home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.protocols.tcp-ip
- Path: sparky!uunet!caen!sol.ctr.columbia.edu!cs.columbia.edu!ji
- From: ji@cs.columbia.edu (John Ioannidis)
- Subject: Re: Firewall usage
- Message-ID: <Bs3vCz.K13@cs.columbia.edu>
- Sender: news@cs.columbia.edu (The Daily News)
- Organization: Columbia University Department of Computer Science
- Date: Tue, 28 Jul 1992 15:36:35 GMT
- Lines: 78
-
- [[Too many articles to follow up on individually -- I do work instead
- of reading news (for a change) and suddently there's a flamewar that
- no firewall will stop (I know, bad pun)]]
-
- I'll probably follow the suggestion offered and write a "Firewalls
- Considered Harmful" paper. Meanwhile, here are some points:
-
- * I'm not advocating that companies should allow uncontrolled access
- to their networks -- that would be stupid.
-
- * A firewall is no excuse for lax internal security. To wit:
-
- - In a large organization, there are bound to be "bad guys" (either
- through malice, negligence, or sheer stupidity) inside the
- organization as well. No firewall is going to protect you against
- those.
-
- - A firewall only protects you against *known* external threats.
-
- - If your internal network is insecure, you are vulnerable to anyone
- who can get physical access to it. Today this involves tapping
- ethernet cables, but tomorrow it may just involve dropping by with
- a laptop with a wireless interface. I have a vested interest in
- seeing wireless LANs take off -- I don't want t hem stifled
- because of security concerns.
-
- - Think of the Maginot Line.
-
- * The network should switch bits and enforce routing policies -- not
- cover up for insecure applications.
-
- * It should not be the job of the millions of system administrators to
- patch known holes -- the vendors should be doing that. There is
- simply no excuse for vendors shipping us insecure code. (Is it true
- that SunOS is still distributed with /etc/hosts.equiv containing a
- single '+"? Why do we still have login programs that only accept
- eight-character passwords, password files that are publicly
- readable, things like NIS that allow uncontrolled access to their
- information, etc? At least we don't get sendmail shipped with the
- debug option turned on any more.
-
- * Having firewalls reduces the urgency (that is, the pressure on the
- vendors) of patching those security holes. It's a vicious cycle.
-
- * We've seen analogies such as putting locks on the front door rather
- than each individual room, and that it's perfectly acceptable
- capitalist behavior to put a firewall gateway in front of your
- network. I claim that this is far from being capitalistic; you're
- beeing communist inside, and hiding behind an Iron Curtain.
-
- * The argument "naive users and administrators don't want to deal with
- security" has been kicked around. I say that the systems should be
- secure from the beginning. I hope it's not too late to do that.
-
- * There are a lot of other security concerns in networked systems that
- should be addressed, that have nothing to do with firewalls. If
- those concerns are dealt with, firewalls will stop making. For
- example, I don't want anyone with the root password to be able to
- read my files, or log onto my machine and spy on what I'm doing.
- That includes the head of the security department, as well as the
- guy down the hall that I just had an argument with and wants to kill
- my files in revenge. While the latter is probably unavoidable, the
- former can be dealt with with proper cryptographic techniques.
-
- Finally,
-
- * Firewalls are an easy solution to a very real and very serious
- problem. My point, if a bit idealistic, is that we should *fix* the
- problem, rather than patch its manifestations.
-
- * Security, like good manners, starts at home.
-
- /ji
-
- In-Real-Life: John "Heldenprogrammer" Ioannidis
- E-Mail-To: ji@cs.columbia.edu
- V-Mail-To: +1 212 854 8120
- P-Mail-To: 450 Computer Science \n Columbia University \n New York, NY 10027
-