Auralization and SonicFinder

Terry Slattery
Principal Architect

I’ve written here (and in our Forum) about visualization and why a lot of visualization isn’t very useful.  An interesting parallel concept is that of auralization, which is to sound what visualization is to sight.  William W. Gayer at Apple developed a neat application he called SonicFinder that uses sound as a key part of the user interface.  There are a number of papers and references to his work, most of which I’ve not yet had time to digest.

Steve Busby at the Univ of Nebraska Medical Systems reminded me about Netcosm by NetQoS.  I’ve seen it before and haven’t yet figured out how, in its current state, it would be useful to me in an operational network of more than about 50 routers and switches.  My problem with it is how to effectively use it in a network of hundreds or thousands of nodes and their associated packet flows.  The video shows too much detail – i.e. it’s too high resolution.

I need something like a network weather map that shows me where the weather is bad and where it is good.  I could then focus my attention on the bad sections, where the Netcosm system might actually be useful.  I’m also curious how much CPU goes into driving Netcosm’s display.  I could see a netflow collector being used to drive a lot of what’s happening.  But there’s more being displayed than netflow.  Where does it get the data for the firewall animations?  What about topology information?  How can it update this in real-time, making it a useful visualization or troubleshooting tool?  Or is it not necessary for it to be real-time?

Anyway, back to sounds.

I can see (hear) that an audio output would be useful.  There was a paper presented at Usenix many, many years ago about using syslog or server events to drive sounds.  The demonstration was pretty interesting.  Normal events of various types generated water bubbling sounds like in a stream.  Other less common events generated crickets or bird calls.  Major events were like a tree falling or an explosion (depending on the severity of the event).  The audio that was demonstrated was really useful.  I could imagine working away and not having to watch what’s happening in a network map or syslog server and have a relatively nice set of background noise.  Then, when something important happened, I’d hear it and could immediately take action.  Having the sound of a brook bubbling in the background, augmented with crickets and birds for other events, plus the occasional boom of something big happening would be interesting.

I can imagine some good musicians assembling sets of sounds that work well together.  Instead of UI ‘skins’, we’d have UI ‘sounds’.



Re-posted with Permission 

NetCraftsmen would like to acknowledge Infoblox for their permission to re-post this article which originally appeared in the Applied Infrastructure blog under


Leave a Reply