Our client software randomly selects a server to connect to. There are four separate addresses the software randomly chooses from. Two addresses are "old" and two are "new". As many firms need to explicitly allow outbound access to IP addresses, there will be some firms (users of some firms) that cannot connect to the "new" addresses.

We record connections over time. Based on when the new addresses were introduced (June 2012), I'm interested in finding a means of indicating which firms (users of those firms), since June 2012 to now, are unable to connect to the new addresses.

(We have no way of explicitly determining this. I am wondering if we can infer it with a probability model of some sort.)

So far I have done the following:

  • Retrieved connection history of all individual users since June 2012
  • Done a count of the number of times each user has connected to one of the "old" addresses
  • Discarded any users from the list who have connected to a new address at least once (since this would suggest there is no problem connecting to new addresses)

In the end, I have a list of users who have only connected to the old addresses, and how many times they have done so.



Your Answer

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

Browse other questions tagged or ask your own question.