IT and networking reinvents itself partially or wholly every few years, and here we are again now - distributed, virtualised, cloud (private and public) based, hybrid networks... So where does this leave traditional Network Management (NetMan) applications?
Every few years, the topic of "next generation" NetMan crops up once more and here we are again right now, thanks to the the distributed, somewhat cloudy and virtualised nature of contemporary network deployments. I mean, just how do you manage a network "entity" if you don't know where it is?
I recently finished testing with a genuinely fascinating start-up called Moogsoft (the report is available from the broadband-testing.co.uk website) who also featured in an article I did recently for Computer Weekly on managing external, virtualised networks (it's on the CW website somewhere!). The work really did bring home to me just HOW different it is attempting to manage a global network in 2014, compared to even 10 years ago. What seems now way back in '99 with the emergence of network optimisation products and security I tried to set up the NGNMF - Next Generation Network Management Forum - with a view to getting the network management specialists - the BMCs, CAs, HPs IBM Tivoli's etc of this world - to outline how they would advance their software solutions to cope with the - then - new generation of networks being deployed.
15 years on and that task is almost infinitesimally more challenging. I was speaking last week with Dan Holmes, director of product management at CA, about this very topic. Dan has historically been through many of the phases of network management development as myself, in his case starting with the - then Cabletron's Spectrum manager, one of the first products to try and bring AI into the equation for resolving networking problems. Dan acknowledged that the change in networking infrrastructure is requiring a change of approach by all the major NetMan vendors and pointed to a lack of standards as being just one of many issues to resolve. He described the fundamental difference now as being, whereas NetMan was previously focused from the inside looking out, now it has to refocus from the outside looking in; in other words, the starting point is the bigger picture and you have to drill down to individual services, threads, user conversations, transactions... That's a hell of a lot of "data" to manage - application monitoring and similar tools can do certain tasks, but they are not the complete answer - checkout the Moogsoft report to see how radical a solution is seemingly required in order to be that latter-day solution.
And if that means the major NetMan vendors are all playing catch-up currently, it'll be interesting to see how fast each or all of them can adapt to the new generation of networks being rapidly built out there in the ether at the moment...