From: hcb@gettcomm.com
Date: Wed May 31 2006 - 14:29:15 ART
Quoting "Bryant, Paul M" <paul.m.bryant@uk.verizonbusiness.com>:
> I work for large ISP we generally put DC, which has many definitions but in
> many cases it is a location with a few connections but total resilience
> within those connections, we would normally connect these a nssa ( nssa just
> in case we need to redist something in the future). The DC routers just want
> to send data out, quickly so a small routing table is a good idea so I would
> have thought even a nice tidy area 0 would have a bit more information that
> is really needed. But as DC can mean many thing to different people this
> might not be ok for your requirement.
In general, I prefer not to put any application hosts in area 0.0.0.0. Even if
the applications are centralized, I'm likely to make area 0.0.0.0 a simple LAN
and put the servers, at the same physical site, in their own nonzero area.
In your DC, is there traffic between servers in it, such as backup? Usually,
this is the case for a network of any appreciable size -- the servers backup,
synchronize, publish, etc. Why should routes relevant only to those servers
leak out of the area?
Remember that I said I don't put application hosts in area 0.0.0.0. That
doesn't mean there aren't hosts there, such as TFTP servers, network management
boxes, primary DNS, or, if the topology makes sense, firewalls. AI select these
as the devices that may be necessary to get the overall network back up after a
major failure, and to which access should never be delayed by application
traffic. What if, for example, there's a major security bug in IOS that needs
an upgrade, or a host OS worm? How do you distribute the software updates?
>
> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
> CCIEin2006
> Sent: 31 May 2006 14:37
> To: Jeff Theunissen
> Cc: Cisco certification
> Subject: Re: OSPF Area Design in Large Networks
>
> Thanks guys.
>
> So if you had two data centers, would you put both data centers in area 0 or
> would you use a different area number for each data center?
>
>
> On 5/30/06, Jeff Theunissen <jhtemail@yahoo.com.au> wrote:
> >
> > I work with a large OSPF Network with over 57'000 clients. This
> > network is a star network utilising area 0 at the core. The sub area's
> > are based on what we call clusters which are the way the sites are
> > brokens down. These clusters are given sections of the IP Address
> > subnetting, this provides a convienient point for summarizing the address
> space.
> >
> > Jeff
> >
> >
> > *CCIEin2006 <ciscocciein2006@gmail.com>* wrote:
> >
> > Hello folks,
> >
> > Being that my OSPF experience is limited to small environments with
> > one area (area 0) I would like to know how big companies do OSPF.
> > For example, when would they decide to create multiple areas?
> > How do they usually assign areas - by geographic area, by department,
> > by function?
> > Are virtual links common practice in the real world?
> >
> > Thanks
> >
> > ______________________________________________________________________
> > _ Subscription information may be found at:
> > http://www.groupstudy.com/list/CCIELab.html
> >
> >
> > ------------------------------
> > The LOST Ninja blog: Exclusive clues, clips and gossip.
> > <http://us.rd.yahoo.com/mail/tag/au/**http%3A%2F%2Fau.blogs.yahoo.com%
> > 2Flostninja>
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
This archive was generated by hypermail 2.1.4 : Thu Jun 01 2006 - 06:33:23 ART