CHOOSING THE OPTIMAL PLAN FOR ENERGY EFFICIENT REENGINEERING OF CORPORATE COMPUTER NETWORK

Abstract

The paper shows that energy efficient reengineering of a corporate computer network is performed in several stages. At each stage, reengineering is performed for a certain fragment of the network
taking into account the amount of allocated funds for this phase of network modernization. To assess the effectiveness of each stage in the work model to calculate the cost of implementing an energy-efficient reengineering of an arbitrary network fragment and a model for calculating the economic effect of reengineering
are proposed. Calculation of the power consumption of the network is carried out at the level of each switch
port and network card. The power consumed by the switch port and the network card depends of the channel
load to which they are connected. In the work the technology of port and network card load calculation is
offered. The technology consists of two stages. At the first stage, using the instrumental measuring tools, the
average value of the frame rate per port and network card and the average value of the transmitted frame
size are determined. At the second stage, using the queuing theory models, the average frame transfer time is
determined. In the models was made the assumption about the simplest flow of requests for servicing frames
in the communication channel of the network and the exponential distribution of their service time An efficiency indicator is formulated and a dynamic programming method for constructing an optimal
reengineering plan for a corporate computer network is proposed. Reengineering of the network is performed for several stages and at each stage, using the dynamic programming scheme, the optimal network
fragment is selected, which ensures the maximum economic effect from the realized of energy-efficient network upgrade.
The proposed method of choosing the optimal plan can be used as a practical tool for network engineers in the work on energy-efficient reengineering of corporate computer networks