back to article Network monitoring is hard... If only there was some kind of machine that could learn to do it

It's difficult not to wish for the "good old days" when workloads stayed put, packets behaved, and firewalls just did basic port-level filtering. Admins knew where they stood. The cloud has changed that. Workloads and the resources they run have become malleable, virtual machines can be shunted from place to place, traffic …

  1. Herring`

    "AI and ML have an insatiable demand for data."

    So what happens when the cause of the network issues turns out to be the AI that's managing the network? Will it switch itself off?

    1. Anonymous Coward
      Anonymous Coward

      'Will it switch itself off?'

      'Just what do you think you are doing, Dave?'

  2. Tom Paine Silver badge
    Stop

    Hmmmm

    How do you troubleshoot a network issue when a so-called AI is reconfiguring stuff every 100 millis? Even if you figure out what it's doing that's causing a problem, how'd you find out why it did it, or how to make it stop doing it?

    Or is the thinking that our new ML overlords will net-eng with such incredible Homo silico skill and speed that nothing ever breaks?

    I think I know which file this one's going in. Actually that file's quite full of other AI / ML bolllocks, it's about time I emptied it...

  3. Herring`

    No Skynet comment yet? I am disappointed.

  4. john.jones.name
    Go

    reason for SDN

    one of the driving forces for SDN

  5. Anonymous Coward
    Anonymous Coward

    What a load of bollocks!

    Yes, the big virtual hosters need more automated, real-time monitoring, but how many enterprises really approach the scale of Google, Facebook, AWS, Azure, etc.? Automation has it's place, but troubleshooting IT issues isn't that place.

  6. The Monitoring Guy

    The AI and ML Rabbit Hole

    I get it. AI and ML are fancy and get great results... when you invest the time and resources. But, companies rarely invest properly in operations components that don't directly contribute to revenue. If they did, the need for AI and ML in monitoring would be virtually non-existent. And investing in AI and ML really wouldn't address the deficit of attention give to IT operations. We need to remember that AI and ML are an enhancement, not a fix for broken monitoring. What IT operations really needs is discipline and a cohesive methodology, not another system to manage.

  7. Pascal Monett Silver badge

    Oh for God's sake

    There is no AI. There is statistical analysis, and vast amounts of data, but we do not have a machine to which we can ask : "tell me what is working".

    There are highly intelligent people who understand statistics and can fine-tune these statistical analysis machines, but AI it is not.

    I would really like for the media to drop the "AI" in their articles, but that is obviously never going to happen because . . marketing. Ai is sexy, and because we don't have it, it is the perfect fantasy.

  8. onefang Silver badge

    I'm still looking for a good solution to monitoring just two computers. Last time I looked, I had to use half a dozen different tools to do what I needed.

  9. 89724102172714182892114I7551670349743096734346773478647892349863592355648544996312855148587659264921

    Greatly anticipating a Reg piece about the flipside of this issue: State sponsored AI hacking must be a reality.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2019