> hmm,  you  make  up  the your straw-man problems and solve them, the
> rest of will fix the real problems :))

I'll  work  on  that,  while you work on not always throwing IMGate at
issues, even with its positive side effects. :))

> The only real protection is internal firewalling.

That's  irrelevant. Firewalling one server from another is meaningless
when  you  have to pass everything through the "primary" server, as in
the  architecture  that was being addressed in that part of my port. I
clearly meant that each subdomain could survive the other's downtime.

> He runs two mail servers as independent mail domains, Imail as 
> students.school.com and Exchange as faculty.school.com.

This  was  already  dealt  with  and suggested in my first e-mail (see
mail1.domain.com and mail2.domain.com) and in Dave's.

> The  email  address  translation  works  both  ways, both in and out
> through Imail.

Just  as  in  my  original suggestion, only with old hardware to cross
your  fingers  over  and  a  new OS to learn, and more alias tables to
maintain  (though,  as  I  have  never  denied,  with  the  well-  and
oft-stated benefits of IMGate).

Why  more  alias  tables?  In  your  setup, IMGate needs to know which
school.com  addresses  are  students  and which are faculty; and IMail
needs  to  know which are which, too, or it will fail delivery (unless
you put one-way peering in the mix, which you didn't mention).

Or were you suggesting that ALL mail go through IMGate to be rewritten
before  looping  back  to  IMail/Exchange?  That'd  be  even as far as
management  load  goes,  but  with  a  terrific waste of resources all
around:  every  loopback message will spend twice as long in the IMail
spool, use twice the disk I/O, etc.. Would you, of all people, want to
put IMail through that?

Or  perhaps you meant that users would actually use the subdomains for
initial  addressing?  If  so,  you've  abandoned a primary requirement
stated in the original post...of course, solutions are a lot easier if
you do that sort of thing. :)

> No $ for new machines.

We've  been  here  many  times before. 90% of the organizations that I
have   worked   for   and  observed  would  not  put  out-of-warranty,
non-fault-tolerant  hardware  in  a  mission-critical  position. Maybe
Chris'  employer  won't  mind, but it's always been cavalier to assume
that  tightly-managed,  -standardized, or -budgeted organizations will
be able to follow this suggestion. In addition, I'd hate to be working
against a raft of student wanna-be-hackers as a *nix newbie. IMO, it'd
be  more  productive (and avoid the recurrent questioning) to add such
caveats   to   the   IMGate  boilerplate,  instead  of  just  charging
forth...many  full-time sysadmins work under what we consultants would
consider  abject  IT misery, in terms of extra boxes that haven't been
scavenged  for  parts,  a  boss  who  listens  (isn't  that  why we're
consultants,  anyway?), purchasing authority, bureaucratic disrespect,
mandatory  project  bidding,  inane  first-level  support  issues that
ambush  self-tutorial  time, et al.

I  can think of several medium-size clients of mine, in fact, who have
great,  redundant  production  systems, but essentially nothing on the
shelf  (they  gave to charity--who knew?). In times past, picking up a
couple of $1000.00 boxes for a compelling new project was a given, but
with profits slipping, that's no more.

> Fantastically  better  anti-mail-abuse  as  a freebie, AND they will
> realize  later that IMGate will turn out to be, retrospectively, the
> biggest benefit of this project. :))

Yes, IMGate is a great answer to another question. It is a great thing
to  have for many reasons. But it doesn't simplify the architecture in
any way I can see.

-Sandy


Please visit http://www.ipswitch.com/support/mailing-lists.html 
to be removed from this list.

An Archive of this list is available at:
http://www.mail-archive.com/imail_forum%40list.ipswitch.com/

Please visit the Knowledge Base for answers to frequently asked
questions:  http://www.ipswitch.com/support/IMail/

Reply via email to