Cisco’s Biren Gandhi hasn’t played around with drones as much as he’d like to, but it looks as though he’s going to have lots of chances to do so given the company’s growing interest in these high-flying Internet of Things devices.
Cisco increasingly has made its presence felt within the commercial drone/unmanned aerial vehicle (UAV)/unmanned aircraft system (UAS) community, with Gandhi and others speaking at and attending industry conferences such as InterDrone (see video below) and NASA events. The company is also working with startups, carriers and others making up the burgeoning commercial drone ecosystem, and of course has been pushing hard into the broader Internet of Everything, such as through its $1.4B purchase of Jasper Technologies. Gandhi, a distinguished engineer & strategist within Cisco’s Corporate Strategic Innovation Group who early in his career worked as an R&D engineer at the Indian Space Research Organization, has also blogged about Cisco’s beliefs about drones over the past year.
While Gandhi examines all sorts of emerging opportunities for Cisco, from deep learning to the blockchain, Network World homed in on Cisco’s interest in drones and the possible impact of commercial drones on the lives of enterprise IT professionals during our interview with him last week.
As Gandhi notes, there are many uncertainties in the commercial drone world as the Federal Aviation Administration clarifies its rules, though that’s not stopping investors from piling hundreds of millions of dollars into startups and companies like Cisco from putting resources and thought into the market (See also: "Commercial drones gaining altitude with enterprise IT vendors")
Commercial drone network infrastructure
The possible angles for Cisco are many, starting with network infrastructure that would support drone-to-drone, drone-to-the-ground and ground-to-the-cloud communications. “That’s the de facto play for Cisco and other infrastructure companies,” says Gandhi, who prior to working at Cisco served as a CTO at game company Zynga and as a senior architect at Facebook.
Drone operators currently have numerous communications options, from Wi-Fi to 3G and 4G/LTE to microwave to satellite. The big question Cisco could help operators answer is: How do I optimize a drone to use those technologies? Satellite communications, for example, might be a last resort because it can be so expensive, but it also might be the only thing available in very remote locations. 4G cellular networks might work fine in some locations for those on the ground, but if the antennas have been tilted downward to cater to those users, drone operators might find the network lacking for their unmanned aerial vehicles flying above the cell towers. Gandhi and colleague Nico Darrow, in fact, recently filed for a patent that addresses exactly these sorts of issues (“Dynamic Network Connectivity Using Unmanned Aerial Vehicles”).
“If Wi-Fi is available, possibly use that as a low cost, high bandwidth option. If it’s not available you could fall back on 3G/4G, then on to microwave or satellite,” Gandhi says. Another option might be to launch an ad hoc mesh Wi-Fi network for customized coverage, he says.
Commercial drone big data
And what about dealing with all of the data that camera- and sensor-equipped drones collect? Cisco has given that plenty of thought, too. The company has demonstrated use of technologies such as WebEx and Telepresence conferencing as well as collaboration offerings like Spark and Tropo to help organizations make better and faster use of drone-eye views and drone-collected big data. One example Gandhi mentioned would be using drones to monitor for fires across acres of open land and giving remote people the ability to essentially sound smoke alarms to alert local response teams.
The realities of bandwidth availability and cost also play into how organizations might deal with their drone-collected data.
“Many players believe that once you capture data from drones you can just shove it into the cloud and manage it from there,” Gandhi says. “Sometimes that works very well, but in many other cases, such as agriculture, it does not, so you have to do the processing of high volumes of data locally on the ground disconnected from the cloud (i.e. fog computing), then maybe send just the results to the cloud later.” This approach to data management and traffic can both save money and allow for faster data analysis, he says.
While much of the mainstream drone coverage has focused on hobbyists using camera-equipped drones for fun, but also causing trouble, Gandhi says it is the commercial use of drones that provides bigger data challenges. Rather than strapping more typical cameras to their drones, businesses are using hyperspectral and thermal cameras, LIDAR sensors, virtual reality systems and other gear that generates rich and bandwidth-hungry data that needs to make its way back securely. “The data is a real valuable commodity,” Gandhi says.
The reality of drones is that while they can automate otherwise dull, dirty and dangerous processes, supporting services are needed on the back end. “Once you have a fleet of drones, there has to be an orchestration system, a management system, a high bandwidth reliable traffic system, and on top of that, security and policy,” Gandhi says. “All those are enterprise IT attributes and you have to take care of them once you go into a scalable mode of operation.”