Networks Need an Athletic Brainiac
It used to take just a few seconds to roll your eyes back into your head and count how many things in our lives needed an Internet connection. I stopped counting when it went over 30, as I hadn’t even scratched the surface.
There are various names for this trend: “M2M” for “machine to machine” connectivity. The “IoT” — Internet of Things. The “IoE” — Internet of Everything. The physical “things” in our lives that crave Internet connections are generating trillions of network events per year. That’s trillions with a T.
Here’s what will happen: Our machines will start clogging up the Internet. This will happen for two reasons:
First, people are the primary users of the Internet — for now, but not for long.
By 2017, we forecast there will be more machines connecting to the Internet than smartphones, or tablets or PCs or TVs. Soon thereafter, we’ll invariably have more machines, in aggregate, dipping into the Internet than we have humans on earth.
Second, machines behave differently than people. They respond autonomically, much like our bodies. When I go for a run, I don’t tell my lungs to work harder. I don’t tell my heart rate to go up. I don’t tell my body to sweat. It just happens. When I cut my finger, I don’t tell white blood cells to go to that spot. It just happens.
Similarly, machines operate much faster — at “machine speed,” if you will, rather than “human speed” — meaning that they can create many different transactions or events in the same amount of time it would require us to do just one. And because so many of the machines will be interconnected, an event with one can cause a cascading series of events with another, thereby causing even more activity on the network.
As a result, networks that must deal with M2M, and all the many other services from mobility to networking, need to be autonomic and elastic to adapt and harness resources to accommodate these ever-changing dynamics.
The care and feeding of an Internet designed to work at the speed of people isn’t necessarily as complex, in terms of agility and processing, as when you’re connecting trillions of actively chatting devices, with their own ceaseless conversations and cadences.
All over the place, the machines that serve us will take dozens of little Internet sips, when they need to, whether we are awake or not.
For example, in an M2M world of the future, the alarm clock (which is to say, the mobile phone) will be able to tell the coffeemaker to not start yet. Snooze happened.
Many homes today are already increasingly able to execute a daily to-do list of automated jobs — securing windows, running webcams and much more. They’ll be talking to the network, all day long, in hundreds of little shotgun-y bursts.
And this trend will only increase because once you start putting mental cycles on what would make your physical world better, it’s hard to stop. This is true of anyone I’ve ever had this discussion with. You know: I’d like this (fill in the blank) better if it (fill in the blank).
Example: A connected bicycle helmet. Should you fall (again?!), it can sense the impact, and trigger a call to your phone. If you don’t respond, it can call 911 or a contact with your exact location.
Another example: Trap-mapping. A professional trapper up in Canada put M2M sensors on his traps. Trap gains an occupant. Ping! A digital flare, with location, so it can be located on an online map. This alone is a genius reincarnation of a 2,000-year-old practice.
But where these examples go next is to take them out of a discrete interaction between the machine and the Internet and into a broader set of interactions through a mashup of IP-side apps. Then the possibilities are practically endless.
Think about what happened with your mobile phone. In the beginning, it was a just a phone. Then a text machine. Now, with Internet connectivity to apps, your phone is a mobile computer that acts as your conduit to dozens — or hundreds — of other activities.
Likewise, in our earlier examples, the interaction can expand to include weather, traffic, locations of nearest food or supplies.
Eventually, the really useful M2M mashup makers will shake out, and somebody new will have a seat in the Internet economy, making the interactions we have on the network today seem as advanced as a dial-up modem seems now.
One thing is certain: M2M and the IoE will change the way networks treat traffic, not just because the transaction rates are going to be huge, but because they’ll require unprecedented levels of integration and orchestration between users, network and data centers.
This is why you hear so much about “software defined networks” (SDNs) and “network function virtualization” (NFV), by the way — because dealing with the IOE needs an athletic brainiac. It needs intelligence, and brawn.
And just as our bodies respond as one, all the responses of the network need to do the same, meaning that trends such as virtualization can’t reach their full potential until all the discrete components are orchestrated together. Or, put another way, just as the network will increasingly be virtualized, virtualization will also need to be networked.
M2M adds just one new dimension to the Web, and to the care and feeding of the Internet. It adds device intelligence, even for devices never built to do anything but what they do as physical machines. When those devices, services, networks and cloud can begin to work more together, it will deliver far more than any of the devices could do on their own. This is where the possibilities of the Internet of Everything really start to take off.
While humans are still the majority users of the Internet, the status quo is gear-built for brawn. Routers move data at predictably faster and faster speeds, and higher and higher densities. The trajectory is always about faster.
However, what networks need — in order to deal with (and monetize) the looming onslaught of M2M traffic, which is happening today and unfolding further every day — is intelligence.
Networks need an athletic brainiac. This mashup will make the infrastructure of the Internet truly programmable, which could spark an explosion of technological creativity that would not only lead to advances in our physical and digital lives, but would also act as a catalyst for the Internet of Everything economy.
Doug Webster is VP, service provider marketing, at Cisco. Reach him at his blog.