Filed under: Performance Monitoring
For organizations that are monitoring and maintaining a large network, bandwidth consumption can cause major issues. Knowing who is using all the bandwidth in the company is particularly helpful when troubleshooting performance issues and all too common complaints that “the network is slow!” Insight into what external web sites employees are visiting, and whether or not those sites are work-related is important. Today, almost everyone uses the Internet for everything from news updates, to dating, to banking, to music, and so on. Many of these will not consume too much bandwidth, but that depends on the service.
For example, when I use PathView Cloud to look at the bandwidth consumption of my company’s network – full of 20-somethings with access to a computer all day long – I can easily see that music streaming was occurring to many machines. This would not be a problem on a very large network with ample bandwidth. However, with smaller networks like ours, if everyone is streaming their own music. then the amount of traffic generated by each individual user can be staggering. I decided to run a test to see exactly how much bandwidth my co-workers were using and what the impact was. The following examples show how bandwidth utilization varies between two common music services. One is the typical Pandora.com which streams music based on the preference the user, to introduce them to new artists. The second page is Archive.org, which is a webpage for live music, allowing the user to listen to recordings of live shows. I ran the test for 1 hour to come up with these results.
Archive.org generates 75 MB
As you can see there are many instances of Pandora.com, based on the server where the music is playing from. Archive.org seems to work in a similar fashion, however there are fewer servers as the play selection is distributed differently. Archive.org is creating three times the traffic in the hourdisplayed. Seeing that we have a T1 or 1.544 Mb, this means that streaming music was using up one seventh of the overall bandwidth! So, getting a grasp on the streaming of music in the workplace can be helpful to verify the use of a previous resource. If Pandora.com can generate significantly less bandwidth than Archive.org, it may be a good practice to block Archive.org. This is exactly why flow analysis is critical to network monitoring. As companies deploy more performance sensitive applications like VoIP, video, VDI and cloud services, performance really matters. The more bandwidth being utilized by the end user, the less bandwidth is available for work related tasks.