nanog mailing list archives

Re: multicast nightmare #42


From: Philip Lavine <source_route () yahoo com>
Date: Thu, 15 Oct 2009 13:06:43 -0700 (PDT)

Thank you Eric you are a genius, that has solved and issue that has plagued me for 3 years.

the problem was exactly as you said over subscription of the 8 ports tied to 1 ASIC




________________________________
From: Eric Ortega <eric_ortega () mmi net>
To: Philip Lavine <source_route () yahoo com>
Sent: Wed, October 14, 2009 9:51:43 AM
Subject: Re: multicast nightmare #42

Depending on the model of
blade there is an 8-to-1 over subscription on the 4500s. I have had all
kinds of headaches with this myself. The 48 port SFP "gig" blade can
only have 1 gig per each set of 8 ports. The aggregate ports are known
as "gigaports". The layout is gigaport 1 = 1,3,5,7,9,11,13,15 gigaport
2 = 2,4,6,8,10,12,14,16 and so on. I bet that if add up the total
bandwidth in each gigaport you might be over the "limit"

Philip Lavine wrote: 

I wish that was the case but the switch is a 4500 and the data
rates are less than 100 mbps on a 1 gig blade/sup




________________________________
From: >Eric Ortega <eric_ortega () mmi net>
To: Philip Lavine
<source_route () yahoo com>
Sent: Wed, October 14,
2009 8:24:59 AM
Subject: Re: multicast
nightmare #42

Are you over subscribing
either the link or the backplane of the switching device?

Philip Lavine wrote:

Please explain how this would be possible:

1 sender
1 mcast group
1 receiver
----------------
= no data loss

1 sender
1 mcast group
2+ receivers on same VLAN and physical segment
--------------------
= data loss






-- 


Eric R. Ortega
Network Engineer
Midcontinent Communications
605.357.5720
eric_ortega () gmail com 


-- 


Eric R. Ortega
Network Engineer
Midcontinent Communications
605.357.5720
eric_ortega () gmail com 


      


Current thread: