To launch DXA click the imaabo

CE HOME KURE HOME DXA HOME CONTENTS THUMBNAILS INDEX

---

OVERVIEW

ORGANIZATION

PERSONNEL

SCHEDULE

CORPORATE

SPONSORS

SUPPORT US

DXA PAGE

QSL ROUTE

CONTACT_US

The K7C project had two main goals: (1) To give emphasis to logging European stations; (2) To implement a new near-real-time online system called DXA.

RADIO OPERATIONS

Radio Operations of the K7C Kure Atoll DXpedition commenced on 25 September 2005 at 0712 UTC and ended on 05 October 2005 at 1000 UTC. The following table gives a preliminary set of statistics of the radio log:

The QSOs were distributed in Zones as follows

The worldwide distribution of these QSOs is as follows:

The distribution in modes was:

The number of unique callsigns in the K7C log is 15,472 (30.9%). No 6m QSOs were logged, although K7C fielded the gear and was ready.

DXA OPERATIONS

The DXA system was assembled onsite Kure simultaneously with the K7C radio stations. The three sites (CW, SSB, COM) were linked by a 802.11 WIFI network. All stations used the N1MM logging program, and all stations could see the entire log simultaneously. The COM station was connected to the internet via an Inmarsat link, dialed by the computer as a normal modem. Data rate was 64kbit/s. Data was packetized and uploaded at intervals between 1 minute and 5 minutes. A dedicated high-speed server was implemented at the DXA website to serve an expected simultaneous 5000 clients. The DXA website was maintained by K6SGH assisted by VE7AVV, and supported by the team of K7C Relay Stations.

The radios and DXA systems were implemented at K7C simultaneously. KK6EK was dedicated to the onsite DXA operation. When K7C went QRV, the DXA clock was started, and at QRT it stopped at 10 days, 14 hours, 47 minutes. The TOTAL QSOs counter incremented as each packet was received. The CURRENTLY ACTIVE bandmode table showed red "ON" lights for each bandmode logged in the current packet. Thus, with a 5-muinute upload interval, the CURRENTLY ACTIVE table meant "currently active within the past 5 minutes." If a user logged in with his callsign, he was assigned his personal bandmode table. Whenever he was logged by K7C, a green square was lighted for that bandmode.

The number of users logged in rise sharply in the 4-5 days before K7C went on the air. On the first day of operation , the number of users apparently exceeded 50,000, causing a complete lockup of the server. The result was twofold: the users could not see the updates of the DXA window, and K7C could not upload new data to the site. Even when the server was rebooted repeatedly, it quickly saturated, and went down again.

The defense against the unexpectedly high demand was to reduce the amount of the content served by DXA for each request. The first step was to disable several options, such as views of the Earth and the grayline map. Next, the size of certain files were reduced. Also, one component of the display was eliminated (the Last Hour callsign listing). Finally, the upload interval of 1 minute was lengthened to 2 minutes, then to 5 minutes. These measures reduced the volume of served content and the rate of serving, and together with a naturally reduced user curiosity, brought the demand within the capability of the server, around the third day of operation. From then on, with minor interruptions, the DXA system worked well and reliably. With the extremely low data rate available at K7C, it proved impossible to catch up with some missing green squares.

The DXA dataset was updated on Oct. 21, 2005. Therefore, as of this date, the "green squares" should correctly represent the K7C log entries. If you find any discrepancy between your log and the display on DXA, please commmunicate it to KK6EK@cordell.org. Please understand that I cannot change the log--I can only change the DXA display. Any discrepancies in the K7C log will be resolved by the QSL manager.

The full statistics of the use of the DXA website are listed here. A monthly listing is presented in this table.

The following table shows the overall statistics of the KURE/DXA website:

Number of unique callsigns logged into DXA	>45,000
Number of unique IP addresses requesting content	>64,000
Visits per day (typical)	20,000
Total number of visits	230,000
Content served per day (typical)	30 GB
Total content served	284 GB
Hits per day (typical)	5 million
Total number of hits	43 million

These statistics suggest the following:

Since there are approximately 20,000 "serious DXers" in the world (those who appear in most DXpedition logs), probably nearly every one of them was looking at DXA. In addition, there were perhaps 45,000 additional "casual DXers" who were sufficiently motivated to log into DXA, perhaps in anticipation of a QSO with K7C. Taken in total, apparently more than 64,000 people worldwide looked at DXA. Almost all the time, about 20,000 hams were watching it.

There are about 3 times as many unique callsigns logged into DXA (45,000) as entered in the K7C radio log (15,000), indicating that for every successful DXer, there were two more who tried but did not succeed inmaking a QSO with K7C.

The implementation of  DXA on a high-value DXCC target such as Kure Atoll was risky, and the patience of the DXing community when the demand exceeded expectations is appreciated. We would welcome your comments and suggestions. Please send them to k7c@cox.net or KK6EK@cordell.org. If you are part of a future DXpedition that might be interested in implementing DXA, KK6EK would be happy to share the experience at K7C.

ACKNOWLEDGEMENTS

The K7C team is very appreciative of the sponsors and supporters tha tmade this project possible. Particularly generous were the Northern California DX Foundation (NCDXF), ICOM, and the ARRL. Please look at our SPONSORS page, and if you are motivated, express your thanks to them.