General notes

• The timing results depend heavily on the load of the machine on which the measurements are performed. Therefore the absolute times are not as illustrative as their relative values.
• The '-p' switch was used with the 'time' command, to get the times in seconds, not in minutes and seconds.
• The following tools have been used for data analysis and presentation:
  • a few shell scripts - to run the experiments in the batch mode;
  • a Perl script - to parse and reformat the output of the 'time' command;
  • a C program for statistical analysis of the data;
  • a few Matlab programs - to generate the graphs, and to do the least squares approximation for some data sets.
• The timing data presented in this report is just a fraction of the total amount of the experiments' results. Timings of the intermediate versions are available if needed.

Table of contents:

• Part 1: Java RMI timing
  
• Part 2: File synchronizer implementation improvements
  
• Part 3: File synchronizer security