5G, edge computing will be the bedrock of the next era of human AI partnerships
If Moore’s law continues to remain applicable for next 10-20 years, and communication networks keep evolving from 4G to 5G to 6G and so on, that time is not far when we would measure the world’s intelligent population as humans and ‘AI-ans’ together, and human and machine partnerships would be rechristened ‘humans AI-ans’ partnership.
Technology is transforming the way we live and work and is changing at an exponential pace. It’s normal to see multiple technology generations in a single human generation. Gone are the days when an industrial revolution would occur once in a lifetime!
There are four primary drivers influencing this unprecedented pace of technology transformation. Smart/intelligent software; humongous data getting generated every second; rapid technological advancements in computing power driven by Moore’s law; and the leapfrogging of speeds and bandwidth in communication technologies from 3G to 4G, and now to 5G.
All these technological advancements are happening simultaneously and in tandem with each other, enabling frenzied technological advancements in all spheres of our lives.
As the world becomes more intelligent and software centric and data becomes the new oil, buzzwords such as AI, ML, AR, VR and blockchain have become a reality, but surely are still in their infancy stage relative to their true potential. We have not even scratched the surface here. Such smart software is driving new and innovative use cases in manufacturing, healthcare, robotics, transportation, fintech, telecommunication, etc. There is hardly any domain that has remained untouched.
Further, there is essentially a deluge of data in today’s world -- 175 zettabytes, 175×1021 or 175 billion 1TB drives. To put it another way, more than 13 Empire State buildings. For the non-engineers out there, that is the amount of data IDC forecasts will exist by 2025.
But the key here is to have multi stage data refining/analysis processes analogous to multi-stage oil refining. We need to separate data which is valuable, which can help gain insights through analytics, which can help train algorithms, and which can eventually make business or social sense; from the humongous amount of crude data getting generated every second.
This data refining happens in a multi-layer network architecture: -- at the sensor, at the edge, at the core and at the cloud with different outcomes and use cases at different stages.
Also it is imperative that this data analysis/ filtering happens at different levels as different use cases require different analysis and response time. An autonomous car needs to take decisions in real time and can’t rely on analysis coming from core or cloud data centers. It would be too late. But, to reveal deeper insights, big data analytics need to happen at the core or the cloud, and that requires large amounts of data from multiple sources.
As we move towards faster adoption of IoT and 5G, embedded and edge computing becomes a critical piece of this jigsaw puzzle.
This first stage of data refining at the edge is the most important one as that is where the real time decisions are going to be taken. With 5G providing seamless connectivity between sensors and edge computing, more powerful data analytics is going to happen real-time at the edge itself and the edge will become the new core in the data network.
5G is going to be a game changer in edge computing as the timely, low latency, data availability is not a challenge now, and hence powerful computing is essentially a necessity at the edge. But to harness the full potential of edge computing, both software and hardware needs to tango as embedded computing to get the best performance.
The way high speed data is unleashed in 5G is through 5 key concepts: millimeter wave, small cell, massive MiMo, beam forming, and full duplex. Powerful high performance edge computing is going to be unleashed by smart software; compact, miniature and rugged form factor based hardware, massive data MiMo (multiple data input, multiple data output) and edge analytics. Essentially 5G and edge computing are going to make edge the new core; core the new cloud and cloud becomes the moon.
Soon, just as NFV architecture is redefining the telecom network, embedded computing is going to redefine edge computing going forward in data networks.
Further, 5G and edge computing together are going to redefine the way humans and machines consume the data and are going to greatly accelerate IoT adoption and make AI/ML enabled machines ubiquitous as human beings. 5G and edge computing are going to be the bedrock of the next era of human AI partnership.
The way cloud computing essentially enabled centralized infinite computing at the large cloud service providers, 5G is going to enable infinite distributed edge computing where billions of extremely powerful edge computing nodes, riding on Moore’s law, would work together to provide virtually limitless computing at the edge.
We are at the cusp of the next technological revolution and 5G and powerful edge computing will undoubtedly be the tipping points.
What a time to be alive!
Rahul Gupta is general manager and regional sales director, OEM, embedded and edge computing, at Dell Technologies India. The views expressed in this article are his own.