Notes IoT-provider: nine problems of the Internet of Things, or why it is not good in Russia
Hello, dear lovers of the Internet of Things!
This article is different from my previous ones. This is not about solutions and cases. I wrote about nine IoT problems that spoil our lives.
I propose to join my reflections and together predict the future of the Internet of Things.
Progress is leaps and bounds. We pay a high price for speed, especially for mistakes.
For example, programmers write code. Based on this code, a library or firmware is created that adapts to different tasks. The same code is used in different projects every time. As a result, there are "tails" - the remnants of the code that migrate from the device to the device.
It is tolerable if these are working “tails”. And very bad if there is a security hole.
Example : launch from the Vostochny launch site of the Soyuz-2.1b launch vehicle. Then the accelerating unit "Frigate" behaved uncommonly. In his program of 20 years ago, all launches were tied to Baikonur. He did not expect the launch from Vostochny. So there was an incident.
I dare to suggest that these very “tails of 20 years of prescription” will increasingly spoil life. Because they get into new programs. And this is a problem.
Lack of qualifications is a common thing. But with the advent of the Internet, it became a problem. The engineer stopped thinking with his head, he is looking for a ready-made solution on Google. As a result, we get a superficial approach and errors.
Example : some “builders” of LoRa-networks have been reading booklets and expecting a distance of tens of kilometers from the base station with a dead antenna, which was installed in the apartment at Kulibin.
This is good if it just did not work. And if it failed, when it started to exploit?
Marketing loves to interfere where they don’t ask. On the one hand, this is understandable. The product needs to be sold. On the other hand, borders must be kept. Marketing has nothing to do in the technical documentation. Otherwise, it turns into an advertising booklet (read - useless thing).
We have a rule in the Industrial Internet Center: do not use what is not tested for performance.
Through us passed mountains of consumer goods, sometimes even with nameplates of serious companies that even remotely do not show passport characteristics.
Example : the wording “before” in LoRa descriptions. For example: "our equipment reaches speeds of up to 50 kilobits / s and runs up to 15 kilometers." Of course, no one mentions that these are different “before” and at the same time they do not work. Either speed or range.
I wrote about it in detail here - Propriety . I will be brief here.
Many dream of creating a sought-after product that only they can sell. This is how proprietary, incompatible standards and devices are born.
Even if the development is conducted by competent specialists and the product is not decorated with tinsel. Anyway, when working with a narrow team, you can miss something.
Open standards tend to gather around them a community of enthusiasts. They are under the microscope studying all the advantages and disadvantages, looking for holes and bugs. The result is an understanding of the project’s problems and their solution.
In the case of proprietary technology is brewed in a limited circle of people. And they will definitely miss something.
Example : NB-Fi standard that is not protected against replay attacks.
This problem is particularly acute in radio communications.
Most of our IoT devices are constrained by non-technical, but administrative restrictions. We cannot exceed the permissible radiation level, we cannot use some frequencies without a license.
IoT devices additionally suffer from a hazy regulatory framework. Not every provider is willing to invest in an IoT network, which requires partnership with a mobile four operator (NB-IoT solution) or has an unclear legal status.
Example 1 : frequency feedback from Yota in Kazan in 2010, when the 4g network pilot was already built.
Example 2 : the draft decision of the SCRF to use base stations LPWAN only of Russian origin. Thus, networks that have already been built up or networks may be outlawed during the construction process. After all, equipment has already been purchased for them.
In contrast to copy-paste engineers, we have a growing community of people who love to break something. Some of them do it selfishly. For example, hacked the payment system - stole money. Others just love to look for vulnerabilities.
There are more cases of attacks without a specific goal. It is terrible that the qualifications of these guys are above average. And because the laws of Murphy is now doubly relevant.
Example : the Mirai virus has infected tens of thousands of devices. Using the login-password brute force method, he got access to the devices using default or near-default settings. Next, infected devices were used for DDoS attacks.
Security requires additional resources and affects the price. But there are no options.
The problem of scalability is the most serious, in my opinion, for the modern IoT.
A technology is born that is rapidly gaining popularity. Technology has limitations that hinder its development. But due to simplicity, functionality or some other reasons, they start to do everything on it.
So we very quickly run into frames, which are extremely difficult to leave.
Example : the wholesale distribution of Wi-Fi 2.4 GHz routers. Yes, the “waffle” is easy to use, reliable and designed for simultaneous operation of several independent devices. But when 60 routers are cut in at the same time, we run into the capacity of the ether.
In the evening, the speed of subscribers decreases, because the frequencies are not rubber. The solution here is a radical departure at 5 GHz with a large capacity reserve and a shorter wavelength.
This problem is similar to the previous one, but it has other reasons.
Often, new technologies are forced to work with old standards or devices. As a result, the technical task of engineers contains strange conditions. Which nevertheless must be observed.
Example : the width of a digital television standard DVB-T2 in Russia. It is 8 MHz. Why not 5 or 10? Very simple. 8 MHz is the bandwidth of the analog channel of the SECAM standard, on which we sat the previous half century.
This solution was used to ensure the joint broadcasting of analogue and numbers without violating the logic of OIRT. However, this year the analog will be disconnected, and 8 MHz will remain.
One of the main, in my opinion, problems IoT.
Progress requires quick decisions. Reliability is checked by time. Today, progress leads, and raw technologies come into production. Some of them do not stand the test of time: they break, do not tolerate scaling, or are simply useless.
The market votes in cash. And it is not uncommon for anyone to use this technology, because it is inconvenient, unreliable or useless. The latter are particularly sinful startups that develop complex, interesting and useless solutions.
Example : we regularly come across abandoned dispatch projects.
One of the illustrative cases is a survey of meters over GSM. We deployed a pilot zone in several basements - everything worked. Broadcast solution on a hundred basements - got problems.
At first it turned out that the connection is not stable everywhere. Then he discovered that the battery lasts a maximum of a year, while the average period is 6 months. Service in the project is not laid. As a result, simply threw a hundred boxes with pulsars, batteries and antennas. It turned out that it is easier to take readings in the old manner, rather than maintain such a network.
In my opinion, these nine problems hold back the development of IoT and spoil our life.
In my opinion, you can not throw in a bunch of technology with an excellent profile. For example, Z-Wave and LoRaWAN are initially different tasks, and therefore it is incorrect to compare them.
Attempting to drag through LoRaWAN RS-485 in transparent mode or build long-range solutions on Z-Wave is an example of using tools for inappropriate purposes. You do not blame the pliers that they are inconvenient to tighten the screws. Although it is possible, it is inconvenient and extremely strange.
I would appreciate feedback in the comments. What agree with what - no. Perhaps I missed something in my review.
Archive of previous articles:
#one. Introduction → # 2. Coverage → # 3. Zoo of metering devices → # 4. Propriety → # 5. Activation and security in LoraWAN → # 6. LoRaWAN and RS-485 → # 7. Devices and outbid → # 8. A little about the frequency → # 9. Case: we are making a LoRa network for TRK in Chelyabinsk → # 10. How to create a LoRa network in a city without a network in one day? → # 11. Notes of the IoT provider: may there be light, or the history of the first government order for LoRa
This article is different from my previous ones. This is not about solutions and cases. I wrote about nine IoT problems that spoil our lives.
I propose to join my reflections and together predict the future of the Internet of Things.
Old tails
Progress is leaps and bounds. We pay a high price for speed, especially for mistakes.
For example, programmers write code. Based on this code, a library or firmware is created that adapts to different tasks. The same code is used in different projects every time. As a result, there are "tails" - the remnants of the code that migrate from the device to the device.
It is tolerable if these are working “tails”. And very bad if there is a security hole.
Example : launch from the Vostochny launch site of the Soyuz-2.1b launch vehicle. Then the accelerating unit "Frigate" behaved uncommonly. In his program of 20 years ago, all launches were tied to Baikonur. He did not expect the launch from Vostochny. So there was an incident.
I dare to suggest that these very “tails of 20 years of prescription” will increasingly spoil life. Because they get into new programs. And this is a problem.
Copy-paste
Lack of qualifications is a common thing. But with the advent of the Internet, it became a problem. The engineer stopped thinking with his head, he is looking for a ready-made solution on Google. As a result, we get a superficial approach and errors.
Example : some “builders” of LoRa-networks have been reading booklets and expecting a distance of tens of kilometers from the base station with a dead antenna, which was installed in the apartment at Kulibin.
This is good if it just did not work. And if it failed, when it started to exploit?
Aggressive marketing
Marketing loves to interfere where they don’t ask. On the one hand, this is understandable. The product needs to be sold. On the other hand, borders must be kept. Marketing has nothing to do in the technical documentation. Otherwise, it turns into an advertising booklet (read - useless thing).
We have a rule in the Industrial Internet Center: do not use what is not tested for performance.
Through us passed mountains of consumer goods, sometimes even with nameplates of serious companies that even remotely do not show passport characteristics.
Example : the wording “before” in LoRa descriptions. For example: "our equipment reaches speeds of up to 50 kilobits / s and runs up to 15 kilometers." Of course, no one mentions that these are different “before” and at the same time they do not work. Either speed or range.
Propriety
I wrote about it in detail here - Propriety . I will be brief here.
Many dream of creating a sought-after product that only they can sell. This is how proprietary, incompatible standards and devices are born.
Even if the development is conducted by competent specialists and the product is not decorated with tinsel. Anyway, when working with a narrow team, you can miss something.
Open standards tend to gather around them a community of enthusiasts. They are under the microscope studying all the advantages and disadvantages, looking for holes and bugs. The result is an understanding of the project’s problems and their solution.
In the case of proprietary technology is brewed in a limited circle of people. And they will definitely miss something.
Example : NB-Fi standard that is not protected against replay attacks.
Administrative barriers
This problem is particularly acute in radio communications.
Most of our IoT devices are constrained by non-technical, but administrative restrictions. We cannot exceed the permissible radiation level, we cannot use some frequencies without a license.
IoT devices additionally suffer from a hazy regulatory framework. Not every provider is willing to invest in an IoT network, which requires partnership with a mobile four operator (NB-IoT solution) or has an unclear legal status.
Example 1 : frequency feedback from Yota in Kazan in 2010, when the 4g network pilot was already built.
Example 2 : the draft decision of the SCRF to use base stations LPWAN only of Russian origin. Thus, networks that have already been built up or networks may be outlawed during the construction process. After all, equipment has already been purchased for them.
Qualified villains
In contrast to copy-paste engineers, we have a growing community of people who love to break something. Some of them do it selfishly. For example, hacked the payment system - stole money. Others just love to look for vulnerabilities.
There are more cases of attacks without a specific goal. It is terrible that the qualifications of these guys are above average. And because the laws of Murphy is now doubly relevant.
Example : the Mirai virus has infected tens of thousands of devices. Using the login-password brute force method, he got access to the devices using default or near-default settings. Next, infected devices were used for DDoS attacks.
Security requires additional resources and affects the price. But there are no options.
Technological stalemate
The problem of scalability is the most serious, in my opinion, for the modern IoT.
A technology is born that is rapidly gaining popularity. Technology has limitations that hinder its development. But due to simplicity, functionality or some other reasons, they start to do everything on it.
So we very quickly run into frames, which are extremely difficult to leave.
Example : the wholesale distribution of Wi-Fi 2.4 GHz routers. Yes, the “waffle” is easy to use, reliable and designed for simultaneous operation of several independent devices. But when 60 routers are cut in at the same time, we run into the capacity of the ether.
In the evening, the speed of subscribers decreases, because the frequencies are not rubber. The solution here is a radical departure at 5 GHz with a large capacity reserve and a shorter wavelength.
backward compatibility
This problem is similar to the previous one, but it has other reasons.
Often, new technologies are forced to work with old standards or devices. As a result, the technical task of engineers contains strange conditions. Which nevertheless must be observed.
Example : the width of a digital television standard DVB-T2 in Russia. It is 8 MHz. Why not 5 or 10? Very simple. 8 MHz is the bandwidth of the analog channel of the SECAM standard, on which we sat the previous half century.
This solution was used to ensure the joint broadcasting of analogue and numbers without violating the logic of OIRT. However, this year the analog will be disconnected, and 8 MHz will remain.
High expectations
One of the main, in my opinion, problems IoT.
Progress requires quick decisions. Reliability is checked by time. Today, progress leads, and raw technologies come into production. Some of them do not stand the test of time: they break, do not tolerate scaling, or are simply useless.
The market votes in cash. And it is not uncommon for anyone to use this technology, because it is inconvenient, unreliable or useless. The latter are particularly sinful startups that develop complex, interesting and useless solutions.
Example : we regularly come across abandoned dispatch projects.
One of the illustrative cases is a survey of meters over GSM. We deployed a pilot zone in several basements - everything worked. Broadcast solution on a hundred basements - got problems.
At first it turned out that the connection is not stable everywhere. Then he discovered that the battery lasts a maximum of a year, while the average period is 6 months. Service in the project is not laid. As a result, simply threw a hundred boxes with pulsars, batteries and antennas. It turned out that it is easier to take readings in the old manner, rather than maintain such a network.
In my opinion, these nine problems hold back the development of IoT and spoil our life.
Why don't I consider the battle of standards a problem?
In my opinion, you can not throw in a bunch of technology with an excellent profile. For example, Z-Wave and LoRaWAN are initially different tasks, and therefore it is incorrect to compare them.
Attempting to drag through LoRaWAN RS-485 in transparent mode or build long-range solutions on Z-Wave is an example of using tools for inappropriate purposes. You do not blame the pliers that they are inconvenient to tighten the screws. Although it is possible, it is inconvenient and extremely strange.
I would appreciate feedback in the comments. What agree with what - no. Perhaps I missed something in my review.
Archive of previous articles:
#one. Introduction → # 2. Coverage → # 3. Zoo of metering devices → # 4. Propriety → # 5. Activation and security in LoraWAN → # 6. LoRaWAN and RS-485 → # 7. Devices and outbid → # 8. A little about the frequency → # 9. Case: we are making a LoRa network for TRK in Chelyabinsk → # 10. How to create a LoRa network in a city without a network in one day? → # 11. Notes of the IoT provider: may there be light, or the history of the first government order for LoRa
Source: https://habr.com/ru/post/438924/