Let's break some estimates down to requests per second and ponder the results...

The assumptions we'll work from in this example:

Peak traffic on the site is from 9am - 7pm; this gives a 10 hour peak window. Let's say that 80% of the total traffic falls inside this window fairly evenly spread across the ten hours.

Let's say that the site gets 10,000 visitors per day and the average visitor views 10 pages. That gives us 100,000 daily page views. Let's say that each page generates 10 requests (hits), so we have 1,000,000 requests per day.

Okay, some basic math:

Daily Requests * Percent of Traffic During Peak = Peak Requests

Peak Requests / Number of Peak Hours = Peak Requests Per Hour

Peak Requests Per Hour / 60 = Peak Requests Per Minute

Peak Requests Per Minute / 60 = Peak Requests Per Second

Using our example data we come up with around 22 requests per second. Is that a lot? Well, it’s more of a burst than most average sites get, but just considering this one data point, it is not a very large number consider what IIS can handle.

Requests per second can be useful, but not on its own. This is only one of several different factors and data points that need to be considered when performing an architecture review for a hosting solution.

~Brad