magnus at rubidium.dyndns.org
Thu Jan 7 16:26:52 UTC 2010
Poul-Henning Kamp wrote:
> In message <4B4603E5.4050200 at rubidium.dyndns.org>, Magnus Danielson writes:
>> john.foege at gmail.com wrote:
>>> Now, I'm not sure what kind of testing they did, [...]
>> What better test of server performance as a ten-fold increase in load?
> And at only $100.000 - undoubtedly a lot extra sales, they get the
> testing for cheap.
> Best way to performance test an ecommerce site I've seen.
Indeed. Creating a (tailour made) test-suite can take alot of man hours,
and while it can bash out a number of performance issues, there is no
replacement to real load. Toss in the advertisement aspect (I've
received the news through three independent sources of various forms of
social networks) you have a real winner.
I don't really object to it. It's a win-win kind of solution. We, the
customers, gets the toys cheap when we act like crash-head dummies for
their new site. Modern technology takes new approaches. This is clearly
one of them. But the fundamental concept isn't entrierly new, but the
level gain of doing has not been seen before.
More information about the time-nuts