Certain rules for calculating the optimal lifetime of computing systems, which are based on rates of technological change in the computing industry, are provided in this paper. Particularly, the study claims that the laws of the development and maintenance of hardware and software are similar, but their specific rates and trends vary vastly. This last issue is, at least, controversial, and it is possible that different results could have been produced. However, the analysis of the replacement of computer systems, based on the choice of the providers, the optimization of the system lifetime, and its price/performance ratio in the market, seems rational.
The authors focus their modeling on the last two issues using vintage capital models (VCM). Such models reveal the nature of renovation of computing equipment. To calibrate and simulate the models, the following data is required: the productivity and price of systems, the forecast of these values, the ratio purchasing price/maintenance cost, and the capital depreciation. The authors provide two examples based on two different scenarios: personal computers and high-performance computers.
The paper will be very interesting for systems administrators, as well as those with financial responsibilities, since system lifetime forecasting is a crucial issue for deployment.