The use of public Internet channels for managing objects in the IoT system can lead to the emergence of security threats not only for this IoT system, but also it can provide cybercriminals with resources to carry out attacks on any other objects of the global network. Therefore, you should use secure data exchange technologies that prevent unauthorized entry into the system when building IoT systems. This technology is discussed in detail in this article. The purpose of this work is to improve safety of IoT systems through the use of a perfectly secure data exchange channel.
In 2020 the number of connected devices to the
IoT exceeded 30 billion, and their annual growth
increased from 3 billion in 2017 to 5 billion in
2020 as shown by the published data of
researchers [
Forecasts up to 2025 assume that this growth
will not decrease, but tends to increase. This
testifies to the rapidly growing need for managing
remote sites and ample opportunities for their
implementation using existing tools and
technologies. However, the rapid growth of needs
and the broad possibilities of implementing IoT in
a short time often leads to insufficiently thought
out solutions from the point of view of security,
which is described in [
However, it is difficult to develop uniform recommendations or standards due to the difference in security requirements depending on the area of use of the IoT. The variety of areas of use is shown in Table 1.
Gemalto's survey of IoT users found that 90% were unsure about security. Thus, it seems to be relevant the analysis of IoT systems from the point of view of ensuring the secure exchange of data over the Internet channels, as well as the technical solutions in this area, given in the work. 2. Analysis of data exchange options in IoT systems
To connect IoT devices to the Internet, one of two schemes can be used, shown in Fig. 1 and Fig. 2 respectively.
The scheme shown in Fig. 1 is the simplest one and can be successfully used in internal computer networks. But such solution has a number of disadvantages in the conditions of the public Internet:
- Connecting the server directly to the Internet facilitates the intervention of unpredictable external threats into management processes. - Cybercriminals are more likely to install their botnets (malware) on your server to implement DoS and DDoS attacks.
- This server requires a dedicated IP address on the Internet, which is associated with additional material costs.
- To ensure the information security of the server, qualified service is required.
Disadvantages listed above are absent in the circuit shown in Fig. 2, where data flows between the terminal and management objects are filtered by the proxy server. This server can simultaneously serve many users, protecting their data streams from malicious attacks. Internet service providers (ISPs) can install such servers, providing customers with cloud-based access to resources. However, the user can install own separate or corporate broker server in case of high security requirements for information about objects managed. In cases when the broker's server is hit by a threat, the information about the managed objects will be kept intact. An increase in signal latency should be noted as a disadvantage of control through an intermediary server in comparison with the first scheme. But this disadvantage can be considered insignificant, since the performance in control systems cannot be high due to the presence of unpredictable network access latency using Internet channels. 3. Technical solutions to secure the IoT
Object management via the Internet does not
require the transfer of large amounts of data and
high-speed messaging. This allows you to use the
most advanced methods of protecting data from
threats of disclosure or spoofing during
transmission over channels. The use of such
methods makes it possible to exclude the
possibility of these threats being realized, which
is mathematically provable. It should be noted that
no expensive technical solutions are needed for
absolute protection. This protection is
implemented using simple software methods. It is
mathematically proven that the absolute
protection of information is provided by the
Vernam cipher, which is called one-time pad [
Condition Generation of random bit sequences (not pseudo-random)
Condition fulfillment A method for random bits generating is implemented, which allows you to generate
random sequences on any
computer, as described in [
Each random bit For each communication sequence can be used session, random bit for encryption only sequences are generated once independently of each other
For the exchange of The exchange of random bit random bit sequences, sequences occurs according to an absolutely secure the Diffie-Hellman algorithm communication with such parameters for channel should be which there is no possibility of used data disclosure in modern conditions
The work [
The beginning of filling the array with N
random bits looks like this:
var N = [504]; // Array of 503
random bits (N [0] is not used)
var T1 = new Date (); // Take the
timestamp for transformations
var TN = T1.getTime (); // TN
the number of milliseconds from
01/01/1970
N [
The block for filling an array MA with powers
of A looks like this:
// Elements of arrays with index 0
are not used
var A = [504]; // Sequence of 503
bits for exponentiation
var B = [504]; // A sequence of
503 bits of the exponent
// Arrays for multiplying the
elements of the Galois field GF (2
^ 503)
var M1 = [504], M2 = [504], R =
[504];
// M1 [], M2 [] - factors R []
the result of multiplication
var MA = new Array (504); // Array
MA [] [] of degrees A []
for (var i = 0; i<504; i ++) MA
[i] = new Array (504);
for (var i = 1; i<= 503; i ++) MA
[
multiplying the elements of the
Galois field GF (2 ^ 503)
for (var j = 1; j <= 503; j ++) MA
[I] [j] = R [j];
} // Put degree 2 in MA [
Our task is to get the same random bit sequences C[] on both sides of the data exchange. This allows to add modulo 2 (XOR operation) bits of the C[] sequence to each bit of data being sent on the transmitting side. With such information coding, absolute protection against disclosure threats in the communication channel is provided. The recipient of the information must add modulo 2 bits of the C[] sequence to the received bits for decryption, which is exactly the same procedure as on the transmitting side.
The transformation process begins by
generating a sequence of 503 random bits on each
side. This is done simultaneously with filling the
array MA[][] with powers of the primitive root of
the Galois field. The number 2 is one of primitive
roots, which should be entered into the array A[].
In our example, the least significant bits
correspond to the lower array indices. Therefore,
we get a primitive root like this:
for (var i = 1; i<= 503; i ++) A
[i] = 0; A [
For raising to a power, a well-known method of simplifying calculations was used, which consists in replacing the operation of raising to a power by a product of powers according to the next expression: (1) where
Since any exponent B can be represented as a sum of values selected from a range of weights 20, 21, 22, 23,…, 2502, to calculate AB it is enough to multiply no more than 503 elements from the array MA.
The block for raising A to power B looks like
this:
for (var i=1; i<=503; i++) A[i]=0;
A[
multiplying the elements of the Galois field GF(2^503) for (var I=1; I<= 503; I++) A[I]=R[I]; } // The elements MA[][] was Multiplied, where B[J]=1.
The function of multiplying the elements of the Galois field according to the rule of polynomials looks like this: function MULT() { // Multiplication using the polynomial X^503=X^3+1 var i, j, r, r1, r2, r3; for (i = 1; i<= 503; i ++) R[i] = 0; for (i = 1; i<= 503; i ++) if (M1[i] == 1) // Select units, because multiplication by 0 gives 0 for (j = 1; j <= 503; j ++) if (M2[j] == 1) r = i + j-1; if (r> 503)
Let's imagine an algorithm for obtaining bit sequences that will be the same on both sides of the data exchange.
Step 1. The client enters a random bit into the first element of the array N, and enters the value of the primitive root of the Galois field into array A.
Step 2. The client executes the block of filling the array MA with powers of A with the simultaneous completion of filling the array with N random bits.
Step 3. The client copies array N to array B and executes the exponentiation block of A.
Step 4. The client sends to the server the result of raising A to the power of B as a sequence of 503 bits
Step 5. The server stores the sequence of bits received from the client in array C and performs actions similar to steps 1-3 of the client.
Step 6. The server sends to the client its result of raising A to power B.
Step 7. The client stores the sequence of 503 bits received from the server in array A.
Step 8. The client executes the block of filling the array MA with powers of A without filling the array with N random bits.
Step 9. The client executes the block for raising A to the power B and enters the result into array C.
Step 10. The server copies array C to array A and performs the steps similar to steps 8 and 9 of the client.
The result of performing the above actions is to obtain the same random sequences of bits in the arrays C of the same name on the client and server sides, which was required for encryption using the one-time pad method. 4. Full-scale model of a secure IoT system
The main element of the IoT system that needs to be protected from false control commands and from intrusion by attackers who can create threats such as DDoS attacks is computer for object management (see Fig. 2). Connecting this computer through the Router without providing a real IP address does not provide the ability to control this computer other than through the console used to install the software or an application program that provides the protection described in the previous section. A well-known minicomputer of the Raspberry Pi 3 type, which has a 40-pin GPIO interface with wide possibilities for connecting objects for monitoring and control, was chosen as hardware. Linux version Ubuntu 20.10 was selected as the operating system, and the Node.js platform version v12.18.2 with the onoff package was used as a programming tool, which allows objects to be controlled via the GPIO interface.
The initial snippet of the CONPIN.js program installed on this computer in the / home / ubuntu / directory looks like this: const HOST = '91 .198.50.144 ';
This client program regularly contacts the server (Mediator server) (see Fig. 2) with a period of 20 seconds to transmit information about the state of objects and receive control signals. The duration of the period of 20 seconds is chosen from the condition of proportionality with the time of entering the Internet. The operation of this program must be protected against possible power outages. To automatically start the program after power-up, add the following three lines to the /etc/rc.local file: #! / bin / sh echo "#################### CONPIN ######################" / usr / bin / node / home / ubuntu / CONPIN &
The SOCKET.js program must be running on the Mediator server (see Fig. 2) located at the ISP (Internet Service Provider) site that provides services in SaaS (Software as a Service) mode. The initial snippet of this program looks like this: // server / SOCKET.js //
With a single intermediary computer with a
single real IP address, the provider can serve
multiple IoT client systems. The number of
supported systems depends only on the technical
data of the computer. The operation of the
SOCKET.js program must be protected from
failures that can lead to an emergency shutdown.
To do this, use the process manager pm2
automatic program restart tool, which must be
downloaded using the npm install pm2 –g
command. After that, the SOCKET.js program
should be launched with the pm2 start
SOCKET.js command. In this case, in case of any
failures, the program will automatically restart
[
The main task of the Mediator server is to
protect the resources of IoT systems from the
penetration of intruders who have as their goal the
implementation of DoS and DDoS attacks. This
requires unauthorized entry into the Mediator
server, which is unlikely, provided the provider
follows standard instructions. Usually this
situation arises due to the fault of the provider's
staff. In any case of failures on this server, the
provider always has the ability to switch to a
backup server or restore the operation of the same
server using copies, which is the norm in the work
of providers [
The exchange of data between users of the IoT system and their objects is carried out via a web interface through intermediate data files. These files are created anew at each data exchange session. Each individual user on the Mediator server is allocated his own directory, where, in addition to the SOCKET.js program with a unique value for the PORT parameter, the vybir.js program is located, the initial fragment of which looks like this: // vybir.js - HTTP Server Ver. 18 February 2021 var http = require ('http'); var url = require ('url'); var fs = require ('fs'); var static = require ('nodestatic'); var querystring = require ('querystring'); var file = new static.Server ('.'); http.createServer (function (req, res) {
In the vybir.js program, a separate TCP port number is allocated for each user. The CONPIN.html file with images of object state indicators and control buttons is also located in the user directory. The user can download this file through the link given to him like http://91.198.50.144:8000/CONPIN.html. All communication processes, including the authorization procedure, are protected using the means described in the previous section. The above link is unprotected as it is only intended to demonstrate the control process using eight binary objects as an example. Authorization data is stored in the same directory in an encrypted file. 5. Conclusions
The reasons for the emergence of security problems in IoT systems are described. Potential security threats have been identified, both for the IoT itself and for the use of its resources by intruders in the implementation of attacks on other objects of the Internet.
Variants of data exchange schemes in IoT systems have been analyzed and the choice of the most secure scheme has been substantiated.
The technical solutions that make it possible to secure data exchange in IoT systems by building an ideally secure data exchange channel are considered in detail. These solutions are presented in the form of text programs in the JavaScript language and can be embedded in any user software.
Using the example of the current model of the IoT system, it is shown that it is possible to eliminate problems with emergencies in IoT systems that arise for various reasons, including malfunctions of programs, temporary power outages or attempts to unauthorized entry into the system. A link to a resource on the Internet is provided to demonstrate the process of managing objects.
The technical solutions proposed in this work make it possible to fully secure IoT systems from information threats. 6. References