wg-multicast - Re: SAP storm from 145.19.1.183
Subject: All things related to multicast
List archive
- From: Bill Owens <>
- To: Michael Lambert <>
- Cc: wg-multicast <>
- Subject: Re: SAP storm from 145.19.1.183
- Date: Thu, 19 Jun 2008 09:51:13 -0400
On Thu, Jun 19, 2008 at 07:55:38AM -0400, Michael Lambert wrote:
> Not being a true multicast practitioner, but would it help to turn off
> MSDP and bid adieu to ASM? I guess this is as much a political
> question as a technical question.
No, although MSDP has certainly been a problem in the past, it isn't the
cause this time. From what I can see, we have two issues:
1. Network equipment that is either configured or defaults to listen to
traffic in the SAP group; for example, Cisco routers with "ip sap listen"
configured on an interface. This causes high CPU utilization, and depending
on the hardware and other confiugration the effects can range from dropped
traffic to BGP bouncing to crashes.
2. Sites with relatively small uplink bandwidth that happen to have a SAP
listener running, something as simple as VLC with SAP Announcements selected.
Although VLC appears to be stable even when receiving high SAP traffic, the
volume in the last two incidents was enough to congest the uplink for someone
with less than 25-30 Mbps of available bandwidth.
Rate limiting the SAP group will solve both of these problems if it is
implemented throughout the network. However, I doubt whether the effort
required to implement that thorough a configuration is greater than the
continuing value of SAP/SDP. Looking at the display in VLC just now, I have
44 working streams (though that's being generous, at least 8 of them had no
a/v content) and 51 non-working. I've long been in the habit of warning
people who want to test multicast that they have no better than a 50/50
chance of picking a working stream at random.
One other approach to fixing this problem might be to convince the fine folks
who 'maintain' the miniSAPserver to stop distributing it. That only prevents
one, accidental DoS vector; the problem remains that some substantial
fraction of the multicast-enabled network is vulnerable if someone sends high
traffic into the SAP group. To a certain extent this is a problem inherent to
ASM, but made much worse by the high likelihood that a site will have at
least one SAP listener. Other than perhaps one of the beacon groups, you'd be
lucky to find a single group that would have so many receivers.
Since this has become a recurring problem I'm going to approach it through
education; telling people not to leave a SAP listener running on their
desktop, and getting them to disable sap listen on their routers. Hopefully
that will be enough. But I wonder whether it is time to start down the road
towards deprecation. . .
Bill.
- SAP storm from 145.19.1.183, Julian Y. Koh, 06/19/2008
- RE: SAP storm from 145.19.1.183, Taylor, Scott J., 06/19/2008
- Re: SAP storm from 145.19.1.183, Martin Clifton, 06/19/2008
- Re: SAP storm from 145.19.1.183, Michael Lambert, 06/19/2008
- Re: SAP storm from 145.19.1.183, Simon Lockhart, 06/19/2008
- Re: SAP storm from 145.19.1.183, Bill Owens, 06/19/2008
- Re: SAP storm from 145.19.1.183, Frank Fulchiero, 06/19/2008
- Re: SAP storm from 145.19.1.183, A . L . M . Buxey, 06/19/2008
- Re: SAP storm from 145.19.1.183, Frank Fulchiero, 06/19/2008
- Re: SAP storm from 145.19.1.183, A . L . M . Buxey, 06/19/2008
- Re: SAP storm from 145.19.1.183, A . L . M . Buxey, 06/19/2008
- Re: SAP storm from 145.19.1.183, Frank Fulchiero, 06/19/2008
- Message not available
- Re: SAP storm from 145.19.1.183, Tim Chown, 06/27/2008
- Re: SAP storm from 145.19.1.183, Marshall Eubanks, 06/27/2008
- Re: SAP storm from 145.19.1.183, Bill Owens, 06/27/2008
- Re: SAP storm from 145.19.1.183, Marshall Eubanks, 06/27/2008
- Message not available
- Re: SAP storm from 145.19.1.183, Tim Chown, 06/27/2008
- Re: SAP storm from 145.19.1.183, Tim Chown, 06/27/2008
- Re: SAP storm from 145.19.1.183, Michael Lambert, 06/19/2008
- Re: SAP storm from 145.19.1.183, Wes Young, 06/19/2008
- Re: SAP storm from 145.19.1.183, Martin Clifton, 06/19/2008
- RE: SAP storm from 145.19.1.183, Taylor, Scott J., 06/19/2008
Archive powered by MHonArc 2.6.16.