December 20, 2019
The IEEE GLOBECOM 2019 conference in Waikoloa, HI, brought together nearly 3,000 attendees with a focus on building next-generation networks. IEEE Senior Member Geng Wu, an Intel Fellow and chief technologist for wireless technologies and standards at Intel Corporation, delivered a keynote address on “Future Generation Mobile Connected Intelligence.” He described how future networks must be architected for data efficiency, flexibility and service.
The nature of data is constantly changing. Over the past 10 years, we saw an increase in social media and video-streaming traffic. In the next 10 years, we are likely to see more client-generated media, such as video surveillance services, machine-to-machine communications and metadata about the services that we are using. Wu explains that as networks evolve and become no longer restricted to only humans, the nature of data and types of communication are going to reshape mobile computing.
Two Challenges of Next Generation Mobile Computing
Wu segments the challenges for the next generation of mobile computing into two areas. First, the complexity of developing modern systems-on-chip is dominated by software, validation, physical implementation and IP qualification challenges. Second, future networks have new use cases such as roadside units and robots, and changing topologies such as drone clusters or autonomous vehicles.
“Tackling these challenges will require a systems view,” says Wu. “We need to holistically consider data, computing and communication in order to build scalable, secure and efficient networks.”
According to Wu, “security is key; one way to address it is to have an open network architecture and implementation.” Wu proposes multiple ways to impose security at the application, network and platform layers in next-generation networks.
First, we must implement protocols to locally share, process and distribute data between devices and the network. Second, we can implement blockchain in the radio-access and core networks to ensure trusted data sharing across security domains.
See also How to Design Ethical, Resilient and Secure IoT Software, According to Google’s Vint Cerf
What is the Future of Next Generation Networks?
Wu believes that future networks must also be designed with AI in mind. To do so, “we need to create new fundamentals so that multiple AI systems can work together,” explains Wu. “Machine learning is moving from offline, to near real-time, to dynamic. We can leverage AI to design systems with many dimensions, however, AI data can be misleading; how do we determine which data is meaningful, and remove the data that might be downgrading performance?”
When asked how the end of Moore’s Law will impact next-generation networks, Wu pointed out that 3D packaging technology and silicon photonics will help to overcome semiconductor scaling challenges in the near term.
In the long term, Wu predicts that breakthroughs in materials and computing algorithms will continue to drive innovation. Wu notes that this will require collaboration across universities, governments and industries, as well as research in new computing models such as quantum computing.
About our author
Mario Milicevic is an IEEE Member and Staff Communication Systems Engineer at MaxLinear.