A Short History of the Reverse Beacon Network
(originally published in the US National Contest Journal, September-October 2012)
In only 6 years, the Reverse Beacon Network (RBN) has become an established part of the contesting universe. No serious assisted or multiop contest station can afford not to use RBN spots, which now are distributed via more than 100 DX cluster nodes worldwide.
CW Skimmer, software capable of decoding many CW signals at once, was released in early 2008. At the time, the author — Alex, VE3NEA — viewed it primarily as a DXing tool, permitting efficient monitoring of pileups. It’s no surprise that contesters soon saw how they could use it to increase their scores, and controversy quickly erupted as to whether contest rules should permit the use of skimmers. After Alex added telnet capability, potentially permitting a local skimmer to feed spots directly to contest logging software, the debate took over the CQ-Contest reflector. Was this “assistance,” like the DX cluster? Should its use be permitted at all?
Meanwhile, the idea of a Reverse Beacon Network arose out of an e-mail exchange in March 2008 between PY1NB and me. Felipe had been running a unique DX Web site, DXWatch.com, for several years. It offered DX spot filtering tools and mapped spotted contacts on a world map. I had been working with VE3NEA since late 2007 to test, develop and refine CW Skimmer. Felipe saw a way in which the basic framework of DXWatch could be adapted to display skimmer spots.
Also in March 2008, Felipe wrote the first aggregator software, intended to receive spots from skimmers telnet servers and transmit them to the Web site for display. The Web site was initially the only way to view skimmer spots, and at first we thought of it primarily as a tool for viewing and studying propagation. We began recruiting contesters and DXers to act as reverse beacons in late March, and the RBN was born.
Through the spring and summer of 2008, controversy raged in the contesting community over whether non-assisted single operators should be allowed to take advantage of this new technology. Just in time for the fall contest season, opinion coalesced around putting CW Skimmer and the RBN in the same category as traditional DX cluster spots, limiting their use in contests to the assisted and multiop categories.
For the remainder of 2008 and through 2009, the RBN focused on developing its network of stations distributed around the globe. Nick, F5VIH/SV3SJ, joined the RBN team in 2010. His computer science background was a great asset, and in March, he rolled out the Signal Analysis Tool, a way to graphically compare signals of multiple stations on multiple bands, as heard by a single skimmer anywhere in the world.
A couple of key hardware and software developments added momentum to the RBN’s development. The SDR-IQ receiver offered by RFSpace (www.rfspace.com), especially when used with W3OA’s SkimScan software, made multiband spotting with a single receiver possible for the first time. Then, Phil, N8VB, released the QS1R software-defined receiver, with a large field-programmable gate array. In summer 2009, VE3NEA released Skimmer Server, software that could simultaneously decode a swath of up to 192 kHz on up to seven bands, using the QS1R. It was, and remains, a programming tour de force.
As more and more of these receivers made their way into the field, the number of RBN stations continued to grow, and because of these developments, the number of spots received grew almost exponentially. By spring 2010, the RBN was feeling serious growing pains. During the ARRL International DX CW that February, the database server was unable to handle the sheer volume of spots, and it crashed repeatedly.
At about the same time — neither memories nor records are quite clear — it occurred to us that it might be beneficial to provide RBN-derived DX spots to the worldwide contest and DX community through a telnet server, using DX cluster software. Initially, we were very concerned that the large volume of RBN spots would “leak” inadvertently into the traditional DX cluster network, provoking a worldwide backlash, but we decided that, if worst came to worst, we could always shut down the server.
The potential benefit of providing spots in a format usable by any contest logging software was just too tempting not to try it, so Nick wrote software to link the RBN’s database server with a Telnet server. In April 2010 the RBN Telnet node debuted, on a server provided by Rick, K4TD. Almost immediately, it proved very popular, to the point where the server quickly reached full capacity. During the 2010 ARRL November Sweepstakes CW, it collapsed under the load, and many spots were never forwarded.
This would not do. As a stopgap, we added a second temporary server, using AR Cluster software and running on a laptop in my shack. In November 2010, just in time for the CW World Wide CW, Dave, KM3T, joined the RBN team, and George, K5TR, contributed an additional server. Nick made important changes to the server and database infrastructure, and Dave, Nick and Felipe worked hard to ensure that the RBN servers would not fail during the contest. They succeeded, and in 48 hours the telnet servers delivered more than 1.7 million spots from 60 to 70 skimmers, without serious incident.
What really saved our bacon, though, was the decision by DX cluster software developers to support “skimmer” and “non-skimmer” modes, so RBN spots could be distributed by many cluster nodes, rather than directly from the RBN. In September 2010, VE7CC and VE1DX began distributing RBN spots through their cluster servers. Shortly after that, AR Cluster version 6 was released in beta with similar provisions and an advanced filtering scheme. In March 2011, a telnet server using AR Cluster version 6, donated by Jamie, W2QO, was added to the RBN’s facilities, spreading the load and allowing for easy distribution of skimmer spots to AR Cluster version 6 clusters worldwide.
In September 2011, Dick, W3OA, joined the team and produced the first Windows aggregator. The beta was a big success, and in succeeding months he delivered increasingly sophisticated versions of the software, which is now in release 2.6, with 3.0 currently in beta. The RBN delivered close to 100 million spots in 2013. Another omen – the number of Skimmers online with the RBN on a typical non-contest weekday rarely dips below 100, and something over 150 different Skimmers were on the network during CQWW CW. .
As far as hardware is concerned, we’re in a period of watchful waiting. At some point the database server will max out. This probably will make it necessary to separate it from the Web server, but we seem to have a little way to go yet.
After CQWW 2013, we replaced the original DXSpider Telnet server with another instance of ARCluster V6, because the older, single-threaded software is no longer capable of handling the volume. Felipe is working on a new GUI for the website, to permit more flexible searching and better performance in general.
While the total number of skimmers is growing satisfactorily, we continue to need more coverage in some areas. 2013 saw Chinese amateurs put three new nodes on the air, including one in Urumchi, close to the most remote spot on earth.
Also, in 2013, the Yasme Foundation graciously funded the establishment of a full-capability node in Bangalore, India. We are currently working with IARU Region 1 to secure one or more additional nodes in equatorial Africa. There is no shortage of things to do.
Probably the most important lesson we all learned is that technological breakthroughs can have entirely unexpected consequences, but that, in the end, the genie cannot be put back in the bottle. Nobody knew that Alex’s CW Skimmer would have such an impact. My e-mail exchanges with him at that time make interesting reading, if only to demonstrate how little we understood where the technology would take us. Each advance in the use of CW Skimmer and the RBN has been due to the willingness of all those involved to “go with the flow.”
Volunteer contributions are the key to the RBN’s success. Each skimmer setup represents an individual contribution to the project, and except for Felipe’s DXWatch server, each of the servers has been donated. We’ve also received hundreds of dollars in PayPal contributions from users around the world. These all have all been plowed back into the cost of expansion.
Coordination of the project would not have been possible without the Internet, and particularly Skype instant messaging, among core participants in France, Brazil and several parts of the United States. When I reflect on how far we’ve come in the last 20 years, I’m amazed. This project would not have been possible two decades ago. All of the pieces — hardware, software, the Internet, and especially volunteers — have made it happen.
Where will we go from here? Stay tuned!