According to the CEO of Cisco Systems, soon there will be one million new connections every hour. So unless something drastic happens, we will live in a world where almost everything will be connected to a network. Gadgets ranging from an Apple Watch to a toothbrush will be generating data and communicating with a server.
This unprecedented growth of IoT connections, which is being caused by the incorporation of sensors and software in everyday tools, will eventually make people’s lives easier and prevent future problems.
However, the alarming growth of devices being connected to the internet is generating certain problems with the infrastructure on which information runs on.
There are two main problems that come with the boom of IoT, and these are scalability and security.
Scalability because the current way that devices communicate with servers isn’t optimal if there are too many devices communicating with the same server—specially if you require advanced functionality on those devices; and, security because the proposed solution for the scalability problem, can make any network vulnerable to malicious attacks.
Solving The Scalability Problem
Back in the day, when a device was connected to a network—let’s say a security camera, there was no “intelligence” incorporated into that device. Meaning that the camera could basically be either on or off, and, while it was on, it would be sending information to the server the entire time.
If you had a few cameras talking to a server this wouldn’t be an issue because the server could process all of the information. But, when you add hundreds or thousands of cameras that are doing the same thing, you start to have major issues even if you have a huge pipe.
The scalability problem is solved by Edge Computing. This is the idea that the “intelligence” of a network can be pushed to the “edge” of the network. Edge computing takes advantage of intelligent devices that are connected to a network by configuring them in such a way that they only send relevant information to the server.
In the security camera example, since old security cameras had no way to determine whether they were sending relevant information or not, they would simply send anything that was recorded as long as they were turned on.
However, with the advancement of IoT devices, these cameras got smarter, and now, using sensors and software incorporated in them, they are able to send, for example, only video where there is motion.
There are several benefits to Edge Computing. Among them, of course, is the capability of significantly reducing the load on the network devices and primary servers, allowing the network to operate more efficiently.
Another not-so-obvious benefit is the increased functionality of common devices. Now that technology has made simple everyday devices “intelligent,” these devices can be configured to perform advanced tasks and make decisions “on the spot” without having to completely rely on centralized servers.
Edge Computing can solve the scalability issue, and those who take advantage of this trend will become very valuable as IoT connections grow worldwide. However, solving the scalability issue is only half the battle, which leads us to our next topic:
Solving The Security Issue
Everything is butterflies and roses until you realize that with every IoT connection you are adding one more “hackable” device to your network.
Consider the fact that (probably) it is never going to be financially feasible to build robust security into every single IoT endpoint and you’ll realize that, at least for now, true security for IoT networks will have to come from somewhere else in the stack—that place is the network itself.
The main challenge with IoT security is financial resources. Since it is not viable to build the same level of security into a sensor that helps open a door than into an endpoint where sensitive data is being stored, many devices will remain vulnerable to attacks.
With this in mind, visibility into the network becomes crucial. However, even being able to look at the network holistically, knowing whether a device is distributing malware or stealing credentials could be very challenging if you have thousands and thousands of devices.
Add to this the fact that most traffic is now encrypted, which makes securing the network even more difficult. According to Cisco, over 80% of the world’s traffic will be encrypted in just a couple of years. You’d think that this increases security levels. However, Cisco also estimates that 70% of attacks will use encryption as well.
So, we find ourselves at a point where too many devices are vulnerable to attacks that can, and probably will be, camouflaged as encrypted traffic. Hence, detecting which devices have been compromised will be a very challenging task—especially since dealing with software updates in so many devices could easily compromise security as well.
The solution to this is something called Encrypted Traffic Analytics (ETA). ETA allows a network to analyze traffic metadata and, with a very high degree of certainty, determine whether traffic is malicious or not.
ETA uses machine learning to build a history of how encrypted benign traffic looks and behaves. This cutting edge technology allows the network to develop a frame of reference that is later used to infer the probability of traffic being “infected” by comparing it to previous data sets.
ETA, of course, threads the needle between privacy and security because the technology could be used for many other things other than detecting malicious traffic. But, perhaps true privacy is something that will inevitably have to go away as IoT connections grow all over the world.
Here’s a great explanation of ETA from Cisco Systems Sr. Technical Security Lead:
The Bottom Line
IoT connections will only continue to grow over the next few years. With technology becoming cheaper and, as a result, more accessible to everyone in the world, it will be just a matter of time before virtually everything is connected to the internet.
This will prove to be beneficial for the world. Imagine how many tragedies could be prevented or how much easier life could be for children and elders. However, in order for IoT to truly takeover, networking technology must advance to the point where this outrageous IoT growth is even possible.
Edge Computing proposes a viable solution to the scalability problem. With the implementation of Edge Computing we will have access to data that has never been leveraged before, which would lead to the simplification of human life.
But this will come at a cost—and that cost could be privacy. With so many intelligent devices that are vulnerable to encrypted attacks, it would be up to solutions such as ETA to help prevent the misuse of critical information.
Unless you’re the NSA, there is really no easy way to know whether a device in a network of hundreds of thousands of devices is distributing malware, or if due to a recent update a device has been made vulnerable to malicious attacks. It is now up to the network itself to prevent these attacks by using cutting edge technology such as ETA.
Cisco Systems is tackling this issue directly. They have announced their Network Intuitive, which uses ETA and Predictive Maintenance to “flag” potential malicious traffic and fix loopholes before they are taken advantage of by hackers.
But even with Cisco’s efforts, the imminent issues we are faced with due to exponential IoT growth are far from being solved. Hence, any IT professional who stays on top of the evolution of IoT, Edge Computing or ETA will become an invaluable asset to any organization.