3 min read
Back in 2019, even before the rise of artificial intelligence (AI), it was clear that the world was facing an exponential increase in data. It didn’t take a crystal ball to see that there was no way the current networking and computing infrastructure would be able to meet the demand. NTT, drawing on its experience as the first company in the world to build fiber to the home, and provides undersea fibre optics cables around the world, realized that photonics could offer a solution. So the concept of a photonics-based networking and computing infrastructure was born.
And, as AI continues to develop, photonics is becoming more and more important to the infrastructure of the future.
Why AI demands new infrastructure
There are many business cases for AI of the future. The possibilities are almost endless. But they require robust technology and networks in order to make them a reality. There is a limit to what can be done with today’s electronics-based technology.
For instance, right now it is not possible to push any more computing through silicon because of the overwhelming heat it would create – but imagine the capabilities if you could have an AI processor in your phone. Or consider the benefits of automated driving, which can’t be fully realized yet because it needs billions of sensors that can’t run on electronics and batteries due to power limitations.
Our current infrastructure simply cannot support the volumes of data and computing required. AI will greatly accelerate the amount of electricity needed to generate computing power. With these demands on energy, as well greater speed and lower latency requirements, AI is changing the way we think about networks and computers.
This is where photonics comes in.
The advantages of photonics
There are three key advantages to photonics:
- It uses significantly less power for the same amount of capacity.
- It’s much faster than traditional networking infrastructure.
- It has low latency. Low latency now has become a critical requirement for AI, especially with robotics and edge computing.
From the data center to devices to cars, photonics can support massive amounts of data, and it can do it efficiently because it focuses on the supply side of power and energy. For instance, instead of requiring alternative energy sources to address the growing need for more compute power, photonics can supply the same power while with less energy – while still providing, or even enhancing, network performance. It also doesn’t generate massive amounts of heat from data centers, thereby removing the need to find new ways of cooling. Think of it like the transition from incandescent to LED lighting. LED’s use significantly less wattage while still producing the same amount of light
That’s not to say that renewable and alternative energy isn’t also critical, because it is. In fact, we are committed to net zero, and 45 percent of that is by implementing photonics-based technology. The company is committed to photonics as its infrastructure, both on the network and on the computing side. The other 45 percent of NTT’s commitment to net zero is through renewable energy.
Collaborating to build the future
Building photonics based networks and computing is not something that any one company can do alone.
That is why NTT partners with 150 of the world’s largest companies as part of the Innovative Optical and Wireless Network (IOWN) Global Forum. IOWN uses an intelligent system that is not just a fiber but is also made up of partners and includes everything including software, hardware and other technologies right through to the devices at the endpoints. The key to this is that the shift to a photonics-based network is not just about using fiber optics for fast transmission but for more intelligent transmission. And that is the infrastructure that will truly enable the AI of the future.