News from PANUG/BizNix - June 21, 2002
http://panug.org - http://biznix.org


FREE/BUSY DETAILS
At last night's PANUG meeting, Tom Rich, the presenter, was asked about
how Caldera's Volution Messaging Server handled free/busy information
for Microsoft Outlook clients. Tom sent us a link to a web page that
answers the question:

http://docsrv.caldera.com:8457/en/volutionmsg_ag/msgag.caladmin.html


WHY SOFTWARE IS SO BAD
or
"Software sucks because users demand it to."
By Charles C. Mann
http://www.technologyreview.com/articles/mann0702.asp?p=1


VERSION FATIGUE
This article, which is not especially well-written and doesn't go
as far as it could/should, points out the downside of products
whose user interface changes with each new version: users are less
willing to learn products when they know that that knowledge will
be made obsolete by the vendor before long. This is not a new
revelation. One PANUG member recognized this a year or two ago and
referred to it as "throwaway knowledge". Perhaps he might want
to write an article on the subject.

http://www.techcentralstation.com/1051/techwrapper.jsp?PID=1051-250&CID=1051-061902B


YET ANOTHER SECURITY INITIATIVE
In reponse to yet another serious security hole in Internet Explorer,
Microsoft announced it will put more effort and generally speed up
plans to rework/remove bad code from many of their products. The move
comes amid criticisms from industry analysts who point out that
Microsoft's "Trustworthy Computing" initiative, announced in 
January, has up to now been more hype than substance.

http://news.com.com/2100-1001-934363.html


APACHE ISSUES
by Ed Sawicki - Accelerated Learning Center / Tailored Computers

A security hole in the Apache web server has been discovered and fixed.
You should upgrade to version 1.3.26 or 2.0.39. The bug has to do
with requests that use chunked encoding. Note that the problem was
discovered by routine investigation by the developers. It was not
discovered and used by a hacker/attacker to attack systems (yet).
The full text of the issue is here:

http://httpd.apache.org/info/security_bulletin_20020620.txt

For those of you that enjoy the technical detail, this web page contains
interesting information about the issue of multi-threading versus forking
child processes - a perennial debate, especially between Linux
and Windows developers. Here's the relevant text from the page:

 In most cases the outcome of the invalid request is that the child
 process dealing with the request will terminate. At the least, this
 could help a remote attacker launch a denial of service attack as the
 parent process will eventually have to replace the terminated child
 process, and starting new children uses non-trivial amounts of
 resources.

 On the Windows and Netware platforms, Apache runs one multithreaded
 child process to service requests. The teardown and subsequent setup
 time to replace the lost child process presents a significant
 interruption of service. As the Windows and Netware ports create a new
 process and reread the configuration, rather than fork a child process,
 this delay is much more pronounced than on other platforms.

This seems to indicate that, in Apache's case, the pre-forking child
process model used by Linux and other Unix platforms is a more efficient
design -  especially when the server is under attack by a child killer.
The more efficient model also helps under conditions of varying load
when the parent (normally) needs to terminate and respawn children.
Another lesson from this is that, if high performance is important to
you, it's best to configure Apache so it keeps an abundance of idle
children around. This helps minimize the time-costly spawning process -
regardless of the platform and model used.