To whom it may concern, Dr. Sugih Jamin and myself (Patrick McDaniel) of the University of Michigan are currently studying the load characteristics of a potential Internet sized key distribution system. Because of the limited amount of these kinds of systems, we are using DNS traffic to model the expected usage. To continue this study, we are in need of DNS traffic traces collected from a wide range of networks. The purpose of this post is to ask for your assistance in this effort. We provide a script for collecting this data. This script will: 1) Automatically collect traces from the named process over a 24 hour period. 2) "Anonymize" all the data collected. No domain or host names will appear in the traces after they have been filtered by the downloaded script. At your request, we will not divulge the source of the trace in any resulting publications. We have submitted a initial version of the study to SIGCOMM 98'. The submission paper is available in at: http://www.eecs.umich.edu/~pdmcdan/docs/keyhier.ps.gz The script used for collecting the data is available at: http://www.eecs.umich.edu/~pdmcdan/automated-dns Instructions for script usage: 0. Login to the authoritative domain name server for your network. (note that you will need to perform this as root) 1. Copy the downloaded automated-dns script to the /var/tmp directory. 2. CD to /var/tmp 3. Run the script with the following arguments: automated-dns <boot file> <named pid> <output-file> Where: boot file : is boot file for DNS server named pid : process ID of named output file : file to output trace information eg. % automated-dns 517 ./dns-trace 4. The script will ask you to confirm the PID of the named process. 5. The script will ask for the signal numbers of the SIGUSER1 and SIGUSER2 signals for your system. This information should be available on the man page for signal (section 7 on Linux, section 5 on SunOS 5.5.1, signals 16 and 17 are most likely correct) 6. The script should execute for 24 hours, then "anonymize" the trace and write results to the output file. 7. Email the output file or contact me to exchange the data. Please include as much information as you can about the size and type of network on which the traces were collected. My email address is: pdmcdan@eecs.umich.edu Thanks for your time. If you have any questions, please feel free to contact me through email. If you run into any problems, I will be more than happy to help. Thanks, Patrick McDaniel ---- Patrick McDaniel University of Michigan Office : 3115 EECS Building [ \/ ] Office Phone : (313) 647-3780 ][\/][ Email : pdmcdan@eecs.umich.edu [__][__]