For me it all boils down to risk analysis. I do not believe that any system
is Hack proof. Spend enough time and money and you will crack anything. My
concern is to assess the degree of additional risk created by opening port
443. I agree that if there is a risk that even one person will gain access
then you must keep your patches upto date. However patches only fix known
bugs :( , hence my current preference to restrict by ip address (as a first
line of defence) as this at least reduces the number of potential attacks.

I accept your point that all it takes is for our "friends" to shift emphasis
to 443 and things will be as bad as ever.

I think the advice has to be VPN

rgds

gmcb

-----Original Message-----
From: Nick Holland [mailto:[EMAIL PROTECTED]]
Sent: 06 March 2002 16:52
To: [EMAIL PROTECTED]
Subject: Re: [gb-users] Security Issues


[EMAIL PROTECTED] wrote:
> 
> I work for a Company that has developed a web based Warehousing system for
> Vendor Managed Inventory. This has been very successful for us, however
with
> the increase in attacks from virus's we made a decision to restrict access
> by ip. This has been very successful from the virus point of view (0
attacks
> in the last year). However our ever enthusiastic salesforce are pushing
for
> a relaxing of this.
> 
> If we were to remove the ip restriction, block port 80 and allow only
access
> via 443 would this in the panels opinion represent a massive increase in
> risk? We have considered VPN, however the thought of talking a non
> IT-literate person through the installation and setup when he speaks no
> English and I speak nothing else and he is umpteen thousand miles away
just
> doesn't work for me.

Scenereo:
Let's make the assumption you are running NT,2000, or XP.  

Let's say you did this a year ago...  CodeRed comes along, and you
snicker and pat yourself on the back, as your port 80 blocking very
effectively prevented your machine from being infected...who needs to
worry about patch kits?

Ah, but the vulnerability is still there.  While your machine may not
even see the CodeRed probes (due to firewall filtering), a person who
uses the CodeRed technique on you, refocused on port 443, will find
your box as much a pushover as any (or even more, if you followed my
"who needs to worry..." (il)logic)


If your web server and application is vulnerable to attack, it is
vulnerable to attack.  Period.  If you wish to filter it to "known"
sites, this helps a lot...but even then, this is not true security.

Scenereo:
100% outside your control, your client has a vulnerable system.  It is
cracked.  It is now used as a springboard to attack your system.  Now
see first scenereo.

Unlikely?  maybe...maybe not.  If someone is going to target
you...why?  Two reasons I can think of: 1) Because they can (or more
accurately, want to show they can).  2) Because you have something
they want.  Who would find you have something they want?  A competitor
of one of your customers, or a competitor of yours.  Or, perhaps, your
customer, looking for information on their competitors (your other
customers).  Or, anyone looking for priveledged information that might
be on your site for any reason.  Filtering by user will prevent you
from being randomly attacked, but the person attacking *YOU* might
just have to be more creative.

You can't filter your way out of broken...or breakable...software. 
The best you can do is isolate to minimize the dammage and maybe make
it take longer to hit paydirt...long enough that hopefully you can
catch and stop it before something really bad happens.  If your
application is vulnerable to attack, you have to fix it, not filter
your way to "safety".  Filtering will help protect you from
vulnerabilities in areas OTHER than your app (i.e., if your OS of
choice has a vulnerable service you don't need, filtering can
eliminate that risk)

Nick.
-- 
http://www.holland-consulting.net

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
To subscribe to the digest version first unsubscribe, then
 e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
To subscribe to the digest version first unsubscribe, then
 e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to