The Communication-Information Systems (CIS) operating in real time scales are undergone of uneven charging for request service because of random nature of their appearance and the different information volume. It results to a necessity of development of unstationary models of the system working in real time scales, which to reflect maximum the behavior of a real network. Ones can be applied to define the optimal system operation modes. The purpose is to model the traffic dynamics in distributed control systems networks on base of techniques using conclusions of the tail-end theory. In such approach, the dynamics of the processes of appearing and arranging in the tail-end and requests servicing can be presented by using unstationary (variable in time) values of average quantities. The mode, suggested in this work can be applied for definition of respective optimizing problems, identifying the traffic distributing approach by relevant algorithms of their solution in real time scales. It is foreseen these algorithms to be tested at the information traffic routing for the TCP/IP network protocol.

The computer networks, and especially, potential implementations of Information Searching Systems (ISS) in Internet are an object of study and optimization in this report. The discussion would be specialized on the network traffic studying and information flows modeling.

The integration of computing machines and information resources in a joint computer network (LAN, WAN) is forced the necessity to solve the problems on opportunely and optimal distribution of information flows in the network [1,2]. The computer networks enable of terminated users to exchange, transmit, and process opportunely information as the terms of servicing these requests is short. The response time is depended by delays, generated by the computer systems high-speeding; routing algorithms; delays of signal transmission upon communication lines; algorithms providing request access to the computer network, and the network topology [1,3]. The delay limitations and optimization of the various factor relationships, causing the delays, allow enhancing service quality in the computer network and its efficiency. These problems are strongly affected by the computer networks scales enhancement, the process dynamics, and the unstationary parameters of information exchange. The problem resolving by hardware devices (for example, by enhancing traffic capacities of the communicational lines) no ever is permissible because of financial or/and organizational reasons. The problem of optimal distribution of information flows can be divided on two missions:

• A mission for accesses methods (when computer can transmit information in the network?);
• A mission for dynamic routing (how the information can be the speediest provided to the receiver);

In order to make such dynamic decision with a global nature, a service mode of information exchange have to be “projected” as in any node to arrive information for entire network condition [4]. On practical reasons, full information for any net node condition is possible to provide in particular cases of small undersized networks only. In the general case, the routing algorithm should to be able to adapt itself in real time scales to conditions of particular net mode [5]. Therefore, it is necessary to resolve contradictions between information, giving net nodes loading and its centralized application on synthesizing optimal routing strategies. In present, the follow math apparatus in modeling traffic processes in computer networks [1,6]:

• Static graph models, where the network is presented as a graph and problems for optimal paths (the shortest, the most reliable, the highest traffic capacity, optimal in relation of two criteria, the shortest paths between any two nodes) or optimal flows (optimal circulation) are resolved;
• A tail-end model by the theory of mass servicing;
• Numerical and analytic description of the network dynamics by differential equations.

The known analytical computer networks models, used in flows projecting, routing, and controlling are the tail-end network models [1,3]. In the references on network models, available network states are focused on which it cans to make precious analysis. The traditional approach of the theory of tail-ends application, defined in the theory of mass servicing, synthesizes the “optimal” routing and strategy of information flows distribution as result of estimated static parameters and defined objective function of the network model. This static optimization problem can solve by the math modeling approaches. A Monitoring station (MS or Network Management Center (NMC)) performs flow-allocation variation for the input traffic variation accounting, network topology, which re-computes periodically the optimal mode. It is a quasi-static approach whereto the network state is discussed as an order of stable states.
 
 

These approaches and ratiocination’s can be applied in full measure for producing statistics of arbitrary Internet server. In Particular, a detailed study of the statistical results and relationships for Netscape enterprise WEB server 3.62 is made. A programme, which can automatically to generate graphical and text interpretations of the log file of relevant WEB server is used. It can visualize in enough degree relationships in a form suitable for user and administrator. On base of these figures, It can to control the site access filter, aiming the most active users to have, in respective, bigger quote for site access.

Below, in graphic form are suggested results of the program implementation as WebTrends Log Analyzer is used.

In conclusion we wish to notice that produced results can use including for the planning, operative control and efficient management objectives of the Special Intended Integrated communications-information systems (CIS). These results, are concerned preferable for functioning above mentioned hierarchical levels of the CIS - SEP and OSC as they allow in real time scales:

• real information charging of the network nodes;
• synthesizing optimal CIS control strategy;
• routing algorithms optimization;
• data base capacity optimization supported in SEP and OSC;
• Comparative analysis performing and selection of the most suitable operating system amongst current the most available ones, such as: UNIX, Windows 98, Windows NT, Windows 2000, etc.;
• Comparative analysis performing on the “functional efficiency” criterion and selection of an optimal database (Oracle, Access, Informix, etc.) control system, used in the SEP and OSP hierarchy levels.

1. Petrova,P., N.Gazibarov, N.Zagorov, Modelirane i optimi-zatsiya na trafika v komunikatsionno-inofrmatsionni mrezhi, izd. PAN – VT, biblioteka “Komunikatsionno-informatsionni sistemi”, V.Tarnovo , 2000, 43 pp.

2. Csopaky,G., T.Moenar, Cs.Szabo, and L.Standi, A traffic generating and monitoring technique for standard and integrated voice/data Ethernet LANs, Computer Networking ed. Csaba L., T.Szentivanyi, K.Tarny, North-Holland, 1990, pp. 231-232.

3. Anantarm,V., S.Verdu, Bits Through Queues, IEEE Transaction on Information Theory, vol.42January 1996, pp.4-6.

4. Douglas E. Comer, Internetworking with TCP/IP, Vol.1: Principles, Protocols, and Architecture. Second edition, Prentice-Hall International, Inc., 1999, pp.109-137.

5. Garbe, K., An experimental OSI Network management implementation for a PC Network, Computer Networking ed. Csaba L., T.Szentivanyi, K.Tarny, North-Holland, 1990, pp.251-253.

6. Bolding K., L.Snyder, Parallel Computer Routing and Communication, First International Workshop – PCRCW’94, Seattle – Washington, USA, May 1994.

7. WebTrends Hand book online – Internet based, http://www.webtrends.com/guide.html, pp.23-35