Clear Sighted Transparency

Posted by Peter Cochrane on July 14, 2012

Invited Keynote: 14th ICTON IEEE Conference, Warwick University July 2012

Peter Cochrane

www.cochrane.org.uk

Positioning

Today's networks mirror their electric telegraph and telephone forebears with working practices, building/facility locations dictated by a copper past. Many of the latest optical networks still utilise old planning and design rules, but TONs offer new opportunities to break the legacy with direct links between countries, islands, cities, towns and villages. They also offer future proof solutions devoid of 'bandwidth blocking' electronics, and very often, do not require optical amplification. The attendant energy demands and reliability risks associated with electronic transmission and switching are also displaced by 'optical only signals' of a hybrid (analogue-digital) form. The operational opportunities and economic gains of TONs combined with non-linear optics cannot be overstressed as they eclipse anything and everything that has gone before.

The Rise of Uncontrolled Complexity

Whilst the electronic revolution has enhanced every aspect of human life, it has also created a world of complexity rivalling that of the biological domain.  This is a new world that we do not fully understand and are unable to fully model or control. We are no longer able to design and build many of the machines and devices that we take for granted, and Artificial intelligences have assumed a design responsibility for much of our integrated electronics and control systems. They also manage robotic assembly lines that produce everything from mobile phones, laptops, cameras, TVs, networking equipment, through to our cars clothing, food production and more.

Our networked dependence is now total and we no longer enjoy predictors of the emergent behaviours these 'man made' systems invoke when connected into global networks.  A dynamic concatenation of complexity is now the norm with a rapidly evolving population (exceeding 5.5Bn) of mobile devices operating over networks that combine technologies spanning three or four decades of electronic and electro-optic development.  

What Might Have Been

As long ago as 1985 – 1995 it was evident that Passive Optical Networks, Optical Amplifiers, WDM, Optical Switching, Optical Signal Processing and Optical Coding and Encryption offered a 'Transparent Future' with bandwidth, and bit rates an insignificant parameter considered to be 'future proofed'.  The cost of such networks at a long-lines and local level looked to be around 10% of the existing copper and microwave infrastructure.     Radio (analogue & digital)  transmission over optical fibre had also been demonstrated, and it was even possible to see how we could switch IP packets optically in a far more efficient manner than any electronic solution.  At a local, short & long haul level optical networks also costed in against copper based solutions by a wide margin. 

What Actually Happened

The PC, mobile phone, and the Internet created an investment focus that saw the majority of resources directed towards electronic computing and software and away for optics.  And so, optical fibre, optical devices and electronic signal processing came together to provide the undersea and intercity long-lines networks we enjoy today.  They did a good job of enabling the internet (as we know it) whilst totally displacing microwave radio and satellite systems, but had little effect on copper in the local loop.  Here the resistance of the old was both extraordinary and resolute.  FTTH penetration has therefore only been significant in Japan, Korea, China, and Scandinavia – but almost nowhere else.  

What Happens Next

Today we sit at a new technological crossroads that will see humanity move from passive absorption to active engagement – a world where the static PC is replaced by the mobile device for hand and pocket. The coming era of Cloud Computing will change everything and demand more bandwidth than is currently available at every level in our networks.  The number one concerns are the 'Last Mile' and mobile networks that are bogged down with legacy thinking, investment, technologies and topologies. In short; no bandwidth – no connectivity – no cloud – no competitive industrial/societal progress.

The Cloud isn't a return to The Mainframe!  It is something entirely new combining new human interfaces with networked devices including embedded electronic sensors and on-line access to data and Artificial Intelligences. Super computing and networking/information/data/knowledge/educational/training capabilities will be available to all along with new forms of group working, employment and interactive entertainment.  It also signifies a big change to more 'things' on-line than people, where data and meta-data will be used and exploited at every level in real time, and networks that deal dynamically with more mobility and more interaction than ever before. 

What We Know For Sure

Legacy thinking; management, employment, world models and the systems of the past; the networks of today; existing electronic bottlenecks and latency traps; and the internet as we know it will not cope or service the new demands about to be created by The Cloud. To satisfy this will require more than greater bandwidth at all levels of our networks, it also demands far greater levels of transparency.  Ultimately, this will mean the wholesale replacement of electronic switching and routers with optical based alternatives.  In some respects this will also mean the displacement of existing 'packet formats' as we know them with hybrid groupings and protocols able to cope with more than one 'all time fixed' signal format.  It will, in effect, have to become a 'format agnostic network' with distributed intelligences built in.

On an entirely different level; the rise of wireless connectivity will demand a move away from the 'fork lift and high power' infrastructure with 1000s of towers spaced 1 – 10km apart in a semi-honeycomb cellular pattern, toward low power networks consisting of millions of local nodes at the end of optical fibres in every home and office supporting communication over 10m or so. 

The Curse of Clustering

Mobility is now one of the biggest generators of complex behaviours in networks.  Clustering happens at every level in the universe and chaos is the natural mode, and on a people scale, we cluster when and wherever we work, live and travel - and therefor so do our devices and network activity.  Human conversation and interaction promotes a correlated demand to communicate from the same place at the same time.  For example; a conference venue or motorway are largely empty and free from telecom traffic demands for the majority of the time.  But when a conference is in progress, or there is a traffic jam, then the demand routinely exceeds the capacity provided.

In mathematical terms; the world we are trying to engineer is fundamentally non-linear, chaotic and almost wholly unpredictable. Strange attractors manifest themselves in ways that are hard for us to imagine:  coffee-time at a big meeting/conference sees a step function in communication, and so does an intermission at a concert, a break between classes, a goal scored at a soccer match, an impending terrorist attack, a sale of prized goods, a phone in programme on the radio and TV, and so on.

The only long term solution we have in our artillery to address this is network transparency and a surfeit of bandwidth. The good news is: we have the technologies and we have the knowledge.  Moreover; bandwidth turns out to be the cheapest of all the IT commodities that we can produce and deploy. By comparison; faster integrated circuits; denser memory/hard drives are orders of magnitude more expensive than optical bandwidth provision and cost of ownership is also far lower.  Even better: by and large all the optical fibre we need is largely deployed and in the ground now – or at worst the duct tracks are in place – and all we have to do is deploy well engineered solutions. 

The Big Challenges

Technological change seldom poses any significant challenge. It is always the people involved! Legacy investments, thinking, working practices are extremely difficult to displace. In some cases you have to wait for a generation to die or move on before the necessary changes can be made.  But we can no longer wait that long.  We are talking about the long and near term future of populations; the enabling of new industries and modes of manufacture and distribution; realising sustainable economies in symbiosis with nature rather than destroying the raw materials & resources on which we all depend.

A Bit/Byte at a Time

Will TONs happen and get deployed fast? No! Will they arrive in time?  Perhaps!  Whilst things are looking good and straightforward at a long lines level with massive connectivity and upgradeable bandwidth capacity in place, the last-mile or local-loop poses the most immediate threat to total transparency.  Progress is patchy and resistance is high. None-the-less it is entirely feasible that widespread deployments will occur in the next decade and could be total within 20 years.  So, is there anyway it could happen faster? Not a chance!  The civil engineering and 'trained body count' required dictate the time to complete along with the investment capital availability.

The second big challenge are the electronic nodes – routers and switches!  Can these be replaced by all optical units? In theory yes; but in practice it could turn out to be a bigger problem than the last-mile. The good news is; the mind sets here are not as retarded, and they are more open to winning formulas.  They also have investment capabilities, and the competitive landscape is far more pressured and volatile.  So what is a realistic time frame for change here?  Again we are probably looking at 10 – 15 years (or longer) before optical solutions overtake the domination by electronics. 

Innovation to Come

The new materials and fabrication techniques coming out of the nano/bio revolution are opening doors into a new world of intelligent optics & intelligent optical fibre.  To date; signal amplification, wavelength translation, packet recognition, switching and sensors just about define the full availability.

“All the innovation, invention and development we have seen with linear optics is likely to be totally eclipsed when we migrate into the non-linear domain and fully understand the potential of what is on offer” 

Soliton generation; basic transmission and processing with non-linear fibres and compressive amplification, wavelength translation are just the start.  When we include polarisation modulation/control, multi-level signalling and all the sophistication enjoyed in the radio domain we have a long way to go.  For example; the intentional non-linear mixing of pulse streams adds new dimensions to the switching matrix.  In electronics we think 'time and space switching', but in optics we can go much further with; time, space, wavelength, frequency (sub-carrier or otherwise) phase and polarisation.

The implications of 'N-Dimensional Switching and Processing' have yet to be fully investigated and the real potential characterised and dimensioned, but at a minimum we can expect to see 'non-blocking' as a prime feature along with 'no-break failures' and self repair lifting network performance to new levels.  Other side effects will include instant fault location, and possible new computation and storage configurations.

“I think we can safely assume that no one understands non-linear systems, and we may never do so, but they are responsible for the creation of the most complex and interesting technology of all - Life” 

Author Bio 

Peter has enjoyed decades of management, technology & operational experience. At BT he progressed from linesman to R&D engineer, Head of R&D and then CTO. His a 1000 strong team engaged in optical fibre, fixed/mobile networks, AI/AL, human I/O design studies. After BT he worked in defence, logistics, travel, retail, energy, healthcare, logistics & pharmacology. He also engaged in founding new tech companies & was appointed  the UK's first Prof for the Public Understanding of Science & Technology @ Bristol in 1998. A graduate of Nottingham Trent and Essex Universities, he received the Queen's Award for Innovation & Export in 1990, many prizes & Honorary Doctorates, and an OBE in 1999 for contributions to international communications.