I often wondered how sites detected the number of users online. After some digging I’ll admit to be no closer to the answer. The problem is that once a page is delivered to someone you have no idea what they’re doing unless, and if, they then do something – go to another page of yours, click a button, etc. They could have immediately left the site and gone elsewhere… you simply don’t get back the level of interaction that would tell you this.
So my solution was based on simple maths. Using Google Analytics and the like I work out how long an average user spends looking at a page on the site. Let’s say it’s 90 seconds.
I’m using PHP and MySQL for this example. So, create a new table named
users_online and create two fields for it – the first named
timestamp and the second
With your database already open execute the following PHP, pre-setting the variable
$online_seconds with the figure we discussed a couple of sentences ago (in my example, 90). This should be at the top of each page.
$timestamp = time(); $timeout = $timestamp-$online_seconds; $insert=mysql_query("INSERT INTO users_online VALUES('$timestamp','$REMOTE_ADDR')"); $delete=mysql_query("DELETE FROM users_online WHERE timestamp<'$timeout'"); $result=mysql_query("SELECT DISTINCT ip_address FROM users_online"); $user_num=mysql_num_rows($result);
This will add the period of time (90 seconds) to the current timestamp and insert a row into the table with that as a “timeout”. We then delete any rows where this timeour has now passed. Finally, we read the number of distinct IP addresses – this is the number of users currently online and is returned as
If nothing else it’s handy to have an idea as to how many people are online before you upgrade your site and potentially break it 😉