Thank you for joining this Global CTO Roundtable on Edge, hosted by the U.S. Department of State, on the last workday of this busy month.
Edge refers to a collection of related topics, such as edge computing, edge networking, and edge AI. The confluence of three forces have turned edge from an interesting exploration a decade ago to an essential pillar of innovation today:
- 5G promises not only higher speed but also lower latency and jitter and much more massive scale as the foundation for its unique value proposition.
- Control and computation, from autonomy to distributed machine learning, are moving closer to the end users, thus offering the benefits of proximity.
- Data collection and cyber-physical interactions in industrial Internet of Things are becoming mission-critical for some vertical application domains.
Sometimes edge operates as part of an end-to-end network, either the old style since decades ago or through a continuum from cloud to “fog.” Sometimes it operates as a self-contained network on its own among connected and autonomous things. Just like open network protocol stacks reimagine the “vertical” decomposition of functional modules, edge enables new architectures on the “horizontal” decomposition physically across the network.
There are obvious value propositions that edge can provide, including the following five that compose an acronym of SCALE:
- Scalability of services and thinning of data transmission
- Cognition of end user needs and the ambience
- Agility of development cycles
- Latency reduction to enable feedback control and real-time learning
- Efficiency through device-to-device or device-for-device communications.
There is also no shortage of challenges. For example, edge as a computing substrate, in contrast to cloud, has to tackle heterogeneity of nodes, variability of links, and various types of resource constraints relative to the tight performance goals. It is not just a smaller and complementary version of cloud, but also opens up the opportunities of new map-reduce, new machine learning algorithms, new and specific use cases. Securing the edge and ensuring trusted apps are increasingly important too.
As represented by the diverse group of 12 companies this afternoon, the development and deployment of edge is often inter-disciplinary. Its success needs advances in multiple sectors: from ultra low-power microelectronics to mobile operating systems and apps, from wireless communication networks to embedded artificial intelligence.
Innovation leadership is a key dimension of our global strategy on 5G. We believe that the technology industry in U.S. and partner countries has an important leadership in this area of innovations: as the pendulum swing from clicks back to bricks, we have an edge on edge.
We look forward to listening to industry leaders on the latest development, competing architectures, and accelerated deployment of edge technology.