The panel comprised Matthew Eastwood, Senior Vice President, Enterprise Infrastructure, Cloud, Developers and Alliances, IDC; R Srikrishna, CEO & Executive Director, Hexaware; and Vishal Lall, COO, HPE Aruba. 

In the opening remarks delivered by Matthew, he admits that the digital transformation of organisations has been accelerated out of necessity fuelled by the pandemic. He believes that 75% of enterprises in industries such as manufacturing and mining will begin to adopt private 5G networks to achieve network reliability and coverage and to maintain data control and security. Given the constantly stretching and expanding IT infrastructure, he explains that scale will become a big consideration in a lot of organisations.

Furthering the discussion about digital transformation, Vishal acknowledges the role of technology in bringing about a change in the workplace. He points out how many of these changes were already underway well before the crisis period accelerated their deployment. As for the adoption of cloud, its scalability and elasticity are important factors, he states. Further, considerations such as customer experience, automation, or cost have been instrumental in driving AI automation. Speaking at length about the implications of remote working, he explains that edge processing is increasing because people are increasingly working in a distributed environment. And from the last 25-30 years cyber security, too, is gaining heightened attention as more and more people are working out of the corporate firewall, making them prone to cyber dangers. The concept of hybrid workplace is here to stay, with employees having the flexibility of working in the office or from remote locations. This new working model will require relentless support from tech leaders, however, in the form of technology and architecture tools. WAN (wide area network) is now as important as LAN (local area network), but it is different from the earlier WAN which connected retail or remote locations into a central office; “We have to think of every single employee’s location as a home branch,” he remarks. 

Srikrishna, aka Keech, highlights the role of cloud for network and security. Rendering the old VPN architectures as inefficient as they were built only to facilitate 5-10% employees to work from home, he propagates a cloud-based VPN for smooth functioning of businesses even on a distributed network. However, from the security standpoint, he believes that nobody is in the trusted zone – every device is untrusted. 

In an effort to bring about digital parity in the post-pandemic world, Matthew believes that we’ll be using the digital tools much differently than we did in the past. He cites his personal experience saying how it wasn’t always fair for the co-workers to attend meetings even as the bosses munched away on chips and sandwiches. Essentially, we’re heading towards a world where people will be working from multiple locations but even if we’re in the same building, these digital tools will be used to enhance the experience for everyone, in the awareness that we must create a level playing field.

Vishal shifts the focus on edge computing and how the architecture is continuing to evolve to support the processing of data at the edge. He highlights the four elements of edge as: connectivity, management, infrastructure and security. The underlying principle of the edge is that it is a very distributed environment, unlike a large dataset that consolidates the data in one place. The first element associated with the architecture is connectivity because you can’t do anything with an edge architecture unless you have robust connectivity. Think of connectivity as a multipolar concept – a combination of 5G, internet, Bluetooth, IoT and related technologies. Secondly, if you have a fleet of devices across the world, how do you manage it from a cloud environment where you can scale and manage these devices on an ongoing basis – that’s a consideration. Thirdly, the infrastructure must be autonomous so that even if connectivity goes down, it can continue to operate because edge processing is revenue generating. Think of a factory; if the assembly line goes down you’re directly impacting revenues. Think of a fast-food store as a micro factory and if connectivity to the cloud goes down you’re again impacting revenue. And these edge locations, infrastructures or platforms are very data heavy so elements of ML and AI come into play: the whole life cycle of acquisition of data, analytics associated with data and then acting on data is important. The fourth and final element, security, starts with visibility. With IoT devices, for instance, identifying the devices, understanding and categorising them, and then lastly protecting them is crucial to successful running of edge computing.

Panelists speaking at the 29th edition of NTLF, held virtually from 17-19 February, 2021.

Tying back to the core theme of the discussion, Keech explains that AI and cloud are, indeed, inseparable. If a group of data scientists is given a choice between very good datasets but moderate AI algorithms, or moderate data sets but very good AI algorithms, they will unequivocally pick the former, because how many organisations can afford too much data if not for the cloud? When the goal is to train algorithms on lots of data, it has to be based on the cloud. Computing power, data and smart algorithms: three of these naturally live in the cloud. Citing examples of AI deployments by Hexaware that have accelerated because of Covid, he mentions an HR chatbot that was deployed for 20,000 employees worldwide within a month in April 2020. It replaces the HR personnels’ role by producing a daily report of employees that are unhappy or at risk. Another chatbot, he says, assess and catches risks in employee mental health.

Speaking of AI from the product perspective, Vishal says that every product that they make and every process that they run at Aruba has some elements of AI involved in it. Al is also involved regardless of whether it’s a datacentre or edge deployment, partly because it enables the elimination of manual tasks, besides other reasons. Other reasons for applying AI-driven automation is to enable the resolution of network issues and the expansion of network capacity on the edge. These are examples of things that were done at Vishal’s company just by working off data collected from the millions of devices in the edge, hence helping to improve network performance and solve issues before they arise. He demonstrates: HPE InfoSight has been around for more than a decade but is learning and self-improving by collecting 90,000 datapoints every 4 seconds. Owing to the many petabytes of data that sit in the database, 90% issues can be solved even before they occur. “We’re using AI to improve customer experience and make these products autonomous so they can learn by themselves,” says Vishal. 

In the closing remarks, Matthew says that in this big IT ecosystem, half of the Global 500 companies are going to build SaaS apps. Therefore, AI and cloud will be at the centre of innovation in the coming times. 

Keech says that talent, which is in short supply, will make all of this innovation happen. Tech unemployment during Covid was only 2% compared to overall 15% in the United States. “Every customer needs to think of talent in a more sustainable way and that's what is going to ultimately help transform this potential to reality,” he concludes. 

Vishal concludes with the realisation that the world can be different and we can sustain differently. “A lot of the impact will be seen in these few areas,” he says of the merger of AI and cloud technologies.

Image from Pixy