to be below 20ms.
Please post to the discussion list if you can contribute further numbers, e.g. latency for remotely controlling machinery using a video link.
The ITU-T Focus Group on Future Networks reported at the end of 2010. Their outputs, and the results of further work on this topic in SG13, have been published in Y.3000 series documents available from here.
A further Focus Group, on IMT-2020, has been formed. IMT stands for International Mobile Telecommunications, and IMT-2020 is the generation which most other groups call 5G.