|
|||||||||||
|
Re: web application access control research
From: Jeff Williams (at) Aspect <(at)>
Date: Wed Apr 23 2003 - 20:30:30 EDT Andy, There is a tremendous body of literature on access control schemes for operating systems and databases. The only difference here is that the set of attributes on which access control decisions are made is a little different in the web application environment. Here's what I look for in a web app access control scheme...
The vast majority of access control components out there allow decisions to be made based on the URL. Period. Not form data, not query string, not time of day, age of user, account number, last page visited, special deal in effect, session data, or anything else. Most developers end up coding a bunch of special rules into their code and it quickly spirals out of control. Access control mechanisms should be centralized. By the way, the first three requirements are properties of a 'reference monitor' -- anyone implementing an access control scheme who hasn't heard those words, should find out what they mean and why they're important. I've always found that thinking through the access control matrix (of subjects and objects) raises many issues and special cases that are often overlooked without a systematic approach. --Jeff
Jeff Williams
>
This archive was generated by hypermail 2.1.8 : Wed Aug 23 2006 - 14:07:50 EDT |
||||||||||
|
|||||||||||