Report on NCF Usage and the impact of Time Limits by Ian Allen Discussion follows the graphs. Annex port usage in February (before binary news groups were removed): *************************************************************************** * * * * * * * * * * * * * * * * * *** * ********************************************** ** *** ***** * ** * * * * * * * * * * * * * * * * ** * * ** * * * * * * * * * * * * * ** * * * ** * ** ** * * * * * * 0h 2h 4h 6h 8h 10h 12h 14h 16h 18h 20h 22h * Annex day=0123456 mach=123 95/02/01-95/02/28 0 -x- 24 0 -y- 200***** Modems are busy from 8am onward. Annex port usage in the month after binary groups were removed: *************************************************************************** * * * * * * * * * * * * * * * * ***************************** ***** *********************** * ** ** * ** * * * * * * * * * * * * * * * * ** * * ** * * * * * * * * * * * * * ** * * * *** * ** ** * * * * * * 0h 2h 4h 6h 8h 10h 12h 14h 16h 18h 20h 22h * Annex day=0123456 mach=123 95/03/15-95/04/15 0 -x- 24 0 -y- 200***** Modems are now busy from 3pm onward. The removal of the binary groups appears to have freed up modems before 3pm. Annex port usage in the month after time limits were installed: *************************************************************************** * * * * * * * * * * * * * * * * * * **** ************ *** ** ********* ** * ** ******* * * * * **** ************** * * * * * * * * * ** * * ** * * * * * * * * * * * * * ** * * * ** * * ** ** ** * * * * * * 0h 2h 4h 6h 8h 10h 12h 14h 16h 18h 20h 22h * Annex day=0123456 mach=123 95/04/18-95/05/18 0 -x- 24 0 -y- 200***** Modems are busy around 4:30pm and from 8pm onward. Many modems are idle during the morning and afternoon before 3pm. Annex port usage two months after time limits were installed: *************************************************************************** * * * * * * * * * * * * * * * * * ** *********** *** ******* **** ** * ** * ***** * * * ********************* * * * **** * * * * * ** * * ** * * * * * * * * * * * * * * * * * ** * * ** ** ** * * * * * * 0h 2h 4h 6h 8h 10h 12h 14h 16h 18h 20h 22h * Annex day=0123456 mach=123 95/05/18-95/06/18 0 -x- 24 0 -y- 200***** Modems are busy from 8:30pm onward. Many modems are idle in the mornings and afternoons before 3pm. Annex port usage three months after time limits were installed: *************************************************************************** * * * * * * * * * * * * * * * * * *** ******** ** **** ********************** ***** ** * * **** ******** * * * * * * * * * * * **** * ** * * ** * * * * * * * * * * * * * * * * * **** * ** ** * * * * * * 0h 2h 4h 6h 8h 10h 12h 14h 16h 18h 20h 22h * Annex day=0123456 mach=123 95/06/18-95/07/18 0 -x- 24 0 -y- 200***** Modems are busy from 9pm onward, early morning use is decreasing, daytime (noon/aftenoon) use is increasing. Many modems are idle aroud 5-7pm. Time Limits Discussion: The instability of usage patterns with the change to summer hours makes prediction of Fall patterns hard. Without a year of time-limit data with which to compare the current graphs, we have no idea what will happen in the future. With daytime use of the modems increasing, removing the peak-time disincentive from this period of the day might be premature, especially if policy required that the disincentive be reinstalled in a month or two. I suspect people will be quick to forget the reduction of peak-time period and quick to criticize its re-installation. ----------------------------------------------------------------------------- ===== Usage ===== Count of Annex non-guest logins: *************************************************************************** * * * * * * ** * * * ** *** ** ** ** ** ****** * * ** ** * ** * ** ***** ** * * ***** ** ** * * ** * * * ** ** * * ****** ** ** * * ** * ** ** * ** * ** * * * **** * *** * * * * * * * ** * * * ** * ** * * ** * ** ** * ** * ** * * * * * * * * ** ** * * * * * * ** ** * * * * * * Mar13 Apr1 Apr21May1 May18 Jun1 Jul1 Jul16 * Col 5 UAnnx 0 -x- 150 0 -y- 10000*********************************** The number of logins per day on the Annex increased, and has stayed elevated, with the introduction of time limits. The number does not appear to be increasing with the increasing number of NCF accounts. Count of Telnet non-guest logins: *************************************************************************** * *** ** * ** * * ** * * ** * * * ** * * * ** * ** ** ** * * ** * * ** * ** * * * *** * ** ** ** ** * * * ** * * ** ** ** ** ** ** *** ** *** ** ** * * ** *** ** ** ** * * ** * ** * ** * ** * * ** * * ** ** * * ** * ** ** * * * **** * ** * ** * ** * * ** * * ** ** * * ** * * * *** * ** * ** * * * ** * ** ** ** * * ** ** ** ** ** ** ** * * * * ** ** * ** * * ** ** * ** ** ** * * * * * * ** ** ** ** * ** * ** * * * * * * * * * ** * ** * ** * * * * * ** * * * * * ** ** * * * * * * * * * * * * Mar13 Apr1 Apr21May1 May18 Jun1 Jul1 Jul16 * Col 6 UTeln 0 -x- 150 0 -y- 6000************************************ The introduction of time limits appears to have reduced the number of times telnet users connect to NCF. Count of Annex Guest logins: *************************************************************************** * * * * * * * * * ** * * ** * * * * * ** ** * * * **** * * * **** **** * ** ** * * * * * * * ** *** * *** ** ******* ** ** * ** * * * * ********** ** *** *** * **** *** * ** ********* ** ** *********** ** * * * * ** ** ** * ******** * * * * ***** * * * * ** * * * * * ** ** * * * * * * * * Mar13 Apr1 Apr21May1 May18 Jun1 Jul1 Jul16 * Col 8 GAnnx 0 -x- 150 0 -y- 800************************************* The introduction of time limits caused a brief rise in guest account Annex usage, then a return to normal. Count of Telnet Guest logins: *************************************************************************** * * * * * * * * * * * * * * * ** *** ** * * * ** ****** * **** * * ** * * * * ** * * ** * ** * * * * ** ** ** * * * ** ** ** * * ** ** ** * * **** ** * * * * * ** * ** * ** * ** ** ** * * * * *** ** * * * * * * ** * ** * ** * ** * * ** ** * * **** ** ** ** * * * * * ** * * ** * * ** ** * * ** * ** ** *** * ** ** ** ** ** ** ** * * ** * * * ** * * * ** ** * ** * ** * * * * * ** * ** * ** * * * * * * * * * * * Mar13 Apr1 Apr21May1 May18 Jun1 Jul1 Jul16 * Col 9 GTeln 0 -x- 150 0 -y- 1500************************************ Guest Telnet login counts are dropping. Count of unique NCF users over a two-week period: *************************************************************************** * * * * * ****************************************** * * * ** **************** * ** * ** * * * * ** * * * * ** * ** * ** ** * * * * * * * * ** ** * * * * * * * * Mar13 Apr1 Apr21May1 May18 Jun1 Jul1 Jul16 * Col 18 Win14 0 -x- 150 0 -y- 20000********************************** The number of unique NCF users is dropping. The above graph is for a sliding window two weeks long, ending on the dates shown. Fewer unique users used NCF in any two weeks in June or July than in any two weeks in April or May. The number of NCF accounts was 40,513 on April 5. It is now 44,721. The effect of this 10% increase in accounts is absent from the above graph; more users are leaving NCF than are joining. Without annual data, we can't say whether current trends are due to seasonal, policy, or other factors.