The problem of optimizing the choice of cryptographic information security algorithm for e-government IT project management in Ukraine by using nonlinear convolution of criteria based on the method of hierarchies taking into account requirements: security, speed, characteristics of the algorithm is solved in the paper. As a result, the optimal cryptographic algorithm is determined, which ensures the integrity and availability of information during the IT project management of e-government, authentication of users and the inability to deny the fact of sending / receiving information.
Electronic Governance means a way of organizing state power through systems of
local information networks and segments of the global information network, which
ensures the functioning of the authorities in real time and makes daily communication
with citizens, legal entities as simple and accessible as possible, non-governmental
organizations. In practice, this means the organization of government management
and interaction with individuals, legal entities and public organizations through the
maximum use of modern information technologies in public administration bodies.
That is, e-government provides the following: any person through information and
communication [
One of the areas of protection in information systems is cryptographic protection of
information, which involves the use of mathematical methods of information
transformation by means of encryption, imitation insertion or digital signature, etc.
Cryptographic protection can be provided during the transmission of information through
communication channels and when processed on workstations and servers [
When transmitting information through communication channels, the following requirements are imposed: • Ensuring confidentiality of information; • Ensuring the integrity of information; • The authenticity of the parties to the information exchange.
Information confidentiality is ensured by symmetric (GOST 28147-89, DES, 3DES, AES, IDEA) and asymmetric (RSA, El Gamal) encryption. The integrity of the information and the authenticity of the parties is achieved through the use of hash functions and digital signature technologies. The set of technologies that ensure the confidentiality and integrity of information when it is transmitted through unsecured communication channels has been called Virtual Private Network (VPN). In the process of network interaction, information security is in particular ensured through SSL, SSH, S-HTTP, IPSec, and the like. The authenticity of the information sharing parties is achieved through the use of X.509, RADIUS, TACACS + and others. These technologies can be implemented by software and hardware. Information protection on workstations and servers can be implemented through file system-level encryption, cryptographic authentication methods (digital certificates, one-time passwords, etc.), cryptographic integrity checks (checksums).
The problem of protecting information by converting it, which precludes its being read by outsiders, only a few decades ago concerned mainly military operations or related to espionage stories but was not widely used. The cause of the rapid development of cryptography, on the one hand, is the use of computer networks, such as the Internet, which transmit large amounts of information of state, military, commercial and private content, which prevents access to it by third parties, and on the other, - the emergence of new powerful computing tools has made it possible to discredit a number of cryptographic systems. Without cryptography, there would be no cell phones, ATMs, digital TV, Internet payments, etc.
Cryptographic information security methods involve both software and hardware use. Software implementation of encryption is cheaper and more practical. At the same time, hardware implementation is more productive and easier to use. Modern cryptographic systems must meet the following common requirements: • The source text of encrypted text can only be played with the decryption key; • Sequentially sorting through the possible decryption keys in order to reproduce the source text requires a considerable amount of computation time or a high cost to implement these computations; • Encryption algorithm information should not affect the encryption resistance of the encryption system; • Slight modification of the encryption key should result in significant changes to the ciphertext of the same text.
1. Encryption with key. The key encryption algorithm is divided into two large groups - symmetric encryption algorithms and asymmetric encryption algorithms.
Symmetric encryption / decryption methods are a method in which the encryption and decryption keys are either identical or easily computed with each other, thus providing a shared key that is secret.
Asymmetric Encryption / Decryption Methods - a set of cryptographic encryption / decryption methods that use two keys - secret (private) and public; none of the keys can be calculated from another within a specified time. Such encryption / decryption is also called public key encryption / decryption.
Until the 1970s, only cryptography with symmetric crypto algorithms was used. Cryptography with asymmetric crypto algorithms is much younger.
Symmetric and asymmetric crypto algorithms have their advantages and disadvantages. Symmetric crypto algorithms have higher speed and shorter key length than asymmetric ones. Asymmetric encryption is used in such an organization of cryptosystems when the use of symmetric algorithms is impossible. And in general, to compare the characteristics of these crypto algorithms would be incorrect: they are designed to solve different encryption tasks.
2. Symmetric encryption method. Symmetric encryption is also called encryption with a secret key, that is, a key that both parties to the exchange of information (secretly from other users) use to encrypt and decrypt messages. In Fig. 1 is a block diagram of a secret key encryption. The main purpose of symmetric crypto algorithms is to encrypt large data sets at high speed. However, due to the need for a secure secret key transmission channel, these crypto algorithms show very low flexibility when creating modern cryptosystems. There are two major groups of symmetric encryption algorithms: streaming encryption and block encryption.
Sender Message
Message encryption The secret key 3. Asymmetric encryption method. The problem of increasing the amount of encrypted information in cryptography is solved by increasing the speed of traditional secret key encryption methods. However, the application of these methods in the context of a constant increase in the number of participants in a joint work (decentralized management structure) and complications in the organization of interaction between them, in particular pairwise exchange of information, is ineffective. This is due to the fact that as the number of participants in the exchange of information increases, the number of secret keys increases. We can show that for N the number of secret keys in such a system reaches N(N-1)/2. In addition, symmetric secret-key cryptography techniques have difficulty in trusting the secret key. In order to reduce these shortcomings, public key asymmetric encryption methods have been developed. Public key encryption is a relatively new field of cryptography. Asymmetric crypto algorithms use different keys for encryption and decryption: for encryption - open, for decryption - secret. Asymmetric cryptography is based on the ideas of W. Diffie and M. Hellman about two-key encryption, which became known in 1976. But the first algorithm of asymmetric encryption, which became practical, was the algorithm proposed by R. Rivest, A. Shamir, and L. Adleman in 1978. It was called the RSA algorithm. In Fig. 2 a block diagram of a public key encryption is shown.
Sender
Message
Mathematical justification for asymmetric crypto-algorithms consists of heavy-duty
(one-way) functions. The theory of complexity calculates the concept that
characterizes the level of complexity of calculations (number of operations), depending on the
size of the input data. Common are the polynomial and exponential nature of the
dependence of the complexity of the calculations on the amount of input data. In
asymmetric cryptography, V. Diffie and M. Hellman encrypt the message in the presence
of a secret key for the polynomial operating time of the computer system, and in the
absence of it, for exponential time. Modern asymmetric cryptography is based on the
algorithms of El-Gamal and Miller-Koblitz. The theoretical basis for the stability of
the RSA algorithm is the problem of factorization of large integers, and the
algorithms of El-Gamal and Miller-Koblitz - the problem of discrete logarithm. Numerous
vulnerabilities of these algorithms are known today. Public key encryption algorithms
have been replaced by more robust encryption algorithms on elliptic curves, proposed
separately by V. Miller and N. Koblitz in 1986 [
Asymmetric encryption algorithms, as well as symmetric encryption, are used to
encrypt data arrays, but their speed is much lower. The main purpose of asymmetric
algorithms is to ensure the efficient functioning of modern cryptosystems. It is these
algorithms that underpin the tasks of user authentication, control of the integrity and
accessibility of information, the impossibility of rejection of authorship or the fact of
obtaining data, etc., in particular, in the organization of e-government. The following
asymmetric encryption algorithms satisfy the most of these requirements: Mars, RC6,
Rijndael, Serpent, Twofish (Table 1).
Asymmetric crypto algorithms have become important in the application of
electronic-digital signature systems. Electron-digital signature is a digital sequence that is
attached to a message to ensure the integrity of information and authentication, and is
formed using asymmetric cryptosystems. The electronic-digital signature uses a secret
key to form signed messages, and a public key to verify the signature. In the process
of organizing e-government, there is a problem of optimal choice of the cryptographic
information security algorithm. Each of a large number of cryptographic algorithms
has its advantages and disadvantages. Therefore, the amount of analysis of
information regarding the evaluation and selection of the cryptographic algorithm that best
meets the requirements of information security in e-governance is quite large. The
selection process involves quantitative and qualitative analysis in the process of
comparing different alternatives. With the increasing number of comparison criteria and
the number of alternatives that can significantly affect the end result, one can make a
choice among such many options (Table 2 [
For e-government, the following benchmarks are crucial: reliable algorithm execution in both hardware and software; rapid generation and matching of keys, their use; minimal memory usage resistance to attacks; flexibility; high bandwidth. So let's define the main criteria for comparing cryptographic information security algorithms for egovernment: security; speed; general algorithm parameters. The criterion "security" is the most important factor in evaluating and comparing such capabilities as the stability of the algorithm to cryptanalysis, the study of its mathematical basis, the randomness of the original values of the algorithm and the relative security compared to other algorithms. The "speed" criterion is another important evaluation criterion that characterizes computational performance across platforms, memory requirements, time spent on encryption and decryption, and attack speed. The third priority in order to evaluate algorithms for e-government is the characteristic of the algorithm, which means: flexibility, hardware, software suitability and simplicity of the algorithm. Flexibility includes the ability of the algorithm to: • Key management, minimizing size; • Implementation of safe and efficient functioning in different types of software environment; • Implementation of hashing algorithm, possibility of providing additional cryptographic services.
The fulfillment of these requirements is necessary in order that in electronic
governance the hardware and software support the implementation of the selected crypto
algorithm. Table 3 shows the comparative characteristics of the five crypto algorithms
according to the identified comparison criteria [
Implementation options (flexibility) imply the ability to perform any optimization algorithm operation for specific purposes in different ways. Most illustrative in this sense are the previously mentioned options for the procedure of the extension of the key of the Twofish algorithm, which allow to optimize the implementation of the algorithm depending, first of all, on the frequency of key change. 5
One approach to assigning “weights” to the final set n compared objects on the basis
of the matrix of paired comparisons was proposed by T. Saati [
T desired measure vector w = ( w1, w2 ,..., wn )
is calculated as the eigenvector of this
matrix corresponding to the maximum eigenvalue. This method of determining the
vector of measure by virtue of the violation in practice of the compatibility properties
[
Let's make it clear. It is well known that the vector of measure w is a native vector
of compatibility (in some sources, the name is fully compatible) of the matrix A,
corresponding to its own maximum value n. Thus, in the case of a compatible matrix, the
measure vector is a specified eigenvector. But when hierarchies are formed in
accordance with the method of hierarchy analysis, it is not necessary to count on the
compatibility of the matrix of pairwise comparisons. This is known to all who are familiar
with the method of hierarchy analysis. This means that in practice, you have to deal
with another situation (model) that is matched by an incompatible matrix.
Nevertheless, according to the method of analysis of hierarchies, the measure vector is again
proposed to be found as an eigenvector of an (incompatible) matrix of pairwise
comparisons, and this eigenvector corresponds to an eigenvalue that is no longer equal
(but strictly larger) n. In the general literature on the method of hierarchy analysis,
there is (at least at the moment) no proof that the required measure vector must be an
eigenvector of an incompatible matrix corresponding to its maximum eigenvalue
greater than n. For this reason, this method can not be called justified, it is a certain
heuristic approach, the logic of which is to recommend to act in such a way in
situations that may be very different from those for which the validity of these actions is
established. This means that the application of the hierarchy analysis method almost
always contains some "model" error of calculating the vector of measure (not taking
into account errors of purely computational nature) and, if this error is large, then the
application of the method of hierarchy analysis becomes simply unjustified.
Therefore, a special numerical indicator of consistency index is introduced, which
characterizes the degree of confidence in the hierarchy results obtained by the method of
analysis. This index is interpreted as a kind of deviation of the original incompatible
matrix from some compatible one. As T. Saati points out [
Let’s note that in the real decision-making process there are related problems that are successfully solved using the method of hierarchy analysis. The method of analyzing hierarchies is a methodological basis for solving the problems of choosing alternatives by means of their multicriteria ranking. The main application of the method is to support decision making with the help of hierarchical composition of the task and the ranking of alternative solutions. Given this fact, it is necessary to recalculate the possibilities of the method.
1. The method allows to analyze the problem. The problem of decision making is presented in the form of hierarchically ordered: • The main purpose (main criterion) of ranking possible solutions; • Several groups (levels) of the same type of factors that affect the rating in one way or another; • Groups of possible solutions; • Communication systems that indicate the interplay of factors and decisions. All of these "nodes" are assumed to indicate their mutual effects on each other (links to each other).
2. The method allows to collect data on the problem. In accordance with the results of hierarchical decomposition, the model of the decision-making situation has a cluster structure. The set of possible solutions and all the factors that influence the priorities of the solutions are broken down into relatively small groups - clusters. The hierarchy procedure of paired comparisons, developed in the method of hierarchy analysis, allows to determine the priorities of the objects belonging to each cluster. This is done using the eigenvector method. Therefore, the complex problem of data collection is broken down into a number of simpler ones, which are solved for clusters.
3. The method makes it possible to evaluate and minimize data conflicts. To this end, harmonization procedures have been developed in the method of hierarchy analysis. In particular, it is possible to identify the most conflicting data, which allows to identify the least clear areas of the problem and to organize more careful selective reflection of the problem.
4. The method allows to synthesize the problem of decision making. After analyzing the problem and collecting data for all clusters, a final rating is calculated, using a special algorithm, which is a set of priorities for alternative solutions. The properties of this rating allow you to support decision making. For example, the highest priority decision is made. In addition, the method allows you to build ratings for groups of factors, which allows you to evaluate the importance of each factor.
5. The method allows to organize the discussion of the problem, promotes consensus. Thoughts that arise when discussing the problem of decision making can themselves be considered as possible solutions in this situation. Therefore, the hierarchy analysis method can be applied to determine the importance of accounting for each participant's opinion.
6. The method allows to evaluate the importance of accounting for each decision and the importance of accounting for each factor affecting the priorities of decisions. According to the formulation of the decision-making task, the priority value is directly related to the optimality of the decision. Therefore, low priority decisions are dismissed as non-influential. As noted above, the method allows to evaluate the priorities of the factors. Therefore, if the priority of decisions is changed when a factor is excluded, such a factor can be considered non-influential for the task.
7. The method allows to evaluate the stability of the decision. A decision that can be made can be considered justified only if the inaccuracy of the data or the structure of the model of the decision-making situation does not significantly affect the ranking of alternative decisions.
If it is sufficient to use only objective data for decision making, then other methods (e.g., target criterion optimization methods) may be predominant in terms of accuracy and speed. The method may be too cumbersome to decide in simple situations, because many pairs of comparisons are needed to collect data. However, if a large-scale problem is considered and the cost of the consequence of a wrong solution is high, adequate tools are needed. The method of analysis of hierarchies allows you to break a complex problem into a series of simple, to identify contradictions. Strategic decision-making often has to rely more on the experience and intuition of professionals than on objective data. In this case, the results obtained by the method of hierarchy analysis may be more realistic than the results obtained by other methods. Ratings of possible solutions are based on "transparent" principles. Therefore, they may be more persuasive than the information to support decision-making obtained through black box models. In such models, the input of the problem will be transformed into the output of the decision-making process according to "opaque" principles and the structure of the decision-making situation is not disclosed. The method of analysis of hierarchies does not require simplification of the structure of the task, which is the a priori rejection of some features. Therefore, it is more effective than other analytical tools to take into account the influence of various factors on the choice of decision.
Composing a decision model can be a cumbersome process. However, if folded, it
can then be reused. It is only necessary to correct this structure and fill it with data. In
this case, the solution of typical problems can be put in the flow. Thus, the application
of the method becomes more efficient. Now, let's look at a specific example of
optimizing the choice of cryptographic information security algorithm for e-governance
with a simplified method of hierarchy analysis based on nonlinear convolution. [
Software tools for the process of selecting the nomenclature of security elements. As mentioned above, we will use the software system to optimize the process of choosing the cryptographic information security algorithm for e-government IS «Vybor». The results of the calculation are presented in Fig. 3-4; Table 4.
Goal level Сriteria level Level alternative
Serpent (0.105)
Security (0.691)
TwoFish (0.169)
Goal Speed (0.218) MARS (0.162)
Options (0.091) Fig. 3. Hierarchy: cryptographic information security algorithms for e-government
Fig. 4. Diagram of the result IT project management is the process of planning, organizing and delineating responsibility for the completion of an organizations' specific information technology (IT) goals. One of the scientific concepts is that project management is an orderly sequence of decision making. The features of life cycle and project management technologies impose specific requirements for the application of management decisionmaking methods in them, which provide instrumental support for decision-making processes in projects. Some popular approaches include the Hierarchy Analysis Method or Analytic Hierarchy (AHP) method proposed by American Professor T. Satie (AHP), and its generalization to the Analytical Hierarchy Method (MAC).
The implementation of projects in specific areas, especially at the level of security organization of the functioning of information systems, requires the priority consideration of their business models for the assessment of situations and decision making. Project activity in the educational and public sphere attracts attention due to the participation of a large number of highly qualified specialists, non-standard value look at the criteria of project implementation efficiency, constant interaction with traditional processes that can resist any and all innovations. The use of decision-making methods in educational and governmental projects is additionally related to the public nature of most of them, which requires the use of clear (including for non-specialist) and constructive (regarding the processing of quality information) decision-making technology. The criteria for maintaining safety in education / government projects by the Saati method are the subject of this study. Government and educational projects have peculiarities in the structure of the phases of the life cycle, which determines the specific nature of the application of management methods in them security-enabled solutions at the IT project management. Different methods of government / educational IT project management can use different methods as an instrumental tool for managerial decision making and one of the most versatile is the Saati method. The accumulated experience of implementing IT projects in areas with high levels of intellectual saturation creates preconditions for comparing life cycles and opportunities for applying management decision-making methods. The life-cycle features of government / educational IT projects are often driven by the high level of their intellectual saturation and the complexity of evaluating intermediate stages of implementation. In this regard, a comparative study of the life-cycle phases of government / educational IT projects and software development IT projects appears promising. Such comparisons can be made in a broad context on the basis of the international standard ISO / IEC 12207: 2008 “System and software engineering. Software Lifecycle Processes” and in a narrow sense based on software development lifecycle models. 6
According to the criteria of evaluation of cryptographic information security algorithms for use in e-governance, it is recommended to choose the algorithm that is the most effective in the integral indicator. The “security” criterion has the highest priority and has the greatest impact on the results, and the “speed” and “algorithm” criteria are secondary to “security”. Guided by the robust security of attack algorithms, MARS, Serpent and Twofish have a high level of security, but RC6 and Rijndael have higher and more secure protection. RC6 and Rijndael generally show higher encryption and decryption rates than average for 128 bit keys but 32 bit platforms, and RC6 has the highest speed. MARS has an average speed of doing the same. For Twofish, the time spent on encryption and decryption is different, but in both cases the level is above average. Serpent showed the lowest performance compared to other algorithms.
Rijndael requires little RAM and is therefore best handicapped. Serpent also provides the right level of encryption and decryption for low RAM. The RC6 has a small amount of RAM, which is a positive thing in the limited space, but there is a downside to the continuous ability to compute decryption subkeys, creating a high RAM requirement for other algorithms. MARS does not meet the requirements in a restricted environment and requires additional resources. Serpent and Rijndael have the best hardware performance for both feedback and feedback. Serpent has the highest performance in feedback, Rijndael offers the best performance in feedback. The RC6 and Twofish have average performance, and both algorithms can run compactly. MARS has high requirements and overall performance is below average. When performing attacks, Rijndael and Serpent's algorithms performed well, quickly detecting and preventing them. Twofish performs longer and with greater complexity, and RC6 and MARS counteract attacks with the greatest amount of time and difficulty. Twofish, MARS, and RC6 require little extra space to encrypt and decrypt. Although Rijndael is inferior in this aspect, it may share some technicalities. Twofish supports continuous calculation of subkey counts for both encryption and decryption. Serpent also supports continuous calculation of subkey counts for both encryption and decryption; however, the decryption process requires one additional calculation of the calculation. The Rijndael algorithm supports continuous computation of encryption subkeys, but requires the previous one-time execution of the full key list before decryption with a specific key earlier. MARS has special features that are similar to Rijndael, but additionally loads the resource for MARS execution. RC6 supports continuous computation of encryption-only subkeys. Each of the algorithms provides reliable security and has advantages in certain areas compared to others. The method of hierarchy analysis based on nonlinear convolution of criteria has been investigated and mathematically substantiated the choice of Rijndael algorithm as the one that best satisfies the requirements of information security in e-governance.