IPP Mail Archive: IPP> Firewalls etc.

IPP> Firewalls etc.

Paul Moore (paulmo@microsoft.com)
Thu, 11 Jun 1998 12:36:14 -0700

It became clear to me during the last tele conf that we need to clearly
distinguish two cases in this discussion.

1. A user inside a corporation sending print commands out into the internet.
This is the one I was always talking about

2. A printer inside a corporation being accessed from outside. It was clear
to me that this is what some others were talking about

I think that these are two very different things and we have not be clear
about what is desirable, acceptable, etc. in each case. Also I think that
some people have been discussing #1 with people who think that they are
discussing #2 (I certainly discovered that I was)
----------------------------------------------------------------------------
---------------------------
Now my views on the two cases.

#1 must work by default. I.e. if somebody on a web site or whatever has an
ipp URL then (provided I have the right client s/w installed) I should be
able to print to it, in the same way I can send email or browse a web site.

#2 Must not work by default. I.e. if I install an IPP printer (whether s/w
on an NT server or embedded in the device) then somebody outside my company
should not be able to print to it. .

This is the solution we have today. #1 works provided that the user can post
onto the internet (usually true). #2 works because most networks dont allow
arbitrary inbound posts. This is true in routed cases and in proxy cases.

The main problem that we seem to have is 'how can an admin allow #2?'. For
example PrintShopsRUs want to make their printer(s) available. Second
example is: I want to set up an IPP printer in my office as an 'internet
fax'. How can their admin configure their router/firewall/proxy to only
allow IPP traffic in? Well today's proxies dont allow inbound HTTP traffic
which rules this scenario out in the vast majority of corporate cases
anyway. For routed/firewall cases it seems that you either allow any POST to
a given IP address (OK for a real printer but not ok for a server), or we
need someway of distinguishing the traffic. If we dont do this then this
will disable this class of usage scenarios. My view is that this is of
secondary importance - given that #2 will never work for proxied
environments we should focus on #1 scenarios, which do work. For #2 people
will have to either a)trust and cross their fingers b) place the printers
outside the firewall (like their web servers often are).

I can never see the Internet Fax one working in a corporation. I mean
allowing me to setup up a printer in my office and just having it work
without any SA work. Most network admins are fanatical (rightly so) about
not allowing inbound connections, each one has has to be approved and hand
configured. They dont even allow ping in, so there is no way an admin would
allow inbound IPP traffic to all IP adresses.

The only issue I see left then is 'how can an admin prevent #1'? My
immediate response is 'why would they want to?'