Photonic chips will rapidly outpace the snail-brained humans

24-01-2018 |   |  By Paul Whytock

Global data communications is running into a big problem. As modern living and ever-expanding commercialism puts immense pressures on data communication capacity and speed, so existing technologies are rapidly falling short of what the future needs.

The question is will photonics technologies save the day? Fortunately there is currently enormous interest and investment in moving silicon-based photonics forward and some of the latest breakthroughs suggest it could be a game-changer.

Take for example work by researchers from Oxford, Münster and Exeter Universities that has developed photonic computer chips that imitate the way the human brain's synapses (the mechanism that permits a neuron to pass an electrical or chemical signal to another neuron) works. Interestingly, the photonic computer chips synapses works a thousand times faster than those of the human brain.

This breakthrough is not mere electronic gizmo-ism of an academic nature. This is important because current computer design is actually fairly slow and merely boosting existing technologies to speed things up consumes increasing amounts of power.

It is well recognised that copper cabling is stifling data centre operations and hindering high-speed computing because of its slow data transfer capability.

What silicon photonics could end up providing is increased bandwidth in servers, faster data transfer speeds and efficient system designs.

Basically, silicon photonics facilitates data transfer between computer chips using laser light via optical fibre of nano dimensions that is manufactured within the actual silicon chip.

The silicon is patterned into micro-photonic components. These operate in the infrared spectrum which is used by the majority of optical telecoms systems. The silicon typically lies on top of a layer of silica in what is known as silicon on insulator.

To give you a real-world example of what such a development with a transfer capability of 50G/sec could do, it would be able to transfer the file of high-definition feature length film in under a second.

But also true of the real world is the fact that even highly innovative and advantageous technological breakthroughs have to be price competitive.

This is not necessary a particularly difficult hoop for silicon photonics to jump through as chips using the technology can be made from standard 300mm wafers using existing complementary metal oxide semiconductor (CMOS) fabrication processes.

Given all these facts it’s not surprising that industry analysts are saying that the global expenditure on silicon photonics is set to rocket.

Estimates vary but back in 2015 the world market for silicon photonics was reckoned to be about $500 million. Today that market is expected to grow by an impressive 22% annually.

Making a substantial contribution to that growth pattern are optical transceivers and today the silicon photonics market is dominated by the USA and Europe which hold 70% of the total global market share. But that could be set to change because Chinese data centre operations are rapidly growing. It currently has around 650 million Internet users and three big companies establishing their own fibre networks. Predictions are that three huge data centres will be implemented there this year.

So the data rates race is well and truly on with 100G/sec already in reach. However, in the long term the target will be 400G/sec.

It is reasonable to say that silicon photonics is still in its infancy and despite the very buoyant market predictions from a sales perspective there are problems to overcome. Adhering to the different communications standards and dealing with heat issues are prime amongst these, as is possible under-funding and slow development of telecommunications infrastructures.

But nevertheless such problems should be surmountable for the human brain, particularly as it has already shown it is capable of developing something that thinks 1000 times faster that itself. Who said AI is just a fairy tale?


By Paul Whytock

Paul Whytock is European Editor for Electropages. He has reported extensively on the electronics industry in Europe, the United States and the Far East for over twenty years. Prior to entering journalism he worked as a design engineer with Ford Motor Company at locations in England, Germany, Holland and Belgium.

Related articles