Q&A: Cisco’s Theresa Bui on the company’s Kinetic IoT platform

It’s been almost a year since Cisco announced Kinetic, a cloud-managed IoT platform aimed at capturing a large and profitable share of the rapidly growing business and industrial IoT market. The executive in charge of Kinetic, Theresa Bui, spoke to us about the platform and how it’s architected, in the wake of a flagship customer announcement – the Port of Rotterdam – and a limited partnership with IBM.

Here is an edited transcript of that conversation.

What’s a customer getting for their money when they buy Cisco Kinetic?

As a whole, the platform enables three core, functional capabilities. It allows you to easily and automatically extract data, and how we do that is we ship a library of automated connectors that help you extract data from various data pipes, put it into a model – whether it’s CoAP or MQTT or whatever the flavor that works for you.

The platform is a combination of cloud-based software and on-prem software with our router and gateway solutions. The second thing it allows you to do because of the bundling with our gateways, is the flexibility to do edge computing of your data, or as near to the edge as possible. A typical oil rig will have 30,000 sensors on it, talking back to the cloud so that you can do historical analysis and predictive analytics. But other kinds of data on the rig could be something like “hey, this particular component is over-rotating by 2,000 percent” You don’t want to wait and send that stream of data back to the cloud, have an alert sent back down to the oil right that says “slow this down.” So most companies want the flexibility of working in the cloud, at the edge or anywhere in between.

The last core capability of the platform is to move and categorize your data based on policy, and share your data based on policy. If I own a factory, and I have a robotic arm in my factory, that robotic arm is collecting a lot of data. I want to share that data back to the maker of that robotic arm, because I want them to be able to do remote monitoring and diagnostics. But here’s the thing I don’t want to do: I don’t want to send data back to that robotic arm manufacturer that could reveal potentially proprietary things about what I’m doing.

How do instructions move around the network, and what kind of network complexity can users expect from Kinetic?

When customers implement Kinetic, they will typically include, depending on how big the customer is; somewhere between five and upwards of a dozen gateways, they will hook up their devices to those gateways. Each gateway will manage a set number of devices. We have an agent on that gateway that allows us to push and pull data, and it’s the agent that has the out-of-the-box libraries for extraction, so we go in, and you tell us what the devices from which you want to get data, and we will install the agent with the right settings from our library. It’s on that gateway where we do the edge computing, if you want it, and it’s all managed and configurable via the cloud.

So implementing our gateways is cloud-based, we try to make it as close to zero-touch as possible, so that you don’t have IT people walking out onto the factory floor and installing equipment. It’s also in the cloud where we have a powerful visual data editor – that was designed very specifically for what we’re calling OT users (operational technology), so that they can visually see “these are the data streams.”

So nothing on a command line, nothing that requires you to have IT experience…

Exactly. And it’s literally drag-and-drop, meaning “here’s the data stream coming out, I want these two data streams to be coupled together, and then I want these two data streams to be sent someplace else. I want them sent back to my partner’s cloud, or I want them sent to my on-prem data center.”

How is the system monitored for faults, safety and reliability?

From a monitoring and management perspective, there’s secure management of the gateway, which is automated and cloud-based. But another way to think about it, when you have your device connected to one of our gateways, your device receives all the protection of that gateway. So the[re’s] data encryption, the ability to identify intrusion to these devices.

The last component in monitoring and management, if you’re using Kinetic to build and deploy specialized apps on the gateway, you can manage those apps in the cloud as well.

Cisco’s a fairly unitary company – they seem to like having all their features under their own roof. Were there any challenges in partnering up with IBM and Microsoft for this? (ed. – Microsoft’s Azure cloud is a  “preferred partner” for Kinetic.)

I would say that, in my experience, we have always partnered where we think it’s the best thing for us, and in this case, IBM Watson has created specialized analytics that are completely customized for a customer. Kinetic has its own visualization capability, but we actually, purposefully, when developing Kinetic, that there’s never going to be a standard visual. When we walk into different companies, everyone’s using a different BI or platform. It doesn’t make sense to ask a customer … to use a second dashboard for their IoT devices.