Security is almost always a trade-off between convenience and security. 
When it is not a trade-off, then security should always be chosen for
sure.

I have an interesting situation with my linode machine.  I was going to
implement a fancy iptables firewall for it until I realized that the
only open ports to the world are ports I want open so I'd have to punch
them through the firewall anyway.  By careful configuration of my
services, I have mysql and postgresql both running, but using unix
sockets only.  Then I have mail which is currently only listening on
127.0.0.1.  I have a couple of misc services which also only listen on
127.0.0.1.  This leaves just web, pop3 (special pop3->hotmail gateway),
and ssh open.  As far as I can tell at the moment a firewall won't help
me.  I think that a firewall is good, but often a crutch for
misconfigured services.  How many times have you heard, "I'm safe--I
have a firewall."  Then find out they are running IIS.

With the advent of IPv6, the need for NAT is eliminated (the way we
currently use it is very broken), many people are a bit scared about
giving up the protection that a NAT gives them.  When dealing with a
network with too many unknowns, a good firewall on IPv6 will give you
the same security features as a NAT, but without breaking the
fundamental end-to-end model that makes the Internet work.

The moral of the story is firewalls are good things, but not always
necessary, and good configuration on services is probably more
important.

Michael


On Fri, 2004-03-05 at 16:22, Andrew Jorgensen wrote:
> I know that this is opening a can of worms, but I'm genuinely interested
> in what people think about this:
> 
> Isn't security really an economic problem? It seems like the real goal
> (since perfect security isn't possible) is to make defeating your
> security measures too expensive to be worth defeating. That is, if it
> costs me more time / money / resources to defeat you than I will gain by
> doing so, then I just won't.
> 
> And on the flip side if it will cost me more to protect my assets than
> I'd lose if someone took them then I also shouldn't bother.
> 
> Of course, really you don't just give up, or leave a system completely
> open just because you don't have anything of value on it. But you might
> spend less keeping your home computer secure than you would keeping your
> department server secure. (When I say spend, I mean effort, time, etc.,
> not just money).
> 
> Also, you aren't just protecting your assets from theft, but damage. 
> And don't forget availability, or liability (if your machine is used to
> attack another system). But my point is that all of these should be
> figured into your potential losses when deciding how much (effort, time,
> money) to spend protecting yourself.
> 
> Why is it okay to use a credit card online? (or anywhere for that
> matter)? SSL has very little to do with it. The real reason it's okay to
> use credit cards is that you can dispute the charges if it gets stolen.
> The reason you make sure you're submitting your info over SSL is to see
> that it's not trivial to steal. If it's as easy as running ettercap for
> a few minutes to 'earn' a few thousand dollars why not?
> 
> I think this is true outside of technology as well. If I had $10 and got
> mugged I'd hand over the money. Why? Because the mugger may do more than
> $10 damage to me if I don't (even if I win or escape).
> 
> So to relate this back to the issue of implementing your own CHAP, it
> seems like if you can make it very difficult to break (though clearly
> not impossible) and you don't have anything vital to protect then your
> own efforts are probably good enough.
> 
> Am I wrong?
> 
> Now you may say that breaking strong crypto isn't just expensive it's
> impossible for the foreseeable future, but that's only if you ignore the
> possibility that there's some other way of breaking it that you haven't
> thought of (quantum computing, that algebraic attack on AES that hasn't
> been proven yet). Still my point is that it would cost too much to break
> for it to be worth the effort. 
> 
> 
> ____________________
> BYU Unix Users Group 
> http://uug.byu.edu/
> ___________________________________________________________________
> List Info: http://uug.byu.edu/cgi-bin/mailman/listinfo/uug-list
-- 
Michael Torrie <[EMAIL PROTECTED]>

____________________
BYU Unix Users Group 
http://uug.byu.edu/ 
___________________________________________________________________
List Info: http://uug.byu.edu/cgi-bin/mailman/listinfo/uug-list

Reply via email to