On Wed, May 11, 2011 at 4:18 AM, Brandon Wirtz <[email protected]> wrote:
>
> What is the Sin coefficient for the pattern of traffic as expressed over what 
> period?
>
>
>  < interesting graph snipped >
>
>
> Both of these use about the same amount of traffic per month, but the first 
> requires a lot more than one VPS is going to handle.  (so is the second) but 
> it is the peaks that cost money. If you could get all of your traffic to 
> arrive on schedule in a perfectly flat usage pattern you’d save a crap load 
> of money.


What's great about current App Engine in the context of those graphs
is that everything happened automatically.

With the new price plan that is no longer the case. If you can predict
a week in advance how many instances you need, you pay only .05/IH
instead of .08/IH.  What is the optimal number of reserved vs. demand
instance hours for your spiky traffic and your sinusoidal traffic
sites? How are you supposed to figure that out? You will have to
automate the calculation with some kind of predictive accounting
system, scraping data out of the App Engine console. There's a
business opportunity there for someone, unfortunately...

Instead of writing code to scale your website you need to write some
bullshit stock market-like adversarial accounting algorithm to fight
over 40% of your cost.

-- 
You received this message because you are subscribed to the Google Groups 
"Google App Engine" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/google-appengine?hl=en.

Reply via email to