Re: gTLD server losses

[Harald Tveit Alvestrand <harald@Alvestrand.no>]

At 12:08 17.04.2001 -0400, Tim Seaver wrote:
>I apologize to VeriSign for mis-stating my gTLD performance results. What
>I was actually seeing in my DNS tests was a regular 4% loss from the two
>gTLD servers closest to me, c and g. However, this loss was due to the
>servers being unresponsive for two 30 minute periods per day during zone
>reloads rather than being overloaded. With the move of g on April 4, it
>no longer shows that behavior, and Matt tells me that c will be moving
>to a new configuration shortly to avoid the reload outages. That's
>good news, since it should remove some 20% of the server retries I see.

A random measurement from a site in Trondheim, Norway, April 15 to now:

[hta@dokka dnsstuff]$ tail -5000 log.log | ./buckets
384 samples
Server               Samp      %  Min  Ave  Max   0 1.0 2.0 3.0 4.0 5.0 6.0
A.GTLD-SERVERS.NET    384 100.00  133  189 5150 381   1               2
B.GTLD-SERVERS.NET    363  94.53  194  351 6665 360                   2   1
C.GTLD-SERVERS.NET    382  99.48  132  161 5126 380   1               1
D.GTLD-SERVERS.NET    384 100.00  133  215 5433 379   1               4
E.GTLD-SERVERS.NET    377  98.18  134  185  383 377
F.GTLD-SERVERS.NET    384 100.00  207  235 1584 383   1
G.GTLD-SERVERS.NET    383  99.74  205  231 1548 382   1
I.GTLD-SERVERS.NET    384 100.00   32   91 5079 381                   3
J.GTLD-SERVERS.NET    381  99.22  309  443 6095 372   1               7   1
K.GTLD-SERVERS.NET    385 100.26   53   93 5068 384                   1
L.GTLD-SERVERS.NET    385 100.26  150  189 6504 384                       1
M.GTLD-SERVERS.NET    355  92.45  375  773 5535 328   2              25

My test seems to be doing 1 retry after 5 seconds.
Apart from B and M, I think they are doing very well.

Same measurement for the root servers:

[hta@dokka dnsstuff]$ tail -5000 arpa.log | ./buckets
500 samples
Server               Samp      %  Min  Ave  Max   0 1.0 2.0 3.0 4.0 5.0 6.0 7
A.ROOT-SERVERS.NET    495  99.00  142  280 5160 482      2          11
B.ROOT-SERVERS.NET    490  98.00  207  602 5486 461      2          27
C.ROOT-SERVERS.NET    479  95.80  140  848 8947 424                 53       1
D.ROOT-SERVERS.NET    499  99.80  133  170 5173 496      2           1
E.ROOT-SERVERS.NET    490  98.00  191  226 5209 487      2           1
F.ROOT-SERVERS.NET    497  99.40  194  278 5214 491          1       5
G.ROOT-SERVERS.NET    448  89.60  145  949 7548 385      1          61       1
H.ROOT-SERVERS.NET    379  75.80  173 1652 6962 286      1          90   2
I.ROOT-SERVERS.NET    500 100.00   33   47  250 500

G and H are not terribly useful to me. I is in Sweden, proving how much
speed-of-light matters in these things; its average is a third of the
minimum roundtrip time of the second best root server.
The reason J-M are not included is that the script actually looks for the 
current root servers of in-addr.arpa....