Archive for month: April, 2018

Podcast: Play in new window | Download

Subscribe: Apple Podcasts | RSS

The number of internet connected devices continues to grow daily. But managing that huge increase in information is no easy task. That’s why I wanted to invite Mrinal Wadhwa, the CTO of Fybr, to join me on the show this week.

Fybr has been developing a number of IoT (Internet of Things) devices that are usually wireless low power units that can be installed in remote or hard to access places. One of the first devices they created was a hockey puck sized device that can be installed in individual parking spots. That device then connects to a central server to show if a parking spot is occupied or not. The devices don’t communicate with one another in the physical world because that would take too much power. Instead Fybr has developed the idea of a “digital twin”. The twin of the device exists on the server side and that allows all of the information collected by the individual devices to be shared across the network.

Developing those digital twins allows the user to create “business rules” for each device. You can tell the device when to go to sleep and when to wake up. You can also change the amount charged for a parking space during different times of the day. Whether those rules live on the server or on the device itself is decided based on the application.

One of the biggest issues in IoT development is over the air updates. A new business rule has been created and the company wants to push that code to the device. There is a fundamental problem, though. If you are constantly updating devices with new code and you make an error you run the risk of bricking the devices. Fybr has solved this problem by creating a tiny virtual machine that lives inside FybrLyn’s. Because it’s a virtual machine that is independent of the actual hardware code it functions as a sandbox. You can push business logic to the edge on a regular basis without fear of harming communication with the remote device.

Fybr’s system specializes in low power wireless devices but that’s not the only kind of IoT device. A customer may have powered devices or large manufacturing machines as well. The SAP Leonardo IoT is very well positioned to bring all that data into one place. The IoT created some required steps that should be performed. This could be to send a service technician to perform work, this would require a work order and here SAP would be integrated and handle those actions. Or it could be the new sales order to a printer ink that needs to be purchased and sent.

As IoT continues to expand there are huge hurdles to overcome. If Fybr’s parking sensors were deployed in every parking spot in America it would require hundreds of millions of devices. The scale of that information will be incredibly challenging to manage. There are also security challenges that have to be solved. A connected machine in a manufacturing facility is very well guarded. It’s very difficult to access that machine.

But a smart city sensors are installed on streets. Someone could spend hours trying to access the sensor without anyone knowing. The sensors will have to include some kind of encryption key. Anyone could steal the device and figure out the key. If a device cannot keep a secret than how can you trust if the device is providing accurate information? That is a huge challenge that is hard to solve for many smart infrastructure applications. Fybr has tried to solve it by carefully monitoring the life cycle of each key in each device.

Mrinal says IoT is a big challenging world. There are a number of unresolved problems. And there has not yet been very large scale deployments. There is tremendous possibilities in this space by combining traditional SAP knowledge with this new class of devices.