I have been watching the progress of the new LEAF configuration mechanism with some interest and while I haven't had the time to follow the various threads closely, I have formed some impressions and I have some concerns.

From what I understand, the new configuration system relies heavily on (attribute,value) pairs. The idea seems to be to define these pairs centrally and then propagate them into the various packages using some sort of configuration API.

In one of my weaker moments, I introduced (attribute,value) pairs into Shorewall configuration management and it was a disaster. The idea was similar to today's sample Shorewall configurations but in my braindead implementation, each sample configuration included an /etc/shorewall/params file that tried to parameterize the user's entire Shorewall configuration via (attribute,value) pairs.

The notion worked extremely well for first-time users with simple requirements -- they were able to get their first firewall working very easily and were very happy; UNTIL... they had to make their first configuration change where I hadn't included an (attribute,value) pair that expressed their particular requirement.

These users were then left with the choice of:

a) Pleading to me for help.
b) Trying to understand BOTH the standard Shorewall configuration method AND my parameterizing scheme so they could extent the latter to conform to the former.
c) Abandoning the (attribute,value) method of configuration and using the native Shorewall configuration technique.

My conclusion was that the native Shorewall configuration method (table oriented) was the correct one for ALL users even if it required a bit more understanding on the part of the user to get their first firewall running. And that requirement for understanding Shorewall configuration has been largely ameliorated by the availability of the sample configurations and the QuickStart Guides.

The bottom line is that (attribute,value) configuration is much less flexible than the table-oriented configuration technique supported by Shorewall and by reducing Shorewall configuration to (attribute,value) pairs, you confine yourself to a very limited set of firewall applications. When users outgrow that limited set, they must undergo a paradigm shift and use a totally different configuration method.

As I said at the outset, these impressions/concerns are formed from an understanding of the current proposals that is far from complete. Corrections and criticisms are welcome...

-Tom
--
Tom Eastep \ Shorewall - iptables made easy
Shoreline, \ http://www.shorewall.net
Washington USA \ [EMAIL PROTECTED]



-------------------------------------------------------
This SF.net email is sponsored by: SlickEdit Inc. Develop an edge.
The most comprehensive and flexible code editor you can use.
Code faster. C/C++, C#, Java, HTML, XML, many more. FREE 30-Day Trial.
www.slickedit.com/sourceforge

_______________________________________________
leaf-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/leaf-devel

Reply via email to