“In Conversation With” is a video series where NTT i3 executives engage in conversation with some of our most visionary partners at NTT’s Operating Companies and global enterprise customers, futurists, and leading Silicion Valley technologists, entrepreneurs, and researchers. The explorations may be around current and near future developments in IoT, big data analytics, machine learning, AI, network virtualization, and security. At times, we also investigate the human impact of technology and the role our values and culture have on how we decide to use what we are inventing.
In this episode, Ravi Srivatsav (CPO and CBDO at NTT i3) joins Shriranga Mulay (SVP at NetMagic Solutions, an NTT Comms company ) on a recent trip to India. The topic: How the convergence of IT, OT and IoT are changing datacenter operations and customer expectations.
So Mulay-san – What are the traditional areas of concern for companies that offer data centers as service? Are all data centers roughly the same? And if they are not, what are the areas in which you can actually differentiate yourself?
I think things have been changing over a period of time. Customer expectations are also changing. So it amounts to SLAs, the various types of services that are required to service the newer architecture of the applications. And that demands different things to be done at the data center level than the normal traditional co-lo data centers.
Apart from that, I think that the IT and OT convergence coming up, the changes are already evidently seen in the network connectivity that we talk about or could be the infrastructure requirements. What it used to be a few years back and what it is now – things have changed in multiple dimensions. Infra, network, and even the applications, within the entire infrastructure.
How do you as a data center prepare for these changes? Are you seeing these changes because of the emergence of IoT?
The convergence of OT and IT infrastructure coming together – The change that has actually come into play is that the boundaries of the datacenter have gone beyond their physical limits. Because now with OT and IT infrastructure coming in, there would be a need for edge computing to come into play somewhere – for local analysis and local controls to be put in at the remote end, and you would have a central infrastructure controlling the overall aspects and giving you an analytical view of what’s going on at the remote end.
The other thing that comes out of this is not just the plain vanilla underlay network. You need a strong overlay as well to connect all these end points to the edge and the edge to the center.
How might one differentiate and provide value to the customer in a post-IoT world from a data center perspective? What differentiates the old way and what is the new way of doing things?
Traditionally, the way data centers have been functioning is you designed the infrastructure to handle x amount of power, cooling, and workload that it would consume. There have been no controls there to monitor it and then put the feedback mechanism in place to control it to the optimal level
If I look at IoT, the way that it will help is to be able to monitor temperature, humidity and all the other parameters which finally attribute to the energy consumption of a data center – apart from the workload energy.
If you look at a data center, especially a public data center where you have customers bringing in their workloads – There are different types of workloads that people bring in where there is a chance of some high-density racks coming in. By monitoring these through various sensors at various locations in the data center, we can actually check out where the heat flow and heat spot is getting created. If we can control the air flow back into those areas properly, we can mitigate losses to equipment life. At the same time I give my better SLA to customers because obviously less losses in the infrastructure means less down time to the customer.
The number of IT devices are being out numbered by the IoT devices inside datacenters and in manufacturing environments. So what tools and technologies should a data center have to address these concerns? And should security be a concern?
The treatment to be given for managing and monitoring OT devices is completely different than how you treat and manage a normal IT infrastructure. And when you talk about a network connecting from one point to the other, security is always a concern. And I think it becomes more complex as you go with the individual IoT devices. There are so many endpoints that you are connecting.
With IoT devices there is heterogeneity in the hardware manufacturers as well as the software that enables them. What kinds of tools do we have to develop to aggregate information from multi-vendor IoT devices?
There are 2 angles to it. One there are manufacturers who have their own proprietary way of encoding the data and sending it across. Secondly there would be occasions at least in in the near future where people may want to continue with their legacy systems in the way they are currently collecting some of their data.
So yes, it’s a challenge – Too many devices with too many proprietary things going on.
Is there a need for you to have a combined dashboard that is AI driven?
So data once collected needs to be analyzed on a real time basis. For that I think we need to pass this data thru real time analytics platform. There are multiple tools that you need to put together as a platform which will give you a view of what it is right now and also you get a trending data of the historical data. Both of this could be used for a particular decision.
Certain things could be automated based on how you tune the system based on the real time data. But you also need some reference back to certain incidents that have happened in the past. So either the system takes a decision on its own or it aids a human who is monitoring and controlling the system to take a decision.
Next: Ravi continues the series and talks with Scott Gibson again. Or, click here to check out all of the videos in our “In Conversation With” series.