I was a bit puzzled by the numbers we got from netsim friday because they didn't match with with the earlier partial replication results I had from my browser testing. I ran some new numbers and updated the source of the frequencies. Basically the numbers before were skewed because they were from a world with a lot more pubs in the social graph, so they inflate the contact numbers. Another thing is that the frequencies before did not properly account for other types. These are roughly 10%, so significant enough to include especially for our partial replication tests. I believe with this change and the blocking bug from https://github.com/ssb-ngi-pointer/ssb-replication-scheduler/pull/5#issuecomment-938392938 fixed we should get a lot better numbers.
Sorry for the lateness of this message, it was just bugging me, so had to get it out of my system :)
I was a bit puzzled by the numbers we got from netsim friday because they didn't match with with the earlier partial replication results I had from my browser testing. I ran some new numbers and updated the source of the frequencies. Basically the numbers before were skewed because they were from a world with a lot more pubs in the social graph, so they inflate the
contact
numbers. Another thing is that the frequencies before did not properly account forother
types. These are roughly 10%, so significant enough to include especially for our partial replication tests. I believe with this change and the blocking bug from https://github.com/ssb-ngi-pointer/ssb-replication-scheduler/pull/5#issuecomment-938392938 fixed we should get a lot better numbers.Sorry for the lateness of this message, it was just bugging me, so had to get it out of my system :)