This investigation focuses on the problem of distribution of information resources between workstations in local networks. The usage of computer networks in corporations allows implementing real-time information links between official structure element, providing efficient and distributed data processing. In this work we present a mathematical description of the algorithm for quasi-optimal distribution of related information resources at designing automated workstations in local corporate network. The undirected graph describing the task of information resources optimal distribution is presented. The computation method of quasioptimal related resources distribution between automated workstations in local corporative networks has been developed based on presented algorithm. Using conditional corporation as an example the modeling of optimal distribution of related information resources has been considered in local corporative network. Based on the obtained results, five sets of the related information elements describing a subject area were identified in the information system of the conditional corporation: The described computation method provides an opportunity to optimally distribute the information resource in the local corporate network, as well as solve the task of building reliable and efficient local networks. The proposed method of quasioptimal distribution of related information resources can be used in corporation of any type.
At present, the most peculiarities in corporation’s activity are the wide use of IT and the effective
distribution of information resources at designing their organizational structures [
The use of distribution of related information resources network technologies in everyday business has become a norm for western companies. The distributed network technologies are used not only for messaging and for access to information resources, but also for carrying out specific commercial transactions. The development of such technologies helps to reduce costs, accelerate all business processes and, as a result, increase the profitability of the company’s operations. The hightechnologies degree and practical implementation of network technology of distribution of related information resources in any modern enterprise considerably determines its success on electronic market. From the level of distributed network application, directly depends the competitiveness of its goods and services.
Manufacturing companies of hardware and software have been the first to use the distributed
network technologies at goods and services marketing and distribution. An example of successful
development of many of them has become a factor of attractiveness for businesses operating in other
areas. Accumulated experience of distribution of related information resources network technologies
is generalized, even the scientific researchers are conducted and materials are published [
The technological aspects of modern economic activity of corporations are extremely important
factors that ensure the effectiveness of its operation. The reliability, safety of technological solutions
is the basis of economic activity, which depends on the effective of distribution of related information
resources. The current level of technology development allows creating high-performance, protected
from external interference information and communication means of distribution of related
information resources. There is also a certain list of organizations that support the reliable operation
of network technologies the main means of information resources distribution of enterprise [
The allocation of computing resources is followed by much greater management decentralization than it could be usually observed in centralized environment, when the data center had strong control over the resources of a large computing system. One of the positive aspects of decentralization is the higher degree of dynamics in multiple areas of computer equipment usage, such as installation and development of applications, connection of peripherals, system expansion, etc.
One of the promising methods to increase the resource management efficiency is implementation of the techniques and tools of new information technology in the performance of the officials. In modern condition the computer, used by authorities as a tool of individual automatization, should be considered as one of the centers of complex technological process of organization document flows of relevant information. The continuity and independence of this technological process requires close attention to the optimization of communication between official’s automated workstations, which use local area networks (LANs).
In this work LAN is addressed as a distributed computing system built on the basis of a common data transmission environment (local area network), which provides physical connectivity of all system components, easy system reconfiguration and comprehensive coverage of its structural elements. Local area networks, having the ability to configure-to-order, allow forming necessary structure of discrete physical elements and processes. This allows fitting into the existing organizational and staffing structure in the early stages of management automatization.
Local area networks are an integral part of distributed in time and space information processing
and storage systems. Analysis of foreign literature has shown that local area networks are the main
component of information systems during the automatization of managerial work [
The usage of computer networks in organizations allows implementing real-time information links between official structure element, providing efficient and distributed data processing.
Distributed database (DB) is a set of logically interconnected databases, distributed via the
computer network [
In general, a Distributed Data Base (DDB) is typically a collection of databases (DBs) that
includes fragments of many databases that are located on the different nodes in a computer network
and are managed by different management systems (DBMS), still remaining available for sharing in
different applications (Figure 1). A distributed database looks like a regular local database to users
and applications. In this sense, the term "distributed" reflects the way the database is organized, but
not its external characteristics ("distribution" of the database is invisible from the outside) [
This section presents a method for determining the quasi-optimal distribution of related information resources at designing automated workstations in a local corporate network. Let us directly consider the mathematical algorithm and method of information resources optimal distribution.
Data set di∈D, i = 1, 2, ...,m, used by N operators, frequency matrix F = ║fij║ (i = 1, 2, ..., m, j = 1, 2, ..., N) of data usage by officials (estimated data flow, unit of measurement - for example, bod). The data should be distributed over sets Oj (j = 1, 2, ...,N), which form a distributed database of the organization O = O1∪O2∪...∪Oj∪...∪ON(Oi∩Oj = ∅for i≠j).
Each set Oj is located in the base computer of the user (operator) oj, the computers are connected to the local network of the organization. There is given a matrix of communication values С = ║сkj║ (k = 1, 2, ...,N; j = 1, 2, ..., N) of the operator k with the "outlying" database j (ckk= 0) – so fee for communication via the network is known.
It is necessary to find a distribution of elements of the set D between sets Oj, which minimizes the total loss from the failed distribution:
N N m S = ∑∑∑ fijc jk xij → min,
k=1 j=1 i=1 Where xij is the membership function, which is defined as follows: xij = ⎩⎨⎧10 ↔↔ ddii ∉∈OOjj , (1) (2)
The key idea of the algorithm for the distribution of vertices included in the data set D between N sets, which are formed near the operator vertices oj, is: the maximum value of fij for all i and j is determined ( ∀ij( fi**j* = miajx fij ). Data di* for which i=i*, j=j* ( fij = fi**j* ), is attributed to the operator set oj (j = j*).
The tuple di (i=i*) is removed from the description of the graph di (i=i*): D1 = D \ di* .. The procedure is repeated until all data is deleted. Since all data must be distributed, the order of processing tuples for di does not matter, so for each di in its tuple fij** = max Pr2 Pr3 d j is determined j and di is attributed to the set of operators j*.
It is better to describe the result in tuples relatively to the vertices included in the set O. 2. General case.
Obviously, when assigning a given d1 to the set o1, the partial loss due to such an assignment is
N s11 = ∑ f1jcj1 . Similarly, for an arbitrary data element di its assignment to the set Oj leads to losses: j=1 In the general case, we have a loss matrix sij i = i, N, j = 1, N.
N sij = ∑ fikcki ,
k =1
sij = fik × ckj ,
The algorithm, which is described above, is focused on using the matrix f , and it can be applied
ij
to using the matrix s . The graph is described by tuples [
In contrast to the others algorithms 1 in this algorithm minimum value of sij for all i and j is determined ( ∀ij(si** j* = min sij )). Data d * , for which i=i*, j=j* ( sij = si** j* ),are attributed to the set of ij i the operator oj (j=j*).
Next, the algorithm completely coincides with the algorithm described in item 1: the tuple di (i = i*) is removed from the graph description: D1 = D \ di* The procedure is repeated until all data is removed. For each di in its tuple si*j* = min Pr2 Pr3 d j is determined and di is attributed to the set Oj*. In j accordance with the mentioned above a method of quasi-optimal distribution of related resources in the design of workstations in the LAN of the corporation is proposed (Figure 3).
In this section presented the results of modeling of optimal distribution of related information resources using the conditional organization as an example. To explain the algorithm, its work on the example of the simulated organization is considered. Analysis of the organizational and staffing structure of the organization official, the tasks to be solved by departments during the preparation of activities allowed identifying groups of users of the distributed database (Table 1). Here, сkj =1/Lkj is taken as the price, and there Lkj is the length of the cable network between the machines. The task is to distribute data between automated workstations (groups of operators).
First of all we solve this problem of distribution, using only the frequency matrix F. According to the proposed algorithm we obtain the following distribution (Figure 4): • D1 refers to the database operator O3, because 175 is the maximum for row D1 and is in column O3; • D2 refers to the database of the operator O2; • D3 refers to the database operator O2; • D4 refers to the database of the operator O1; • D5 refers to the database of the operator O11. In the general case, using the proposed algorithm, we find the loss matrix sij = fik × ckj .(Table 4).
S=
D1 D2 D3 D4 D5
O1 59,69 51,31
6,6 20,45 6,672
O2 35,18 30,4 4,962 11,19 3,04
O3 43,38 46,13 7,001 14,64 5,278
O4 55 71,87 9,584 19,46 4,859
O5 67,68 58,42 8,467 23,51 7,514
O6 63 68,64 9,349 19,32 6,311
O7 72,24 69,59 11,46 24,82 8,632
O8 89,98 81,91 15,12 29,49 6,747
O9 53,96 52,79 8,267 20,08 6,571
O10 66,29 67,55 9,579 22,5 8,764
O11 57,49 58,48 7,227 20,5 6,228
O12 70,53 67,93 7,549 23,54 7,487 Acting according to the algorithm, we obtain the following distribution (Table 4): • D1 refers to the database operator O8; • D2 refers to the database operator O8; • D3 refers to the database operator O8; • D4 refers to the database operator O8; • D5 refers to the database operator O10.
The first and fifth results do not match. The structure of the database is significantly influenced by the structure of the network. The use of the proposed method of related resources distribution is possible in organizations of any level.
Based on the analysis of management concepts of information systems and effectiveness of the LAM functioning, criteria are determined that can be used as a basis of the model for solving the tasks of optimal related resources distribution in information systems. The use of the described algorithm allows to optimally distribute the information resource in the local corporative network, as well as to solve the problem of building a local network with the highest reliability and high efficiency. A method of quasi-optimal distribution of related resources in the design of workstations in local corporative network is proposed. When constructing the optimal structure of a distributed database, many factors that influence the final decision are easily taken into account by including their impact in the cost of communication between resources.
This study was supported by the Ministry of Education and Science of Ukraine, Project No. 0112U000635, “Development and implementation of the modern information systems and technologies in the socio-economic activities”.