My company uses a HP 3000 mainframe for our customer information and
billing systems. The HP uses devices called DTCs to provide remote
terminal services and print services. Its well known that the protocol
(called AFCP) that the HP uses to communicate with DTC is not
routable. Therefore, bridging has to be enabled across WAN links to
provide DTC services to remote subnets.

We've long since replaced terminals with PCs and terminal emulators
but we still use printers attached to the DTC. I'm told that migrating
to a network printing solution is too expensive because of the
software cost for the HP 3000.

The environment this needs to work in is pretty simple: 2 sites
connected via a 10mbps Wireless Ethernet WAN with 2621s at each site.
One site has the HP 3000, the other has 2 subnet hanging off the 2621
each with a DTC on it.

My problem is that the bridging is causing bandwidth problems on my
WAN links. However my attempts to filter just the HP related traffic
have come up short. I have searched the MPE/iX lists and groups
extensively looking for solutions.

I have tried the following:

1.      Filter by MAC address of the HP 3000 and DTC by enabling bridging,
adding permanent forwarding table entries for the HP 3000 and all
remote DTC, and turning of dynamic forwarding table entries by issuing
"no bridge X acquire"

Ex. from my central site:

interface FastEthernet0/0
 description Connected to Merrillville LAN 10.100.0.0/16
 ip address 10.100.0.4 255.255.0.0
 no ip directed-broadcast
 appletalk cable-range 160-161 160.2
 appletalk zone post-trib
 bridge-group 10

interface FastEthernet0/1
 description connected to Wireless WAN
 ip address 10.102.0.1 255.255.255.0
 no ip directed-broadcast
 appletalk cable-range 162-162 162.1
 bridge-group 10

bridge 10 protocol ieee
no bridge 10 acquire
bridge 10 address 0800.09DB.109F forward FastEthernet 0/0 (Central HP
3000)
bridge 10 address 0800.0924.1530 forward FastEthernet 0/1 (Remote DTC)
bridge 10 address 0800.0910.0526 forward FastEthernet 0/1 (Remote DTC)
2.      Filter via creating access-lists that at my core site permit
traffic that has "0xfcfc" or "0xf8f8" data at offset 0xE bytes into
the frame (which I presume is HP multicast data), and another acl that
permits the MAC of my HP 3000.

Ex from the central site:

access-list 1100 permit 0000.0000.0000 ffff.ffff.ffff 0000.0000.0000
ffff.ffff.ffff 0xE 2 eq 0xFCFC
access-list 1100 permit 0000.0000.0000 ffff.ffff.ffff 0000.0000.0000
ffff.ffff.ffff 0xE 2 eq 0xF8F8
access-list 1100 deny 0000.0000.0000 ffff.ffff.ffff 0000.0000.0000
ffff.ffff.ffff

access-list 700 permit 0800.09DB.109F 0000.0000.000 (HP 3000 MAC)
access-list 700 deny 0000.0000.0000 ffff.ffff.ffff

interface FastEthernet0/0
description Connected to Merrillville LAN 10.100.0.0/16
ip address 10.100.0.4 255.255.0.0
no ip directed-broadcast
appletalk cable-range 160-161 160.2
appletalk zone post-trib
bridge-group 10
bridge-group 10 input-address-list 700
bridge-group 10 input-pattern-list 1100

Neither of these solutions work. Print jobs never make it to the
destination printer. I came across a few posts that suggest that DLSw+
could be used to encapsulate the traffic over IP. I can't find any
documents that describe how one would go about doing this. I'm not all
that familiar with DLSw+ but I know its most often used in SNA
environments. If someone has an example router config that I can take
a peak at that would help out a lot.

Has anyone successfully bridged JUST the HP related traffic across
their WAN? I'm dying to know how. Any suggestions would be very
appreciated.

TIA,

John Matney
Network/Systems Analyst
Post-Tribune Publishing
CCNA

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *