Mike McGrath wrote:
Very true. From the data we've gathered at tummy, though, I'm not nearly as concerned about throughput as latency. Latency makes every connection to the database or to a proxy server that's not in the local colo much more expensive no matter how large or small the data returned. With tummy, in fact, I know that making fewer requests trumps the amount of data returned by a large margin.latency and throughput measure different things though, we'll have to test multiple things.
The bugzilla script I optimized today to work from tummy changed how I cached data from FAS. Before I was storing people information in a cache whenever a new person was found (This hits almost everyone in cvsextras once and only a handful who are not in cvsextras). The new cache requests all of the people in FAS in two requests (getting different information about all the users from each request). The old script took hours to complete. The new one finished in under 8 minutes. (Even when the script accessed the fas1 database directly it took hours to complete... although that ran from the Red Hat Boston office, rather than tummy).
-Mike On Wed, 23 Apr 2008, Toshio Kuratomi wrote:Answering myself, ping times from app4 (phx-Arizona), app5.vpn (tummy-Colorado), and app6.vpn(telia-Germany) to db2(phx-Arizona) [toshio app4 ~]$ ping db2 PING db2.fedora.phx.redhat.com (10.8.34.120) 56(84) bytes of data. 64 bytes from 10.8.34.120: icmp_seq=1 ttl=64 time=0.249 ms 64 bytes from 10.8.34.120: icmp_seq=2 ttl=64 time=0.285 ms 64 bytes from 10.8.34.120: icmp_seq=3 ttl=64 time=0.764 ms 64 bytes from 10.8.34.120: icmp_seq=4 ttl=64 time=0.280 ms [toshio app5 ~]$ ping db2 PING db2.vpn.fedoraproject.org (192.168.1.19) 56(84) bytes of data. 64 bytes from 192.168.1.19: icmp_seq=1 ttl=64 time=43.7 ms 64 bytes from 192.168.1.19: icmp_seq=2 ttl=64 time=42.9 ms 64 bytes from 192.168.1.19: icmp_seq=3 ttl=64 time=42.9 ms 64 bytes from 192.168.1.19: icmp_seq=4 ttl=64 time=42.9 ms [toshio app6 ~]$ ping db2 PING db2.vpn.fedoraproject.org (192.168.1.19) 56(84) bytes of data. 64 bytes from 192.168.1.19: icmp_seq=1 ttl=64 time=196 ms 64 bytes from 192.168.1.19: icmp_seq=2 ttl=64 time=181 ms 64 bytes from 192.168.1.19: icmp_seq=3 ttl=64 time=184 ms 64 bytes from 192.168.1.19: icmp_seq=4 ttl=64 time=186 ms So each request from an app server in phx to the database or to another app will be 200x faster to set up than from an app server at tummy and nearly 1000x faster to set up than telia. If we deploy apps at telia it will be essential to minimize the number of requests if we want to keep things fast. -Toshio_______________________________________________ Fedora-infrastructure-list mailing list Fedora-infrastructure-list redhat com https://www.redhat.com/mailman/listinfo/fedora-infrastructure-list
Description: OpenPGP digital signature