1

I was considering running a small test using Bitcoin transactions when I ran into the stress-test / DDOS that Coinwallet.eu conducted in July 2015 (and maybe September too, I am not really sure) where they bombarded the network with small transactions.

Were there any similar tests done? Are their results available? I am trying to find more information on what happen to transaction propagation times, block propagation times, etc, in presence of high volumes of transactions. Pointers to the right directions will be highly appreciated.

P.S. I am studying the behavior of crypto networks, so testnet tests are not really representative, in general, because of their small scale and different bandwidth characteristics, to the best of my understanding.

1

The 2015 stress test aimed to hit 200MB of backlogged tx, they only hit 15MB as reported by coindesk.

Blockchain.info reports that the backlog of tx hit 120MB May 16th 2017. It's mostly been above 20MB for the last 60 days as of this post.

How big of a stress test is big enough for you?

  • I am actually less interested in the amount of backlog created, and more interested in any measurements of transaction propagation in high volumes. One example is how long does it take for a new transaction to reach an arbitrary node at the when a lot of transactions are flowing through the system? Sorry if I did not present the issue clearly enough - I will edit. – UriCS Jun 19 '17 at 18:36
  • Sounds like something you'd likely need to gather for yourself. I don't know of an open source that publishes latency between nodes on the btc network. To do this you'd likely want to set up many full nodes around the world, sync their clocks as close as possible then just log all of the tx that enter the mempool and their tx time. – placeybordeaux Jun 19 '17 at 18:39
  • Agreed. Actually, that's exactly what I have been working on when I stumbled upon Coinwallet's stress-test, which made me wonder if any other similar test was conducted. I have received very detailed information about the network characteristics from the IC3/Cornell people, which I am going to utilize, but I was wondering if any similar tests were conducted. This seemed like a good place to ask... Thanks for your help and time! – UriCS Jun 19 '17 at 18:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.