Shoutcast Traffic


Yesterday, somebody asked me how much traffic a Shotcast server produces. And I though I might as well share the computation:

Stream @ 128 kbps (Kilo Bit per second) on in average 30 days of a month
n = average number of users that listen 24h per day
s = 1.1 = Shoutcast overhead factor (10%)

So to calculate the traffic in Terabyte:

(((128 / 8) * 60 * 60 * 24 * 30) * n * s) / 1024 / 1024 / 1024

Examples:

n = 10
((((128 / 8) * 60 * 60 * 24 * 30) * 10) * 1.1) / 1024 / 1024 / 1024 = 0.42 TB = 420 GB

n = 50
((((128 / 8) * 60 * 60 * 24 * 30) * 50) * 1.1) / 1024 / 1024 / 1024 = 2.12 TB

The amount of traffic rises linearly with each additional user. So operators: plan ahead!

Information and Links

Join the fray by commenting, tracking what others have to say, or linking to it from your blog.

Information
May 16th, 2011
2 Responses
Feeds and Links
Comment Feed
From This Author
Del.icio.us
Digg
Technorati
Categories and Tags
Chaos

Other Posts

Write a Comment

Take a moment to comment and tell us what you think. Some basic HTML is allowed for formatting.

Reader Comments

Comment Number:
1
Written by:
dagobert
Posted on:
May 17, 2011 at 10:04

Nice calculation :)
On the otherhand, you basically just changed the question! It is now: how do I calculate the average 24h listener (n)?
This might be even more tricky than your simple multiplication, bacause you need to know the average listening time per day of the users over one week. It is important to average it over the whole week because specially on weekends the behavior might be very different.

Greets

Comment Number:
2
Written by:
henning
Posted on:
May 17, 2011 at 14:05

Yes, interesting!
You might ask that question in correspondence to your maximal allowance of traffic on the server you use.

Let’s say we have a traffic limit of traffic = 5TB per month; the maximum number of averaged #24h-listeners would be: n = (traffic * 1024^3 / ((128 * 1.1)/8)) / (30 * 24 * 60 * 60) = 117.68 users.

Of course this may be shifted around to balance workday vs. weekend. I think it is actually best to limit the number of listeners to the number the calculation spits out to then observe the behavioral patterns of the users. There might be configurations out there that allow for dynamic listener count limits.