Skip to Content.
Sympa Menu

wg-multicast - A multicast-enabled cash register

Subject: All things related to multicast

List archive

A multicast-enabled cash register


Chronological Thread 
  • From: Bill Owens <>
  • To:
  • Subject: A multicast-enabled cash register
  • Date: Thu, 12 Mar 2015 09:00:15 -0400

Or, properly a 'till' since it's in the UK.

This morning I was trying to troubleshoot some odd log messages on one of our
Juniper EX9208 routers (which I still haven't figured out) and decided that I
would try tcpdump on the router, since Junipers can do that. After applying a
few filters to clear up traffic that I knew was legit, I started to see a lot
of this:

12:37:47.044973 In IP elvyn-bar-till-1.lut.ac.uk.19264 > 224.1.1.1.8051: PGM
0x000000000000 SPM seq 55283 trail 4965 lead 4964 nla 131.231.22.78 [0]
12:38:04.382642 In IP elvynbar-2.epo.lut.ac.uk.12586 > 224.1.1.1.8051: PGM
0x000000000000 SPM seq 15619 trail 2809 lead 2808 nla 131.231.22.92 [0]

I didn't expect to see *any* multicast traffic to the router, so first I
wanted to see what was actually using that group, and why. A check of MSDP
SAs showed three sources, all at Longborough University in the UK:

224.1.1.1 131.231.22.78 199.109.0.73 212.219.212.232 Accept
224.1.1.1 131.231.22.92 199.109.0.73 212.219.212.232 Accept
224.1.1.1 131.231.179.207 199.109.0.73 194.82.152.254 Accept

I decided to run wireshark against the traffic, so I joined the group from my
Mac and started capturing. Most of the packets are pragmatic general
multicast and they all have the router alert option set, which I think is why
I'm seeing them come to the CPU:

Options: (4 bytes), Router Alert
Router Alert (4 bytes): Router shall examine packet (0)
Type: 148
1... .... = Copy on fragmentation: Yes
.00. .... = Class: Control (0)
...1 0100 = Number: Router Alert (20)
Length: 4
Router Alert: Router shall examine packet (0)

Most of them appear to be heartbeat packets, though I admit that I haven't
tried to really understand PGM; however, one has data and it's quite
interesting:

<?xml version="1.0" encoding="utf-16"?>
<OperatorData xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance";
xmlns:xsd="http://www.w3.org/2001/XMLSchema";>
<OwnerId>
231891
</OwnerId>
<OwnerName>
Jack Lakin
</OwnerName>
<IsLeftHanded>
false
</IsLeftHanded>
<ListShowOperatorId>
true
</ListShowOperatorId>
<ListShowTransactionName>
true
</ListShowTransactionName>
<ListShowCovers>
false
</ListShowCovers>
<ListShowSubtotal>
true
</ListShowSubtotal>
<ListShowDrawerEvents>
false
</ListShowDrawerEvents>
<ListShowTime>
true
</ListShowTime>
<TransactionIdHistory>
<unsignedLong>
145236
</unsignedLong>
</TransactionIdHistory>
... and so on.

I was very confused about what this all meant until I looked up the PTR for
the source:

78.22.231.131.in-addr.arpa. 86400 IN PTR elvyn-bar-till-1.lut.ac.uk.

Yes, there is a bar at Longborough called the Elvyn, and it presumably has a
till. Which, for some reason, is using multicast, with a global (and
reserved) group address. And setting an apparently unnecessary flag that
makes the packets go to every router, everywhere.

I think it's extremely unlikely that I would ever be able to find Jack Lakin
or a networking person at lut.ac.uk who would be interested in reconfiguring
their cash register, so I'll be filtering that myself...

Bill.



Archive powered by MHonArc 2.6.16.

Top of Page