Article View: comp.security.misc
Article #1045Re: finger @direct.ca\!
From: Chris Lewis
Date: Mon, 30 Jan 1995 21:20
Date: Mon, 30 Jan 1995 21:20
63 lines
3583 bytes
3583 bytes
In article <3ghioe$q83@trance.helix.net>, Eric Gisin <ericg@trance.helix.net> wrote: >$ finger @direct.ca >[direct.ca] > Thank you for connecting to Internet Direct. >The service you attempted to connect to on this host is for the >internal use of Internet Direct only. We ask that you do not >attempt to connect to it again. >All connection attempts are logged. If unauthorized connection >attempts continue from a single source CERT/RCMP/CIA/Scotland Yard >will be informed. >Thanks for your co-operation. >----- >Who do these people think they are? >Finger is a *public* service, there is nothing "unauthorized" in using it. >If they don't want the outside world to use it, they should just disable it. >Would CERT even listen to Internet Direct? This is probably idiosyncratic of how their protection facilities work. Probably TCP/IP security wrappers. Say they have only one machine and it's connected to the Internet. If they turn off the port entirely, then their internal users can't use the service at all. If they turn off the port to outside requests, then people get connection failures (or maybe their software needs to connect first before killing it), without knowing why, and keep trying. So they try to be "nice" and give you a "go away, we mean it!" message. The "we mean it" part seems rather excessive about "finger", but imagine that this was telnet or ftp. They probably could try to use different messages for different services, but maybe their software doesn't allow that (it may simply be "if we get an external connect on any port other than x,y,z, issue the following message), or they explicitly decided to reduce support complexity by using the same one. I can't imagine how you can claim that use of finger against a privately owned machine is a "public service" that an Internet-connected machine is somehow required to perform. It is not. No machine on the Internet is obligated to do *anything* for anyone. Further, finger is a security hole, and smart administrators have their Internet firewalls prohibit it. Ours certainly do, but we're not "nice", and simply refuse connects - if they only have one machine, they may not even be able to do that - the connection may need to be established first before their software can decide whether to kill it or not. If someone started messing around with indirect.ca's finger port, even with it disabled, that caused service impairment at indirect.ca (eg: ping storms of the finger port that impaired bandwidth over indirect's IP link), you can bet that CERT will be interested - once they've satisfied themselves that indirect is doing the right thing technically. -- Every once in a while I write something that someone might use in a signature. So I go back and make it real spiffy. No luck. Then some guy goes and uses *that* dog. Argh!
Technical Details
Message-ID:
<3gjl7c$9jk@bcarh8ab.bnr.ca>Path:
arc1-ca.pugleaf.net!swrinde!cs.utexas.edu!utnut!nott!bcarh189.bnr.ca!bcarh8ab.bnr.ca!clewisReferences:
<3ghioe$q83@trance.helix.net>Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: ff723,56914790b4b5eb68 X-Google-Attributes: gidff723,public X-Google-Thread: 10aaae,56914790b4b5eb68 X-Google-Attributes: gid10aaae,public X-Google-ArrivalTime: 1995-01-30 14:34:10 PST Path: arc1-ca.pugleaf.net!swrinde!cs.utexas.edu!utnut!nott!bcarh189.bnr.ca!bcarh8ab.bnr.ca!clewis From: clewis@bnr.ca (Chris Lewis) Newsgroups: comp.security.misc,can.infohighway Subject: Re: finger @direct.ca\! Date: 30 Jan 1995 21:20:44 GMT Organization: Bell Northern Research, Ottawa Canada Lines: 63 Message-ID: <3gjl7c$9jk@bcarh8ab.bnr.ca> References: <3ghioe$q83@trance.helix.net> NNTP-Posting-Host: bcarh2ec.bnr.ca