Re: heartbeat monitoring/failover list

On Fri, 21 Aug 1998, Peter Marelas wrote:

> > > At the recent WWW caching workshop in Manchester [1], there was quite
> > > a bit of interest in developing a free software (or should that be
> > > "open source" ?) system for doing the sort of heartbeat monitoring and
> > > transparent failover characterised by various Unix/NT/switch vendors
> > > "high availability" and "clustering" offerings. There doesn't appear
> > > to be anything out there at the moment - unless you know better!
> > >
> > > If you'd like to help out on this (or you know better :-), drop a
> > > line to heartbeat-request@net.lut.ac.uk to join in the fun.
> >
> > Check out the Alteon offerings (AceSwitch, AceDirector) which have health
> > check functionality. (http://www.alteon.com)
> >
> > lbnamed springs to mind as a free offering (Load Balanced named).
> >
> > RYO may be another way ;-)
> >
>
> The difference between High Availability implemented in the OS and
> High Availability implemented in a switch is that the switch is still a
> single point of failure.

Ok. So the vendors develop failure-over strategies and implement them in
their switch offerings (Alteon, Cisco Local Director - which I forgot to
mention, etc ...) ... cool for them, coz' they get to sell more boxes.

Now single point-of-failure is a mute point, but these solutions are only
available at considerable expense.

> The other difference is High Availability in the OS will be more beneficial
> to the inet community in general if we get together and help out these guys.

No argument here. On with the show! This issue comes up time-and-time
again and in almost every deployment situation I've been in. Good idea,
Peter; let's get heads together!

J.

----------------------------------------------------
John V Cougar | Voice: 1800 065 744
Cache Manager |-----------------------------
Telstra Internet | E-Mail: cougar@telstra.net
----------------------------------------------------
Received on Thu Aug 20 1998 - 18:30:09 MDT