Welcome to the world of Infosys Engineering! It is a half a billion plus organization that takes pride in shaping our engineering aspirations and dreams and bringing them to fruition. We provide engineering services and solutions across the lifecycle of our clients’ offerings, ranging from product ideation to realization and sustenance, that caters to a cross-section of industries - aerospace, automotive, medical devices, retail, telecommunications, hi tech, financial services, energy and utilities just to name a few major ones.

« AR-VR at the intersection of SDN-NFV, 5G and Mobile Edge Computing - a practitioner's perspective | Main | The importance of engineering cloud-native and service based architecture layers for 5G Core Network Functions »

IOT Edge of Tomorrow

The Edge of Tomorrow is a 2014 sci-fi movie starring Tom Cruise that's based on a time-loop and you can read reviews here. However, this blog is about the IOT edge of tomorrow which is as exciting about edge capabilities that are ever expanding. Before getting into a prediction mode, let us recap about what drives the edge in the IOT world - some of the factors are low latency, local computing and real time decision making, lack of reliable connectivity and proprietary protocol protection. I have elaborated on these aspects in my previous blog here. Industry experts, analysts, and my own experience the edge is here to stay and if anything will continue expand. 

One way to peek into the future is to analyze it through various lenses that provide us different perspectives. A synthesis of these images can then provide a "tomorrow' view of what to expect. I chose to analyze the IOT edge from 5 different lenses - market needs, mobile network operator, hardware vendors, IOT cloud providers and lastly my experience of having worked with clients across multiple industry verticals. Let us dissect each of these further
IOT Edge Of Tomorrow.png
McKinsey has done a market analysis based on real use cases that are being implemented across industries. Rewind back a couple of years, and most of these were concepts and so it is encouraging to see the market adoption of these applications. Broadly, the applications fall into these categories (they are not comprehensive but clearly indicate that edge compute touches all aspects of the business value chain)

  1. Enhancing human, machine productivity
  2. Improving operations across the entire value chain
  3. Improving worker and workplace safety
  4. Proactive, predictive and condition based maintenance
  5. Improving security through smart surveillance
  6. Increasing resource management efficiency (especially natural resources)
  7. Optimizing inventory
  8. Enhancing customer experience
  9. Improving targeted marketing
  10. Usage based feedback into product engineering

The second lens is the Mobile Network Operator - Verizon recently tested edge computing in its 5G network and demonstrated latency reduction of upto half through its Mobile Edge Compute equipment. The particular application was using AI enable facial recognition and the images were analyzed right at the edge rather than hopping into the data center. This is a significant milestone in the broader Intelligent Edge Network initiative, of course there are business benefits that Verizon expects to realize from the initiative

The hardware vendors are the third perspective and there are many of them which offer gateways. The vendors also provide software stacks that run on their hardware. In few cases it is an integrated solution of both hardware and software stacks. Virtualization and containerization technologies are extending into the device world, these in conjunction with the platform and cloud providers to provide an end-to-end solution stack. Intel has released Neural Compute Stick (NCS 2) to enable AI and vision applications at the edge. It is the size of a USB drive that can be plugged into Raspberry PI or Linux PC and complex neural network based algorithms can be offloaded to this chip

The fourth lens is the cloud and platform providers. The big 3 (AWS, Microsoft and Google) have taken massive steps to provide a complete solution of building, deploying and managing devices through their solutions. In addition, the cloud providers are enabling AI, ML and cognitive services to run at the edge, combined with workload simulations for continuous integration and continuous deployment. The platforms are opened up for marketplaces to facilitate open innovation and create a rich ecosystem for industry specific solutions.

Lastly, my personal experience of working across automotive, transportation and heavy equipment industries. Each of these industries are exploring and implementing edge use cases for various reasons such as real time decision making, reliable connectivity and on occasions aiding human decisions through intelligent recommendations. The availability of larger compute power, ability to partition to run multiple software stacks, ability to execute open source projects, remote management of devices have contributed to the significant increase in edge compute cases. It's a matter of time that AI/ML models run at the edge in scale will pick up in the enterprise use cases.

As I said earlier, when you synthesize these views from the 5 lenses of market needs, mobile network operator, hardware vendors, IOT cloud providers and my experience it is apparent that edge computing is here to stay and scale. Edge is an integral component of the overall IOT solution and specifics of a technology stack will depend on multiple factors. Architectural decisions have to be based on the present and future needs. Unlike other software technologies, there is an unique aspect of the device which needs to be considered - this is being alleviated through software management of these devices. It is likely that protocol and communication standardization will happen over period of time. Cross industry and technology patterns such as from the telecommunications industry and internet technologies will be adopted. One thing is for sure - Edge Compute will accelerate as IOT adoption scales.  

Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

Please key in the two words you see in the box to validate your identity as an authentic user and reduce spam.

Subscribe to this blog's feed

Follow us on

Blogger Profiles

Infosys on Twitter