Reinventing Network Management

Network management is the proverbial bus syndrome – nothing shows up forever and then there’s a whole queue of them  – in this case interesting technologies, but here’s the really interesting one I’m about to start testing with – Moogsoft.

Think radical invention of network management and you’re getting there – the name and website give some clues I guess that this isn’t mainstream, me-too stuff… 
So here’s the problem – network management, even in its modern incarnation of application performance monitoring and all the variations on a theme, is all about some concept of a network configuration being stable and predictable. So the idea is that you, over time, build up a “rich database” of information regarding all elements of the network – hardware and software – so that there’s a level of intelligence to access when identifying problems (and potential problems). OK, except that, if you have a network deployment that is part cloud (or managed service of some description), part-virtualised and otherwise outsourced to some extent, how can you possibly know what the shape of that network is. Even as the service provider you cannot… 
It therefore doesn’t matter how much networking data you collect  – effectively you have to start from scratch every time, because the network is dynamic, not static, so any historical data is not necessarily correct. And we all know what happens in you make decisions based on inaccurate data… Moogsoft therefore says, forget about the existing methodologies – they don’t work any longer. Instead it uses algorithm-based techniques to establish concurrent relationships between all aspects of the network when an anomalous situation is identified – looking at every possible cause-effect possibility. And it works in a true, collaborative environment – after all, network management is not detached from other aspects of the network in the same way that user A in department X is not detached from user B in department Y. So every “element” of the “network” is relevant and involved.
Sounds like an impossible to scale scenario? Well how about handling 170m incidents a day? And taking resolution time down from hours and days to minutes? Sounds too good to be true? Maybe so, but these are recorded results involving a famous, global Internet player.
Watch out for the Broadband-Testing report on the Moogsoft solution – should be somewhat interesting!!!!
Enhanced by Zemanta