Systems & Software Talk | |
"Other or Supplier" BottleneckRecently I ran a performance test that left a few of we the test watchers/analyzers scratching our heads. I scratched a few more hairs off my own and revealed even more bare skin. This test was executed across a full DS3 hose - 45 Mbps. The results, specifically the bandwidth/throughput chart - showed a flattline at 27Mbps, despite increasing simulated user load that graphically stairstepped through and above that line. Investigation showed we had full DS3 locally. So, we called our internet provider and asked if they had any issues during the time frame of the test. They replied, "None!" Twenty four hours later, we ran the same test. We went above and beyond the previous night's 27Mbps "limit" on the hose to the true demand of the test at about 38Mbps. Mysterious? Yah, you betchah! Did our provider think we would never know we were getting less than what we were paying for? Probably! Learnings:
|
About MeMy Profile Archives Friends My Photo Album LinksCorey GoldbergEffective Testing? Bj Rollison I.M. Testy Blog Alan Page: Software Testing & Rants Dmitry's LoadRunner and QTP Blog Veterans History Project Air Traffic Control Watch Music Making Fun My home 1972-1975 CategoriesFunctional TestingPerformance Horror Development Performance Testing General Tools Tips Warped Humor LoadRunner Tips and Tricks Recent EntriesThey Need To Test More...Software Disorder LoadRunner (tm) & RTE 4 Func/Regr LoadRunner (tm) Random Think time Function Are Rock, Paper, Scissors-based Decisions Obsolete? FriendsLauraScharpphilk10 richardw100 aalhait jimhazen strazzerj Lynnem bru EklecticTester jgottlieb leakybrain michaeljf prainbow rajeshmathur rstens Yury zeeslo whollymindless SyndicationRSS Site Feed |