Re: Number of Routes and network stability

[Jing Shen <jshen_cad at YAHOO.COM.CN>]

I think the number parallel adjacent router per
interface  should be the point that node degree should
be limited; but, to current situation, even with L3
switch, each interface does not exceed 10 adjacy in
real networks. 

So, IMHO the total number of routes should not be a
hard limit for a OSPF area; but, the covergence time &
node degree should be limited. 

regards

Jing



--- Brett White <brett_white at AGILENT.COM> wrote:
> Hi Mitchell,
> 
> You say that a router shouldn't have an issue with
> 3000 links to other routers in an area.  What I am
> interested in is how you see the breakdown of this
> figure on a per interface basis?  Do you expect that
> the device should be able to support these 3000
> links on a single interface or spread across say 128
> ports?
> 
> Regards,
>  
> Brett White
> 
> -----Original Message-----
> From: Mailing List
> [mailto:OSPF at PEACH.EASE.LSOFT.COM] On Behalf Of
> Erblichs
> Sent: Friday, 3 December 2004 13:49
> To: OSPF at PEACH.EASE.LSOFT.COM
> Subject: Re: Number of Routes and network stability
> 
> Andrew Smith,
> 
>         I don't think I agree with 90% of this.
> 
>         Available memory has to do with the max size
> of
>         the LSDB, siumultaneous rexmit queues, etc,
> THAN
>         the max number of routers supported in a
> area.
> 
>         Assuming that enough memory is available,
> link-stability
>         and dead-router interval are more important
> than CPU
>         capability wrt network stability. Next comes
> with the
>         number & sizing of input queues, link-speed,
> and network
>         topology.
> 
>         If the number of hellos and OSPF control
> pkts
>         overload a CPU, then I hate to see the SPF
> timeframes,
>         or what would happen if a large number of
> pkts
>         are targeted to the router.
> 
>         If one says that they can't handle close to
> 3000
>         simultaneous links with other routers in a
> area,
>         then IMO, they have a few issues with their
> code.
> 
>         I am NOT stating that 3000 simultaneous full
> adjs can
>         be formed. But if you are dropping adjs
> because their
>         are too many routers in the area and you
> thus are
>         dropping (or delaing the processing of)
> hellos, then
>         their is something wrong with the code or
> hardware.
> 
>         Support for a large number of routers within
> a area
>         more has to do with IMO, the ability to
> decrease the
>         hello interval and hello multiplier and
> still not drop
>         adjs. Secondly, each router then COULD
> support fewer
>         links and a failed router would have less of
> a topology
>         effect.
> 
>         Mitchell Erblich
>         -----------------
> 
> > Andrew Smith wrote:
> >
> > The recommendation for number of routers per area
> is usually derived 
> > by considering the amount of processor power and
> memory available to 
> > the routers in an area. Usually you consider the
> lowest common 
> > denominator. This means the smallest router (the
> router with the 
> > fewest resources) limits the size of the area.
> >
> > A good rule of thumb is no more than 100 routers
> per area. If you try 
> > to keep an initial design to 50-70 routers that
> usually allows room to 
> > grow and recover from a network crisis.
> >
> > The vendors that say they can do 3000 routers per
> area are giving a 
> > scalability number for that box. This particular
> router can achieve 
> > 3000 routers per area "under the proper
> circumstances". This is good 
> > to know for how much the box can take, not how
> high to scale your OSPF 
> > areas.
> >
> > Regards,
> > Andy Smith
> >
> > -----Original Message-----
> > From: Mailing List
> [mailto:OSPF at PEACH.EASE.LSOFT.COM] On Behalf Of 
> > Jing Shen
> > Sent: Thursday, December 02, 2004 4:46 AM
> > To: OSPF at PEACH.EASE.LSOFT.COM
> > Subject: Number of Routes and network stability
> >
> > Hi,
> >
> > Some vendors recommended that number of routers
> within one OSPF area 
> > should not exceed 50, some others recommended 
> number of routes within 
> > one area should not exceed 3000.
> >
> > To my understanding, OSPF stability is directly
> related to link 
> > stability, radius of an OSPF area and OSPF
> implementation( router 
> > hardware, methodology of link status processing).
> But, is there any 
> > research done on this topic? or how are those
> recommendation derived ?
> >
> > thanks
> >
> > Jing Shen
> >
> >
>
_________________________________________________________
> > Do You Yahoo!?
> > 注册世界一流品质的雅虎免费电邮
> >
>
http://cn.rd.yahoo.com/mail_cn/tag/1g/*http://cn.mail.yahoo.com/
> 

Jing Shen

Data Communication Center
HangZhou TeleCom 
HangZhou ZJ 310027
P.R.China

' spamcontrol '

_________________________________________________________
Do You Yahoo!?
注册世界一流品质的雅虎免费电邮
http://cn.rd.yahoo.com/mail_cn/tag/1g/*http://cn.mail.yahoo.com/