Well, sorta.  For the sake of discussion, let's ignore the DTC48s.
Mostly because I'm not sure what
they will and will not do when interacting with the latest DTCs and
dtcmgrs.  Straight AFCP is not
routable, period.  Additionally, most dtcs do not use a routeable
managment protocol (boot and
config activities).
 The DTC 16 and 48 MX models do require an openview dtcmgr console or a
9000 with dtcmgrux
on their subnet to download the dtc config (code?).  However, the
"Routeable Avesta Flow Control
Protocol" (RAFCP) does permit talking to and from them in a routed
fashion.  You can have routed
printers and terminals you just cannot boot the MX's across a route.
  The DTC 16 RX, however, can also be managed across a routed network
with the help of
the appropriate UX version of DTCMGR (DTCMGRRX?).  Thus you can use
DTCs across a
thoroughly routed network.  You just need to get 16RX dtcs and a little
HP-UX box (or two) to do the
management.
  The general problem with DTCs across a routed network has been (since
the intro of the
original RAFCP) not them talking but their management which was still
not routable.  The newer
RAFCP uses a bunch of DNS entries to point the remote DTCs to their
appropriate front ends and
back again.  The older RAFCP (with which I'm not very familiar) used
socket 168 (or some such) and
a few DNS entries.

It can be done,

Noel



On Fri, 16 Apr 1999 17:06:54 -0400, Joe Geiser wrote:

>Since my name was mentioned, even if only by initial - thanks Terry...
>
>AFCP in itself is not routable.
>
>There are DTCs now (newer ones, the models escape my gray matter at the
>present) that can have IP addresses assigned to them for routing purposes,
>but for purposes of using printers and terminals over a WAN or other routed
>network, bridging is the only option.
>
>BTW - my experience (others may have other experiences) has been that the
>performance penalty of having the this traffic bridged is not all that
>great.  In a former life, we ran 15 sites over a private frame relay
>network, each site had at least three 600 LPM printers, and at least 50
>terminals (not PCs!) nailed to DTCs at the far end.  We always had no more
>than 1-2 second response time.
>
>Regards,
>Joe
>
>
>====================================================
>Joe Geiser, President and Senior Partner
>CSI Business Solutions, LLC
>Phone: +1(215) 945.8100   Fax: +1(215) 943.8408
>Toll Free (US/Canada): (877) 945.8100
>====================================================
>
>> -----Original Message-----
>> From: Simpkins, Terry [mailto:[log in to unmask]]
>> Sent: Friday, April 16, 1999 16:37
>> To: [log in to unmask]
>> Subject: Re: openview DTC manager
>>
>>
>> Larry Simonsen asks about "routable AFCP"
>> ---------------------
>>
>> AFCP is NOT routable.  You have to bridge it unless you use
>> the openview DTC mgr.  It will "encapsulate" (my word) the AFCP inside
>> an IP packet and that makes it routable.
>>
>> At least this is how it has been explained to me at least three times
>> by three different people that I trust (and Joe G. too ;-)
>> And, you have to have an OV station at each site to do the "packing"
>> and "unpacking", so it gets a bit expensive and a pain to manage.
>>
>> *******************************
>> Terry W. Simpkins
>> Director ISIT
>> Lucas Control Systems
>> [log in to unmask]
>> *******************************
>>
>