-
Notifications
You must be signed in to change notification settings - Fork 14
Sec 5.1 #92
Comments
@lisongxu, any update? |
@larseggert sorry for the delay, will catch up soon |
Thanks, @markkukojo You are right that the analytical model "gives average window size with fixed random loss probabilities and unlimited bandwidth." This demonstrates to some extent the AIMD friendliness (now referred to as Reno Friendliness in the document). The original Cubic paper [HRX08] shows the experiment results when Cubic is competing with AIMD/Sack. Figure 10 shows that Cubic is friendly to Sack in low bandwidth or low RTT networks (although not exactly fair share). For example, in Fig 10(a), bandwidth=400Mbps, RTT=10 or 20ms; also in in Fig 10(b), bandwidth=10 -- 400 Mbps, RTT=10ms; Cubic is friendly to AIMD/Sack. |
@markkukojo, please comment on the above? |
No response from @markkukojo in three weeks, closing. |
Markku Kojo said:
The text was updated successfully, but these errors were encountered: