In the next year, the Internet will connect to about 1.5 billion new “things,” and by 2020, over 20 billion devices will be online. In theory, every one of them will be able to connect to every other.
The Internet was not created to handle this population explosion of data-generating devices. The Internet will be threatened by its own growth, and we will see cyber-attacks extend to more and more of these things.
So is the Internet headed for collapse? No, but it will change. Here’s how:
Things themselves won’t break the Internet
The Internet wasn’t designed to handle data being generated by 20 billion devices, but soon it will have to. However, we’re not about to face the collapse of the Internet because of this.
Unlike general-purpose computers and smartphones, The majority of IoT devices and sensors are not sending and receiving large volumes of data. IoT devices are special purpose and normally send and receive only small bits of data to a single destination. The Internet can handle the data of billions of the new sensors as long as the sensors aren’t too chatty.
And while the original specifications for the Internet only allowed for about 4.3 billion devices, we’ve already re-architected it to handle a lot more: about 340 trillion trillion individual things. That should hold us for a while.
But while the fabric of the network itself may be robust, hackers can exploit its flexibility.
Security: Expect worse
IoT-based attacks, like the botnet that brought down Twitter and Netflix this year, are going to get more ambitious and more damaging. Attacks on infrastructure (power grids, traffic systems) or on IoT consumer devices themselves (appliances, for example) are already being attempted. Eventually, one will succeed at scale.
Fortunately, the Internet can, and will, be made safer. Security can be built into the network itself.
One way is to restrict the type of traffic that network equipment will accept from an IoT device. While personal computers and smartphones get mostly unrestricted access to the Internet and the Web, there’s no reason a webcam or a temperature sensor should have the same rights. For example, a thermostat shouldn’t be allowed to send millions of Web page requests a second.
Networking companies and standards bodies are working to let device manufacturers securely assert what their devices should be allowed to do on the network, and network equipment companies will encode the following of those rules into the fabric of the network. The Internet doesn’t have to be wide open to hacks from billions of things that have only very specific jobs to do.
The “edge” of the Internet will become a huge growth market
Some IoT devices do have the potential to swamp existing networks, especially if they’re deployed in large numbers. Video cameras send a lot of real-time rich data. New jet engines are laden with sensors (GE’s new PW1000G engines have over 5,000) and generate 10 gigabits per second when running, terabits per flight. Cars also are now recording massive amounts of information.
Devices that do record huge amounts of data will need new systems to connect to that can process their data.
If an engine that’s recording gigabits of data a second sees a fault or other issue that should be addressed in the moment, its computers should have the intelligence to communicate to the right people and machines, on the plane itself and via radio to the ground, immediately. But most of the data these sensors read is recorded, not transmitted. It is only transferred to a collection and analysis system when there’s a robust and private connection. The airplane engines can offload their data when they’re on the ground and connected to the airline’s own high-speed wireless network.
But if you’ve got hundreds/thousands/millions of planes (or cars, or power plants, or video systems) all uploading terabytes of data at the same time, it’s still too much data to dump onto the Internet to send to the cloud for storage and processing. What we need — and what’s getting built now — is intelligence at the edge of the network: between the devices collecting data and cloud data centers.
If there’s one part of the global network that needs rapid upgrades to serve this new wave of things, it’s the edge — the border between IoT devices and the computers on the Internet. The massive amount of datathese things generate needs to be processed, reduced, and analyzed before it hits the Internet. It’s a big opportunity.
Your toaster will get smarter. Maybe too smart.
Still have a simple toaster made up of a few heating elements and clockspring timer? Maybe save it for an eBay auction one day, because appliances and machines, from light bulbs to ovens to drill presses, are connecting to the Internet. And they’re getting really smart.
The cost of putting Internet connectivity into a device of any kind is dropping fast, and at the same time device manufacturers are fighting each other to put more and more features into their devices. So we’re getting “smart” devices that you can access via a Web site or mobile app.
The problem for ordinary humans is the experience of using all these interfaces. Nobody wants one app for their lights and another for their thermostat. Most people don’t want to use an app for those things at all.
Soon, these smart devices will get really chatty with you. They’ll communicate through voice (either with their own speakers and microphones or through something like Amazon’s Echo), or through chat in a text or Facebook message. One morning, you may wake up to an emoji from your toaster.
There’s an opportunity to mediate or aggregate what’s going to be a crowded cocktail party of gadgets all vying for your attention. Players in the baby “chatbot” industry will fight hard for your attention this year.
Car tech will fuel wireless upgrades
Who needs traffic lights when your car itself knows when it should stop and go? Audi is experimenting with cars that communicate directly with traffic lights in Las Vegas. That’s just the beginning. Within the next few years, there won’t be such a thing as an unconnected vehicle.
We all want safe and more efficient cars, and that requires that they are networked – to each other, to roadway and city infrastructure, and to data collection services in the cloud. Better automotive networking will also help push the rollout of self-driving technologies.
The current infrastructure doesn’t guarantee the fast, low-latency, and reliable connections that cars need. So car companies and wireless carriers are hustling to standardize new networks: Dedicated Short Range Communications (DSRC) and 5G cellular.
In 2017, the automotive industry will drive new networks.
Billions of new things are already straining the Internet and forcing the redesign of protocols. Inventions, economics, societal changes, and even creative criminality are also forcing changes to the Internet. To adapt to these changing pressures, we need to weave newer, stronger communications onto the existing Internet so it can be the foundation for billions of new devices and for businesses we haven’t yet imagined.
Rowan Trollope is SVP and GM of Cisco’s IoT and Applications business.