Otkritieinfo.ru



The Strategies

of Modern Science

Development

International

scientific – practical conference

Yelm, WA, USA

29-30 March 2013

Science Book Publishing House

Yelm, WA, USA

2013

Scientific Publishing Center "Discovery"

otkritieinfo.ru

The Strategies of Modern Science Development

Proceedings of the International scientific–practical conference

Yelm, WA, USA, 29-30 March 2013

The Strategies of Modern Science Development: Proceedings of the International scientific–practical conference (Yelm, WA, USA, 29-30 March 2013). - Yelm, WA, USA: Science Book Publishing House, 2013. - 210 p.

The materials of the conference have presented the results of the latest research in various fields of science: information technology and engineering, agricultural science, economics and jurisprudence, educational sciences end art criticism, social and political sciences. The collection is of interest to researchers, graduate students, doctoral candidates, teachers, students - for anyone interested in the latest trends of the world of science.

ISBN 978-1-62174-024-7

section 1. Information Technology

THE USE OF THE PRIORITY MODEL

IN OPTIMIZATION OF CORPORATE DATA NETWORK ADMINISTRATING

Alex V. Andreev, Maria M. Monakhova, Denis V. Mishin

Vladimir State University, Vladimir, Russian Federation mariya.monakhova@

The main integrative platform of the modern automated management system for the enterprise (AMSE) is the corporate data network (CDN). CDN is distributed infrastructure which is organized component complex such as functional elements (end devices, telecommunication devices, protocols and data network services) (FE) and telecommunication links. In this context CDN administrating is target manager actions on FE implemented by administrator to ensure reliability, capacity and security of CDN as the objective goal of administrating [1].

CDN administrator is a person-machine system implementing set of network management functions. One of the actual tasks in providing need quantity of automated management system for the enterprise (AMSE) operation is designing a queue of set of all needed to execute administrating functions. With limited administrator’s amount optimization of such queue is one of the most actual tasks in providing needed quantity of AMSE functioning. Authors main the approach to the problem in ranging FE of CDN for extend of their share in providing transmission means to the certain finite set of the information processes of the AMSE. Problem solving of the ranging FE of the CDN is represented as developing information systems prioritization under the math model of the FE priority.

In this context priority is a common for every FE of the CDN index of objective importance in the set of information processes of AMSE. Theoretical tasks of network administrating are difficult scientific problem related with developing science-based models and algorithms of the adapted control. Analysis of the actual investigations in this sphere of existed on the market specialized administrating software environments permits to state absence of the effective ranging FE of CDN mechanism. Today using peer reviews of FE of CDN with continuous modernization of CDN and changes of AMSE applications become ineffective because of their low dynamic. Absence of the theoretical and practical developments in ranging FE sphere determines relevance of developing science-based methodology of quantitative priority calculation of FE of CDN. This work introduces the notion of FE of CDN, suggests priority model and method of calculating based on degree of FE participation in implementation set of information processes of AMSE.

Designate set of FE of CDN as SCDN={s1,s2,...,sj}, where sr[pic]S – functional element. Priority (R) is an index of importance of functional element for implementation of information processes of AMSE. Designate a quantitative priority value sr as R(sr). Designate set of all information processes of CDN PCDN={p1,…,pm}, where information process designate pi[pic]PCDN

Information process pi is an information interaction of two and more subjects (users, organization departments, corporations).

The aim of information process is a changes existing information if only one of them. Implementation of information process is related with using FE.

In general case information process is represented by: [pic] where Hi – quantitative assessment of range pi, defined by expert group according to the degree of importance and urgency of a process; Ai={ai1,…,aik} – set of senders functional elements (FE) pi, Ai( SCDN; Bi={bi1,…,bie} – set of receivers functional elements (FE) i, Bi( SCDN; Wi={wi1,…,wip} – set of primary oriented way (all potential distances of network traffic between the subscribers), wiq(Wi – way from aij(Ai to bij(Bi, wiq( SCDN.

Priority numeric value of FE for pi is in proportion to coefficient of FE participation (() in implementation information of information process and range of a process (H): [pic] where Ri(sr) – priority numeric value of the element sr for the process pi; γi(sr) – coefficient of relative participation of the element sr in implementation pi that is determined by the count of its emergence on Wi.

For calculating γi(sr) the set PCDN should be represents as related connected undirected graph NCDN(S,L), where vertices of a graph are the elements SCDN, arcs of a graph are the telecommunication channel.

[pic]

[pic]

Fig. 1. Undirected graph NCDN (up), directed graph Ni (down)

Each information process of CDN as an ordered set of functional elements is a directed subgraph of the sought graph NCDN, Ni – directed graph pi (Fig.1). Set of all alternative combinations between senders’ FE and receivers’ FE in a process pi is called the set of ordered pairs {aij,bij} and is designated as Mi, aij(Ai, bij(Bi.

The set Mi is found as a Cartesian product [1] of the set of subscribers (Ai, Bi) by formula [pic] Define quantity of pairs {aij,bij} as |Mi| (cardinality of the set Mi). For the quantity of pairs |Mi| of “sender-receiver” of process pi find all ways of interaction Wi={wi1,…,wip}, including appropriate senders and receivers. |Wi| denote number of found ways. Denote |Wi| as a number of found ways. Set of ways through sr denote as Wri, Wri(Wi. |Wri| denote number of found ways Wri. Denote |Wri| as a number of found ways Wri. Parameter (i(sr) determines number of |Wri| emergence on the set of ways Wi by formula (4).

[pic] [pic]

Set the result (4) to the formula (2) and calculate numerical value of Ri(sr) (5). After calculating functional elements priority of the set SCDN on the whole set PCDN, a matrix of functional elements priority will be got (Table 1).

Table 1

| P |p1 |p2 |p3 |

|S | | | |

|X=2% (KA=KF=1) |31 |70 |99 |

|X=65% |1 |2,5 |3,8 |

|X=100% |0,66 |1,65 |2,5 |

[pic]

Fig. 1. The ratio of the run-time query

Fig. 2 shows a plot of the average time of the query column in database of the number of processors for different ratios used in the query attributes (10%, 50%, 100%) as well as run-time query in lowercase database. The graphs show that the fifteen-second mark of line database of the average time the query is achieved when the number of processors n = 10. Column Database, this mark is reached at the ratio used in the query attribute 10% (10 out of 100 attributes), even when n = 2 (and that in the absence of compression columns, kC = 1). Saving computing resources is available.

[pic]

Fig. 2. The average query time in the database column (page).

The resulting mathematical model can be used to make well-founded technical solution of choice in a database. It is expected to continue studies of and to obtain estimates of time queries with more complex plans realization.

Bibliography

1. Андрей Арсентьев. Хранилища данных становятся инфраструктурным компонентом №1. CNews аналитика. 2010. [Эл. ресурс]. []. Проверено 27.06.2011.

2. Michael Stonebraker, Uğur Çetintemel. «One Size Fits All»: An Idea Whose Time Has Come and Gone.: [Эл.ресурс]. []. Проверено 27.06.2011.

3. Michael Stonebraker, Chuck Bear, Uğur Çetintemel. One Size Fits All? – Part 2: Benchmarking Results. 3rd Biennial Conference on Innovative Data Systems Research, 2007, Asilomar, California, USA.: [Эл.ресурс].[]. Проверено 27.06.2011.

4. Michael Stonebraker. My Top 10 Assertions About Data Warehouses. / Перевод Сергея Кузнецова, 2010 г.: [Эл.ресурс]. []. Проверено 27.06.2011.

5. Григорьев Ю.А., Плутенко А.Д. Теоретические основы анализа процессов доступа к распределенным базам данных. Новосибирск: Наука, 2002. - 222 с.

6. Ю.А. Григорьев, В.Л. Плужников. Оценка времени выполнения запросов и выбор архитектуры параллельной системы баз данных. МГТУ,2009.

7. Григорьев Ю.А., Плужников В.Л. Модель обработки запросов в параллельной системе баз данных // Вестник МГТУ им. Н.Э. Баумана. – 2010. - № 4. – С. 78-90.

8. Григорьев Ю.А., Плужников В.Л. Оценка времени соединения таблиц в параллельной системе баз данных// Информатика и системы управления. – 2011. - № 1. – С. 3-16.

9. Григорьев Ю.А., Плужников В.Л. анализ времени обработки запросов к хранилищу данных в параллельной системе баз данных // Информатика и системы управления.–2011.-№ 2. – С. 94-106.

10. Григорьев Ю.А., Ермаков Е.Ю. Модель обработки запросов в параллельной колоночной системе баз данных // Информатика и системы управления. – 2012. - № 1. – С. 3-15.

11. Григорьев Ю.А., Ермаков Е.Ю. Модель обработки запроса к одной таблице в параллельной колоночной системе баз данных и анализ ее адекватности // Информатика и системы управления. – 2012. - № 2. – С. 170-179.

THE ADAPTED ALGORITHM OF KUN-MANKERS

IN ADMINASTRATIVE TASKS ECM-COMPUTING ENVIRONMENT

Maria M. Monakhova, Denis V. Mishin,

Sergey D. Luchinkin

Vladimir State University, Vladimir, Russia

mariya.monakhova @

Analysis of current research in the field of sustainable operation of distributed systems (eg ACS) suggests that one of the components of the quality of corporate functioning of ACS is the problem automate recovery of its data-processing environment (KRIVS). KRIVS recovery is made, as a rule, in the processes of administration, the decomposition of which can provide a lot of precedents F (operations) - functions of the administration (PA). Analysis of the administrative system KRIVS company suggests that the task of distribution is decided by the FA is not effective. This paper proposes a method for appointing the FA artist, based on an adapted algorithm for assignment problem Kun - Mankres, also known as the Hungarian algorithm.

Mathematically, the optimization problem assigning functions by administration in the form of a bipartite graph G '= (A', F '; Y) (Fig. 1), where A' = {a'1, a'2, ..., a'n }, A'( A - a subset of available administrators KRIVS, F' = {f'1, f'2, ..., f'm}, F'( F - subset FA, need to be addressed in the current phase of the administration, Y = {yij} - the set of edges connecting the vertices of A 'to the vertices of the set F',,, i> 0, j> 0.

Figure 1 - The F '

We will assume that each administrator can perform any FA. This corresponds to the fact that each vertex of a'( A' is related to all the vertices f'( F', the incidence βa each vertex a'( Ak in this case will be equal to the power subset FA ACS KRIVS, βa’=|F'|. Similarly, each FA can be done by any of the administrators. This corresponds to the fact that each vertex f'( F' is linked to all. will be equal to the power available to a subset administrators KRIVS, βf’'= | A' |. Thus, a bipartite graph G '= (A', F '; Y) is complete (Fig. 1), [pic]a'(A' and f'(F' [pic]yij(Y.

In solving the problem of destination by the FA need of many possible choose a matching (F'-A '), which best meets the specified performance criterion integral T. Since each edge yij(Y can be made several performance metrics t * ij, task assignment by the FA can be reduced to finding a maximal matching vertices of F 'to the vertices of the set A' so that the total performance of all combinations is optimal (Figure . 2).

Integral criterion efficiency T will assume the total execution time in the cycle of the FA administration. Thus, it is necessary to appoint administrators to set A 'as the administrative control of F' in such a way as to achieve a minimum time of complete recovery performance KRIVS:

[pic] Figure 2 - the optimal assignment

[pic] (1) (1)

We take the value of the efficiency, which determines the weight of each edge yij( Y graph G = (A ', F'; Y), the predicted execution time t * FA fj( F 'administrator ai( A'. Index t * - a complex function, which is calculated based on the individual competencies administrator settings the performance of a particular FA. In general, predictable execution time FA define as: t*ij=f([pic]) (2), where: - the average execution time fj administrator a i;- An indicator of the competence of the administrator ai to implement fj; [pic] - the interval between the time the last run fj administrator a i [pic], and the time of this appointment fj administrator to perform a i;.

The competence of the manager is an integral indicator of the level of knowledge, skills and experience. In the simplest case, with the value of the index K competence assume probability of the FA administrator for a time not exceeding normative, where - standard execution time fj, - the value of the last execution fj administrator ai. We assume that the increase relative to the average time function is proportional to the probability of default functions: (3). Then the predicted execution time fj ai administrator will calculate the following: (4). Due to the loss of experience, the value increases as a function of the value of the exponential law: (5), where the interval between the time the last run and the time the next appointment fj administrator to perform ai. As part of the required tasks, the predicted time to time current assignment functions fj will be calculated as: (6).

The result is a matrix of the projected фcting A'F 'represents a binary relation sets A' and F ', each ordered pair of which is expressed by a numerical value of the integral index predicted time t * ij execution fj ( F 'administrator ai ( A'. If the administrator ai ( A 'can not execute a function fj ( F', the value of t * ij =.

Formulation of the problem. With limited administrative resources (power set A ') subject to the limitation that the performance of each f, fF', one is given the administrative resources necessary to ensure the optimal allocation of resources by the FA, ensuring minimal time, incidence KRIVS.

Feature of the optimal use F 'to A' is that one cycle F ', corresponding to | A' |(of administration will run a subset of F'', F'' - the number of available administrative resources. A subset F'' will be formed depending on the conditions determined by | F '|:

• Condition 1: | F '| | A' | F'' = F '. Recovery performance KRIVS can be achieved in a single cycle execution stage as the administrative management;

• Condition 2: | F '|> | A' |. Expected full recovery KRIVS performance can be achieved in more than one cycle of the execution stage FA. Formation of F'' ( F ', for each cycle based on the priority values ​​F'. Function with higher priority are allocated to earlier cycles, which provides priority restoring items with the highest relevance to performance KRIVS. This feature allows you to achieve the best possible performance gain KRIVS at each stage of the execution cycle FA.

Solution. Form the matrix of the projected run-time FA:

• If | F'' | = | A '|, we construct a square matrix of size A'F'' | A' | = | F'' |;

• If | F'' | ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download