On Tuesday 01 April 2003 09:47 am, Horatio B. Bogbindero wrote:

> let them use it. it is my belief that security is more about implementation
> and maintenance. the tools used just makes it either easier or harder to
> implement and maintain. of course, i would not like to place my money in a
> bank that uses insecure systems (even the preception of being insecure
> could cause problems for the bank).
>
> besides i remember that a bank even used plain-text protocols for their
> online transactions across the public internet!!! argh!!! so even if that
> bank used the most secure network tools in the whole wide world, it would
> still not make that banks network anymore secure.

No problem with that. Securing systems starts with software and ends with 
implementation. Windows, Linux, and other OSes must use that approach. Thus, 
security starts at the software level and ends in security implementation 
level.

Here is the problem with Microsoft. The OS's vulnerability always starts at 
the software level. This complicates the task of securing systems. It's 
already tough implementing secure measures with Windows or Linux. Windows 
further compounds the problem with weak points in software. 

I'm not saying that FOSS has no vulnerabilities, but let me illustrate the 
points with a hypothetical example. Note that I somehow use the terms 
"patches" and "fixes" synonymously.

Let's say a bank implements two redundant systems. System Penguin and System 
Redmond. The sysadmin finds that a cracker is probing his infrastructure for 
possible entry weak points, so he implements tougher security measures (like 
new and longer passwords, stricter permissions, etc...) for both systems. In 
both cases, the tougher security implementation strengthens both systems. 
System Penguin and System Redmond wins.

The cracker, undaunted by the tougher security measures, probes further and 
finds software flaws in both System Penguin and System Redmond which allows 
him opportunity to bypass the security implementation level and compromise 
the system. 

The sysadmin hears about  these vulnerabilities in Systems Penguin and 
Redmond. So he buckles down to work by isolating the affected systems, 
informs himself then searches for patches. 

He finds all patches for System Penguin within the span of 3 or 4 days at the 
most. He finds some patches for System Redmond. Sometimes he doesn't find any 
patches, because it will be incorporated 6-12 months later in a service pack, 
which the local distributor of the OS says the bank should pay for this 
service pack in the future. 

In this case, System Penguin wins because all patches are immediately 
available. He applies the patches to System Penguin. It's easy to apply 
patches to Penguin because software is discrete, which means that only the 
affected software are patched. The upgraded software will not break other 
apps in System Penguin due to the overall design of the OS. The end result is 
that System Penguin immediately becomes secure at the software level while 
System Redmond remains vulnerable.

But wait, suppose he finds patches for System Redmond. He applies this patches 
and gets the system on its feet. He somehow feels a sense of accomplishment 
after patching the vulnerabilities.

Several moments later, he finds that the patches break other stable 
functionalities in System Redmond. A user mentioned him about something going 
wrong in the system when previously it worked fine. The sysadmin tries to 
isolate the problem, and soon realizes that he can't. System Redmond is 
monolithic. The patch did fixed the vulnerability but created more pressing 
problems. In short, the patch didn't work. 

The next thing he does is reinstall System Redmond, hoping that it will never 
be probed again. And if it goes down again, he'll prepare in advance a 
hot-pluggable backup, or another redundant System Redmond B just in case. If 
he goes for installing System Redmond B, the bank incurs additional expenses 
(additional hardware, media, etc...) in preparing the backup system, plus the 
sysadmin effort is wasted. The cost of implementing roundabout measures 
compounds the problem. Notice that the troubles started from a patch that 
didn't work when it should work without breaking others.

The real problem remains unfixed and the bank must wait for 6-12 months for 
the service pack. In the meantime, the sysadmin must unecessarily devise 
temporary measures to minimize the problem, relying on sheer luck the system 
will remain unprobed. System Redmond remains vulnerable.

In short, when the security implementation level for both OS is the same, 
System Redmond loses because it has more unaudited weaknesses at the software 
level. Software level problems for System Redmond will not be a problem IF 
patches are available immediately, these patches work, and don't break things. 

But we are tired of patches that take too long. Sometimes the company that 
makes System Redmond will not admit there is a problem until the exploits 
gets worse and becomes a publicity problem for their company. We are tired of 
patches that don't work. 

Patches for System Penguin always work. 

When we can see through the innards of FOSS (including patches) and audit it, 
that's trust at the highest ideal comparable to the banking system's 
perception of trust. This trust spills over to customers, whether they 
understand the underlying technological platform or not.

As Yoda might say, trust in banking earns market share, and market share leads 
to profits. System Penguin wins.


Just thinking aloud with a light saber, 

mikol


_
Philippine Linux Users Group. Web site and archives at http://plug.linux.org.ph
To leave: send "unsubscribe" in the body to [EMAIL PROTECTED]

Fully Searchable Archives With Friendly Web Interface at http://marc.free.net.ph

To subscribe to the Linux Newbies' List: send "subscribe" in the body to [EMAIL 
PROTECTED]

Reply via email to