the power of broadcast
[a revolution in mass-scale content distribution]
the k2b_measure channel
each kast beta 1.0 distribution comes pre-subscribed to the k2b_measure channel. This channel is not meant for "real" content or recommendations---we will be using it to measure konspire2b's network performance during the beta test.
if you do not want to take part in these measurement tests, you can opt-out by unsubscribing from k2b_measure
how the experiment works
we want to know if konspire2b's real-world performance matches our theoretical model. According to the model, a channel owner should be able to reach (2^N)-1 subscribers by sending a file directly to only N of them, and the entire distribution of the file should take only N time steps (where one time step is the time it takes for one of the subscribers to download the file). For example, with a file that takes 5 minutes to download, a sender should be able to reach 100 subscribers by sending it directly to only 7 of them, and all 100 should have the file after about 35 minutes.
in the real world, however, there are many factors that interfere with the theoretical performance of a networked system. Our experiment is designed to test how much these factors degrade performance.
to measure performance, we will wait until the network seems "busy" (many people online), and then send a reasonably large file (probably about the size of an average mp3) out on k2b_measure. Each kast node that receives a file on k2b_measure is set to send us a small message when this happens. By tracking when we receive these messages, we can measure how long it takes for "everybody" to get the file that we send.
if you participate in our experiment, your node will send us a message every time you receive a file on k2b_measure. We do not track your IP address or any other information about you: we simply note that one more person received our file.
however, you may still not feel comfortable with your node sending messages back to us. If this is the case, please unsubscribe from k2b_measure.
we do not consider this part of kast to be "spyware", since we are not collecting personal information. This experimental part of kast will be removed in the final public release.
participating in this experiment will help us improve the performance of konspire2b. We hope that many people will remain subscribed to k2b_measure, despite these minor privacy issues.