Last Modified: ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?



Homepage / Publications & Opinion / Archive / Articles, Lectures, Preprints & Reprints

NETWORKS - THE DELAY IN REMOVING DELAY
Peter Cochrane

Prologue
Who would like a three click, one second, no handbook world? Drill down to anything you want in three clicks of a mouse, and it appears on your screen in under a second! No need to read a handbook, no training - just the application of intuition - an obvious and easy to use interface for everyone. The only prospect of realising this dream relies on 'end-to-end' optical fibre and significant improvements in network and computer protocols, interfaces and software.

The Past
Only 20 years ago delays in telecommunication were endemic with a mixture of human, mechanical, electro-mechanical, and electronic switching, plus analogue transmission over cable and satellite systems. The post dial delay of 15 seconds was followed by concatenated switching delays that could add a further 15 seconds or more. The processing speed of computers, information storage systems and printers also introduced seconds or minutes for a screen fill, card, tape or print out.

Today
Now we get irritated if we do not hear the familiar ring tone immediately after we press the last digit on the key pad. Similarly, when we click on an icon and have to wait seconds for an application to load and open, or access from floppy or disc. The generic problem is having to wait for a period that is too short to do anything else, but long enough to break our concentration, and disrupt our interaction and creativity.

Bandwidth
We now have an abundance of bandwidth with optical fibre, solid state processing and storage power, and the promise of increasingly higher levels of circuit density and clock speed at costs that are becoming truly insignificant. We are thus potentially approaching the realisation of a dream - to access everything everywhere within 3 clicks of a mouse and have it on the screen within a second. For real-time, interactive communication, the dream is for realism to ear, eye and touch by making delays much less than 100 milliseconds.

Satellite
Trying to interact with anything at less than natural human speed soon becomes counter productive and extremely irritating. For example: using a telephone operating over a geo-stationery satellite, which introduces over 300 milliseconds of delay. The delay introduces double talks and a stilted form of conversation because we react to the delay period when we perceive that there is no voice or anyone a the far end. Concatenating this with a GSM telephone, which has a 97ms Codec delay built in, and/or an ATM network with uncertain and variable delays, just adds to the fun!

Coding
For anyone trying to communicate effectively over a modern video conferencing circuit, where the Codec and signal path can introduce delays in excess of 500 millisecond in the visual image, with a different delay in the speech circuit- it is even more obvious. Even with a high degree of practice and familiarity such circuits are a far from realistic replication of true human presence and communication. You just cannot communicate and work effectively with such long and variable delays.

Touch
In the not too distant future our ability to transmit sight and sound will be augmented by an ability to transmit touch. It is interesting to note that sticking a pin into the end of our finger results in a message arriving at the brain some 30 milliseconds later. This approximates well to the 30 millisecond period for sound discrimination of the human ear and the movement of objects by the eye. So in a telepresence future a surgeon may wish to reach out and touch a patient in the operating theatre in New York whilst actually being in London. The effective distance between the brain and finger tips will then have doubled from the present 25 to 50cm. Without the synchronisation of sight, sound and touch it will be impossible to interact effectively with a distant world in any co-ordinated way.

Internet
There is no doubt that Internet is a path finder and precursor to the information superhighway. However, organic growth with concatenated electronic bottlenecks, restricted bandwidth, and long unspecified transmission delays make it resemble a crowded footpath rather than a highway. Moreover, the organisation and assembly of information make it something of a bizarre bazaar. Somewhere is the information you are looking for, if you could only find it! What is actually required is the information shopping mall with information organised in a regular way, where you can find what you are looking for, and where there are established pedigrees, prices and values.

Information Super Highway
The super highway is not:Internet, Cellular Radio, PSTN, ISDN, ADSL, ATM, Satellite or Cable TV. Neither is it information transported over networks that are asymmetric, or incur excessive delays, or constrains bit rates to less then 1 Gbit/s. It is ultimately about instant gratification and access to information at anytime in the right form and at the right price. As we are able to absorb information at up to 1Gbit/s and perceive delays of the order of 30ms, longer delays become increasingly irritating and disruptive. The only prospect of realising adequate delay performance lies with optical fibre to provided a true super information highway with minimal electronic delays, bottlenecks, and processing and bandwidth restrictions. It also requires a new attitude to software, and the processing and storage of information. All the technology required is available now: all we have to do is implement the solution.

Terminals
Personal computers are doubling in power every one to two years. Common clock rates are now 25, 50 or 100 MHz, and at the top end, power PCs are using 160MHz, with the next generation aiming at over 300MHz. The amount of RAM is now commonly 8MBytes and rising to 32MBytes with disk storage up to 500 or 1000MBytes. Given these trends, it is clear that current network technologies are likely to introduce significant bottlenecks. Any inability of telecommunications networks to respond to the demands of their peripheral terminals will fundamentally constraint progress in the first part of the next millennium. The right technology options and designs have to be selected now.

Interfaces and Protocols
There is no doubt that the fundamental delays now facing us in terms of instant gratification are solely related to interfaces and protocols. The industry continues to cling onto copper related relics such as X25, X400 and programmes of unbelievable complexity compared to the task they address. It is not unusual to find a standard application requiring 5 - 6 MBytes of RAM, with <0.5MBytes of core software and the rest devoted to making the screen pretty - and introducing more delay!

The Future
Perhaps at some future epoch we will decide that delay is really the only key parameter of importance remaining, and then take the brave step to optimise the complete information system. The network element should prove straightforward - we have the technology, and we can see how to do it. Software, protocols and interfaces present the ultimate challenge - in that order!

Word Count = 1162

Biography
Peter Cochrane is currently the Head of Research at BT Laboratories and a visiting professor to UCL, Essex and Kent Universities. He is Fellow of the IEE, IEEE and Royal Academy of Engineering. Since 1973 he has worked on analogue and digital switching and transmission systems, hardware and software design, and network studies. His team received the Queen's Award for Innovation & Export in 1990, and he received the Martlesham Medal for contributions to fibre optic technology in 1994; the IEE Electronics Division Premium in 1986, Computing and control Premium in 1994 and the IERE Benefactors Prize in 1994.

All materials created by Peter Cochrane and presented within this site are copyright ? Peter Cochrane - but this is an open resource - and you are invited to make as many downloads as you wish provided you use in a reputable manner