Sunday, March 21, 2010

Load? What Load??



When I wrote the sad story of our server buckling under the load of the ARRL DX Contest, I'm sure people were curious about the quantitative story. Because of the problems that weekend, we don't have the numbers for it, but we do have them for the Russian DX Contest, which just finished.

As you can see, we went from a rate of around 50 spots per minute received at the RBN, in the time just before the contest started at 1200Z, to a peak of 600 spots per minute at 1400Z. There was a brief drop-out between 0300 and 0400Z due to problems with a backup routine, but then the flow of spots continued at a rate of 250-400 spots per minute until the end of the contest. At that point the server breathed a sigh of relief, as it dropped back to around 30 spots per minute.

For the 24 hours of the contest, the database received 545,747 spots. Our guesstimate is that a major contest such as CQWW or the ARRL DX Test, with their much larger number of participants, would yield at least twice the spot rate. The programming team has done a lot of good preliminary work, to stabilize things at the current level, but the effort continues. The next big test will be the WPXCW at the end of May.\

73, Pete Smith N4ZR

Monday, March 15, 2010

Tell us what you'd like in the database search interface.

While Felipe and Nick are doing the real work, I'm inspired to try to stimulate some interest by raising some ideas about improving the user interface for the Reverse Beacon Network database.

It might be possible for there to be a single interface for all searches. Alternatively, there may be good reasons, such as simplicity, to break it into multiple forms, depending on the purpose. What I'm trying to do here is suggest some things that it would be good to keep in mind for various fields:

For example:

DE should provide for specifying a variety of different limiters - they might include the specific callsign(s) of the spotters as well as countries and ITU zones. Multiple callsigns or countries should be accommodated.

DX should have the same range of possibilities.

Any Date field should allow for date ranges as well as single dates

The Signal Comparison Tool should, if at all possible, only offer Skimmer sites that have data relating to the comparison that is being run. As it stands now, when you specify a date, the next step in the tool gives you a list of Skimmers active on that date, but many will not have data for the stations or bands of interest for any given search. It would be great if the skimmers could be offered for selection based both on the band and on the stations to be compared, but I realize that may be too burdensome; even just the bands would be a step forward.

What would YOU like to see?

73, Pete N4ZR

Progress Report

Since I last wrote, Felipe, PY1NB and Nick, F5VIH, have been hard at work putting together a development environment that they (hopefully joined by other contributors as well) can use to develop better software for the RBN site. Stay tuned for more details on that in the coming days, including an invitation to participate.

In the meantime, they have made some changes in the current software that should, we hope, make it better able to handle heavy loading. The Russian contest, coming up next weekend, should provide a good first test.