Rough Tickrate analysis of EU/NA servers.

I've seen multiple posts on NA servers being more responsive for EU servers even for those in the EU. So I decided to see if I could measure any difference in the performance/tickrate of the servers in both regions.

I used wireshark to log the number of packets received and sent in the first 5 minutes of 10 games on each region. I would land somewhere on the edge of the map & go AFK for 5 minutes with wireshark running.

Here are the results:

Region IP Location Up pps Down pps
NA 107.21.85.181 Virginia 32.4 12.8
NA 54.87.188.40 Virginia 35.1 15.4
NA 52.206.113.93 Virginia 35.8 15.7
NA 34.201.70.93 Virginia 38.5 17.3
NA 34.205.131.2 Virginia 37.8 13.5
NA 54.164.139.243 Virginia 37.3 13.7
NA 34.201.25.100 Virginia 37.9 17.2
NA 34.201.241.204 Virginia 34.3 13.7
NA 54.164.139.243 Virginia 32.3 12.7
NA 34.201.241.204 Virginia 30.7 12
Tickrate: 35.21 14.40
EU 52.29.119.132 Frankfurt 35.9 13.6
EU 52.57.236.220 Frankfurt 34.5 16.4
EU 52.59.155.199 Frankfurt 33.9 12.3
EU 52.57.165.120 Frankfurt 30.6 10.1
EU 52.29.203.46 Frankfurt 30.8 11.1
EU 52.59.165.171 Frankfurt 29.4 11.9
EU 52.29.119.132 Frankfurt 34.6 15.9
EU 35.157.224.170 Frankfurt 34.9 14
EU 52.57.249.193 Frankfurt 26.5 9.4
EU 52.59.155.199 Frankfurt 25.5 10.4
Tickrate: 31.66 12.51
CSGO Valve DM 63 71.3

   

  • Firstly I recorded 5 minutes of a CS:GO Valve deathmatch to test if my recording method was working correctly. On a 64 tick server Wireshark reported 63 packets sent per second & received ~71 packets per second, which correlates with the known tickrate.

  • Assuming the PUBG is targeting a tickrate of 40hz client->server & 20hz server->client the number of packets sent and received is quite significantly under that. Particularly on the server->client side where the lowest measurement was less than half of the target 20hz tickrate.

  • In Battle(non)sense's analysis (https://www.youtube.com/watch?v=zjIJTantgWE) Chris measured, using a similar technique, a tickrate of 40/20hz during closed beta. The above measurements would indicate that performance has dropped since that time.

  • The average tickrate of EU servers was slightly less than NA over the runs I tested. This is despite me connecting from the EU.

  • Over 20 games, I was connected to the same server instance 4 times (Assuming each instance has a unique public facing IP). With 100 instances running there would be a 7% probability of this happening. 50 instances ~30%, 25 instances 75%.


Original post

One Reply to “Rough Tickrate analysis of EU/NA servers.”

  1. What’s up,I read your blogs named “Rough Tickrate analysis of EU/NA servers. – Rebelgame” on a regular basis.Your story-telling style is awesome, keep up the good work! And you can look our website about تحميل اغانى شعبى.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.