What is the danger of the Internet of things, and is it worth it to create at all?

How do we see the Internet of things now? A world in which devices and devices of all kinds and stripes collect information about the environment and exchange data with each other in order to increase their own efficiency and productivity, proactively inform about problems, optimize all kinds of processes and preserve people's health. Not peace, but idyll.

Well, the goal is wonderful and worthy of respect. The media has already begun to flicker victorious reports about the advent of the Internet of things as a fait accompli. But the concept of an information network pervading the wholethe world is not only ardent fans. The voices of numerous skeptics who express legitimate doubts cannot be dismissed. The most critical citizens are even convinced that the Internet of things is not just not there, but it will never be created.

Why is he so objectionable to them?

From the point of view of skeptics, most of the promoted concept of the Internet of things is made up of the good old technologies of automation and inter-machine communication (machine – to – machine, M2M), which have been widely used for ten years, if not more. And the remainder falls on proprietary home automation technologies that solve trivial tasks and have a bunch of security holes.

Do we need to connect everything with everything?

If the vast majority of objects in an apartment or house will be included in a single network, then how to solve the almost inevitable problems of mutual influence? After all, the essence of the Internet of things lies precisely in the fact that each component performs its functions depending on the data received from other network participants. Imagine that your door lock mistakenly decided that you were not at home, and forcibly turned off the entire house, with the exception of, for example, a refrigerator. Or vice versa - would you activate the preset programs for different devices deciding that you are at home, although this is not so? You can come up with any scenarios when certain devices and devices will not work as intended, or will not be at all. Finding the root causes will turn into an incredibly complex task, given the sheer number of dependencies and relationships. And half the trouble if some kind of failure only gives you an annoying inconvenience, and if there is a threat to health or even life? Say, the light in the bathroom turned off due to a glitch on the home server, and your grandmother slipped in the dark?

And now let's remember the most important thing - the deliberate hacking. Every day, a huge number of computers around the world get infected with all kinds of trojans, backdoors and other rubbish. The use of botnets has long turned into a full-fledged shadow business with a large turnover. And what, you will feel calm, even if the doorknob in your home acquires a network module and sensors, and household appliances - processors and RAM? Yes, we are sure to be swept by a wave of all kinds of infections and failures, when attackers can gain access to our local home “clothing” networks. It’s one thing when you catch a cryptor on your home computer, on which there is nothing particularly important, and it’s quite another when someone can control the mass of devices in your home — a heater, a refrigerator, a stove, a kettle,physically . Oh yes, after all, in a few years, unmanned vehicles promise us.

Agree, an exciting perspective looms.

You now probably want to say about encryption algorithms and other firewalls. But do you think that many ordinary people will - and will be able - to do this? Yes, at least change the master password?

Yesterday's technology

Today, what we are given for the upcoming Internet of things is a bunch of disparate proprietary systems that massively use cheap sensors, cheap computing modules, and even cheaper communication modules.

A similar situation did not exist from scratch. Just the vast majority of the proposed scenarios using the Internet of things do not promise anything new. Even such seemingly modern and innovative ideas as managing building heating and ventilation systems, diagnostics of aircraft engines, choosing a plowing scheme depending on the weather and soil quality, monitoring the state of perishable products during shipment, determining the pre-emergency state of industrial equipment, managing production lines on based on data on the supply of raw materials and components - all this was proposed ten years ago. Just then it was not called "tasks for the Internet of things."

The difference between “then” and “today” is only that all these and many other ideas can now be implemented much cheaper and adapted for a wider range of areas. Now we have cellular communications, local wireless networks, cheap cloud services for processing and storage of transmitted data. Each of us also has a client device in our pocket, a smartphone or tablet, which can be used to manage “clothing” networks.

And although standard APIs and protocols have not yet been created for the Internet of things, there are already the appearance of peculiar centers of attraction: Apple HomeKit , consortia AllSeen , Open Interconnect , and, possibly, Google Threads and Google Works with the Nest protocol.

Specialized Networks

Perhaps you should not come up with new names for old technologies and ideas, but continue to do what we already do - to create a large number of separate specialized networks from specialized components? In them, at least it is possible to optimize various processes and technology stacks, to ensure security. It is extremely unlikely that it will be possible to create a universal stack of technologies; nothing good will surely come of these attempts.

Yes, in the dwellings it makes sense to connect some devices into the network: thermostat, stove, fire detectors, door locks, surveillance cameras, alarms; or connect televisions, game consoles, computers to each other. Of these clusters can benefit. But why connect heterogeneous clusters into a network? In an extreme case, individual devices can be used to control each of the clusters - a smartphone, tablet, computer.
For commercial and industrial applications, the independence of heterogeneous network clusters is especially important. Otherwise, you may find yourself in a situation, for example, when through a Wi-Fi router in an airplane it becomes possible to penetrate the on-board network and cause trouble. This is possible only in one case - if different clusters are combined in a single network: aircraft avionics and wireless communications infrastructure for passengers. A similar short-sighted solution, as it turned out, was used in the Boeing 737 MAX . In fact, an approach is introduced here, which is characteristic just for the Internet of things.



The vulnerability of generalized networks is already too high. And the best option would be to continue to use the approach with separate specialized networks, and not try to harness a horse, a trembling doe, a swan, a crab, a pike and further down the list of the zoo.

In some cases, it is enough to limit ourselves to dividing into virtual local area networks. In some situations, it is advisable to separate the network physically, it all depends on the specific situation. It happens that it is best to use even different network technologies. Say, some critical components (for example, smoke detectors) are better connected with each other through a separate network channel, and not through a common Wi-Fi. Otherwise, if the router does not work, then you will be left without a fire alarm. Such a deep separation of networks, again, is more appropriate to apply in industry or commercial organizations.

By the way, in addition to the "iron" component of the Internet of things, it is necessary to mention the software one. After all, the exchange of data between different devices should be provided using the appropriate APIs that support all possible data types and thousands of types of devices. For example, your music player will need to be able to exchange data with a microwave. Well, what can they say to each other so useful? Why network them together? And how difficult will the development and support of this universal interface be able to work in any devices and devices?

Considering all the above arguments, the skeptical point of view is as follows: probably, under the “Internet of things” we will still get not some common open network, but numerous proprietary clusters promoted and supported by different manufacturers. Surely the struggle for the client will be expressed in the fact that each manufacturer will try to “hook” users on their own system. Perhaps someday there will be compelling reasons and reliable technologies for switching to a common network, but so far this does not seem appropriate.

What is your opinion?