My ingress is only slightly higher at 463GB, but I have 1.5TB egress. I also have more than 15TB of data stored on that node though. This matters a lot. But it doesn’t explain this big of a difference.
No one can guarantee that, there are too many variables and connection speed is just one of them. I personally highly doubt that this is actually the bottleneck in this case. We’ve seen node operators with lower upload speeds have better performance than that. As @Alexey pointed out before, the speed that matters is the speed from receiving the data request to sending the full piece. This includes upload speed, but it also includes HDD speeds and anything else impacting that process.
It is for 2 nodes that were receiving data during the same time period though. For the most part at least. @Stob’s node provides a useful comparison for this as it is the same amount of months old.
I just noticed something comparing the results of the success rate script from both @peem and @stob. @peem has around 30k downloads initiated and @stob has 300k. The success rate is close, but the number of total downloads to begin with is 1/10th. I can’t really explain how that could possibly be. Anyone else have an idea?