On Jul 24, 8:54 pm, Alex Gaynor <alex.gay...@gmail.com> wrote: > "Seen under this light, the additional 150ms latency resulting from non- > persistent DB connection is huge - it implies almost 30% fewer > customer orders. And it has nothing to do with traffic." > > I'd just like to take a moment to point out that that simply *cannot* > be, else the only logical conclusion would be that .5s of latency > results in 0 sales, which plainly makes no sense.
Actually it doesn't mean that at all. First of all, the 10% decrease in order for each additional 50ms increase in latency is *multiplicative*, NOT additive. For the 150ms increase in latency that non-persistent connection costs you, the decrease in order is 1 - 0.9^3 = 27.1%, hence I wrote "almost 30%" Hence for a 0.5s increase in latency, *if the relationship still holds over a range that big*, the corresponding drop in order would be 1- 0.9^10 = 65%. Which isn't exactly impossible. But my guess is for increase in latency that large, the original "50ms results in 10% drop" relationship will no longer holds. Regardless, the point is that latency is a big deal whether your site is large or small. Even a small latency increase can have a big impact on user behavior. Telling every mom & pop site owner out there "Go set up pgpool or hack the Django signal handler if you don't want high latency" is not very user friendly. Especially when the solution (providing optional persistent DB connection) is relatively simple. --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Django users" group. To post to this group, send email to django-users@googlegroups.com To unsubscribe from this group, send email to django-users+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/django-users?hl=en -~----------~----~----~----~------~----~------~--~---
