An array of unprecedented global initiatives and strategic partnerships have quietly taken root over the past several years. Examples include such large-scale, analog-digital ventures as the United Nation's "17 Sustainable Development Goals," powered by the Turing Institute; the partnership between geospatial software giant Esri and Microsoft to create a Planetary Computer; the Graphene Flagship and Human Brain Project, both massive EU-based, public-private collaborative ventures; The Global Partnership for Sustainable Data Development, which – with more than 250 partners – is harnessing the data revolution for sustainable global development; and a U.S.-based, insurance-industry-led consortium, RiskStream, described as a 40-partner collaborative that plans to unlock the potential of blockchain across its industry.
These "big idea" collaboratives take advantage of the borderless nature of digital communities and have already begun the hard work of mapping, monitoring, measuring, comprehending, and resolving large-scale problems in areas as diverse as world health, land and wildlife management, banking and insurance, and data and data-traffic management.
While diverse in mission, global initiatives in 2020 and beyond will be joined by similar dependencies: access to light-based or photonic technologies, and the need for a computing-communications network that reliably maintains connections and efficiently gathers, processes, and shares data.
There is an urgent need to become educated about photonic technologies, and – because data traffic is exploding and stressing the current infrastructure – there is a second urgent need for a computing-communications system capable of meeting the problem-solving demands of the future.
Beyond human cognition
If photonic technologies are going to successfully integrate with and support our day-to-day lives and lift the types of transformational projects mentioned above, a paradigm shift will be necessary. One such shift is the need to accept that technological baselines will be set to perform at rates beyond human cognition (BHC).
Currently, most digital technologies are optimized for the human cognitive system – for example, the frame rate for video is determined by what the human eye is capable of seeing – but in order to evolve, technologies cannot be constrained by human cognition. Removing this limitation will enable innovation across fields.
A better communications infrastructure
In the years to come, the success of day-to-day work and visionary analog-digital initiatives will depend on a core global communications infrastructure. It is important to collaborate to improve communication capacity, latency, data efficiency, and computing scalability.
To address these issues, and ultimately, build a better communication and computing system, technology and communications industry stakeholders will need to take the following steps:
Move toward data-centric communication and computing
The industry will need to have a fundamental shift from today's application and internet-centric infrastructure. In the current infrastructure, driven by sensing and artificial intelligence technologies, the data generation rate will always outpace communication network capacity. By shifting to a data-centric approach, the network and computing infrastructure can be designed around data in a way that optimizes data efficiency.
Accelerate full-stack communication
In recent years, we've seen great advancements in optical communication and radio access technologies, but capacity barriers remain. For example, using TCP/IP reduces speed as communication round trip time increases. What's more, network virtualization and the projected increase of communication throughput by as much as 100 times will overload the existing infrastructure. These bottlenecks would require new technologies that achieve significantly higher capacity and speed.
Expand computing scaling across device, network, edge, and center cloud
Computing scaling has become increasingly complex as new devices like accelerators and persistent memories have emerged. This results in challenges with data-intensive computing applications like deep learning. These applications cannot be implemented without accelerators, but the performance benefit of an accelerator depends on the speed at which the application can transfer data to the accelerator. There are existing solutions to this issue, but they are meant for data transfer within a single data center. As the network becomes more interconnected, it will be critical to develop a computing scaling platform that can expand across device, network, edge, and center cloud.
Create sustainable growth with energy efficiency
For continued sustainable growth, energy efficiency should be considered across communication, data, and computing in both the component and system levels. In the component level, key factors contributing to energy efficiency include: advanced DSP schemes, continuing enhancement on processor design, further shrinking of process nodes, silicon photonics with advanced packaging, energy-conscious air interface design, energy-conscious access network, and transport network design.
Achieving a quantum leap in energy efficiency will happen sooner than we think but will require fundamental architectural work at the system level on computing and communication system infrastructures to leverage the component-level advancement.
Communications, data, and photonic technologies are making human and technological connections faster, smarter, stronger, and more economically feasible. Without making these changes to the global communications network, exciting new ideas and visions may lose momentum and languish.
Dr. Katsuhiko Kawazoe is president and chairperson of the Innovative Optical and Wireless Network Global Forum.