[Cialug] Is this monitoring tech really new?

jim kraai jimgkraai at gmail.com
Tue Nov 4 00:23:45 CST 2014


I'll say it differently.  It looks like a technique that was known years
ago in telecom is finally getting its moment in a different industry.  I'm
happy for them.

I interpret it as a dashboard on a complex control system that trains
itself to recognize what's normal and thus what to ignore so human
monitors/operators can focus on unusual exceptions--I think.

I was curious whether people are currently or have used similar software.

It's no big deal.

Thank you, Todd.

--jim
 On Nov 3, 2014 10:38 PM, "Todd Walton" <tdwalton at gmail.com> wrote:

> I didn't get much out of the press release you linked to. There was vague
> language all over the place! It "moves between state values that were
> measured at different points in time". Okay... that could mean almost
> anything.
>
> I'm not sure what it is you're comparing to the system you worked up in '93
> and '04, Jim.
>
> --
> Todd
>
>
> On Thu Oct 30 2014 at 4:20:07 PM jim kraai <jimgkraai at gmail.com> wrote:
>
> > It's a read and goes into good descriptive detail
> > http://m.phys.org/news/2014-10-lots-capacity-algorithm.html
> > No, I won't TL;DR it, that introduces my own biases.
> >
> > But is this really 'new'?  Surely someone here has done something
> similar.
> >
> > I wrote a multi-log-file watcher back in '93 that used what sounds like
> > eerily similar techniques, then in '04-ish, wrote something similar for
> > control systems that used wavelets with shifting windows across weekly
> and
> > monthly calendars for storage- and compute-efficient compare and
> alerting.
> > They seemed like obvious applications of what we had available at the
> time.
> >
> > Both times, I started working on a hierarchical system of watcher
> watchers
> > (no laughing, it was awesome), but just didn't have the time to do more
> > than spec them out. The next step was to see if machine learning algos
> > would add any value, again, no resources to go there.
> >
> > If this is new now, then aarrgghh, ouch, and all that
> > _______________________________________________
> > Cialug mailing list
> > Cialug at cialug.org
> > http://cialug.org/mailman/listinfo/cialug
> >
> _______________________________________________
> Cialug mailing list
> Cialug at cialug.org
> http://cialug.org/mailman/listinfo/cialug
>


More information about the Cialug mailing list