Investment and Risk Management Analysis of Proactive as against Reactive Network Maintenance

This study analyses the factors affecting investment decisions concerning end-user support (EUS), which is even more complex than that of centralised data processing (DP). The alternatives of proactive and reactive end-user support (EUS) in a modern LAN environment are compared by analysing the failure risk of one specific component (HUB) of the LAN. The additional preventive actions and management costs are considered as well. Both proactive and reactive EUS are analysed in terms of investment and risk management. Arguments for and against enhanced support capability and a fault tolerant system structure are sought for optimising risk/spending levels. The reliability theory with its demand patterns demonstrated by electronic components has been used in the spirit of inventory-production theory in order to quantify the risk both with deterministic and random failure intensity parameters. Exact Bayesian tests have been developed in order to evaluate the two vendor MTBF figures, only the first of which turns out to match the observation. Bayesian posterior failure distributions have been estimated both in the deterministic and random failure intensity cases, the last of which results in a binomial distribution with an annual mean of 16 HUBs and a standard deviation of 4 HUBs. The cost parameters estimated by NOKIA from the field have been used, and the costs of alternative policies have been assessed. The annual expected reactive cost, with a lower hourly developer cost option of kFIM 563, exceeds the annual proactive cost of kFIM 552. An additional investment, i.e. systems for network monitoring equipment of kFIM 210 with a three-year pay-back period can be recommended. However, the expected reactive cost would not carry the investment where proactive policy would require an additional management effort compared with the reactive one. The additional preventive actions, costing kFIM 672 annually, could be covered only by assuming that developer capacity is a sales-limiting facxtor. Recommendations and sensitivity considerations can easily be provided for the day-to-day decision-making of IT and SBU management, using the developed calculation models with updated cost and failure statistics data. A key notion of the study is to develop a management tool for NOKIA’s DP management.