Freerider wrote:
> http://news.yahoo.com/s/cmp/20060606/tc_cmp/188701822;_ylt=AkrO5_Hlzsv7G4NNtBbYt9qor7oF;_ylu=X3oDMTBjMHVqMTQ4BHNlYwN5bnN1YmNhdA--
>
> >Windows 2003 Server is a more reliable server operating system than
> >Linux, a research firm said Monday.
>
> >According to the Yankee Group's annual server reliability survey, only
> >Unix-based operating systems such as HP-UX and Sun Solaris 10 beat
> >Windows on uptime. Windows 2003 Server, in fact, led the popular Red Hat
> >Enterprise Linux with nearly 20 percent more annual uptime.
I'm curious to see how this survey was conducted, how many servers were
checked.
I'm looking at the Netcraft Survey of Update for most frequently
requested sites.
http://uptime.netcraft.com/up/today/requested.html
Microsoft did pretty well as you might suspect - Average uptime 75
days. I would assume that this was due to routine reboots at 60 and 90
days depending on the server. Who am I to argue with Microsoft's
procedures. These routine reboots were probably excluded from the
Microsoft downtime numbers.
Lycos had average uptimes of 232 days. SCO had 198 days. Most of
these reboots were rprobably to install kernel upgrades to Linux,
usually to get radically better performance.
Of course, FreeBSD which needs almost no kernel updates has some of the
longest uptimes. Geocities has 287 day average updates, and Netcraft
with 290 day updates.
One has to be very careful to look at what is being measured. Taking a
machine down to install kernel upgrades is probably a reasonable thing
to do and could be excluded from the "downtime" calculation. On the
other hand, if you have to reboot your system every time you enhance or
upgrade a library or other minor application changes, this is a very
different situation.
> >On a broader note, said Yankee analyst Laura DiDio, the major
> >server operating systems all have a "high degree of reliability," and
> >have showed marked improvement in the last 3 to 5 years.
Looking at Longest uptimes from Netcraft
http://uptime.netcraft.com/up/today/top.avg.html
Actually, FreeBSD hasn't improved all that much, because it hasn't
needed to. There are FreeBSD and BSD servers that haven't been
rebooted for over 3 years. I do find it interesting that there are
some Windows servers that haven't been rebooted since 2003.
What makes it really interesting is that they would haven't been
rebooted since BEFORE Windows 2003 was officially released. A bit
suspicious if you ask me. Clearly things have improved, especially for
Windows 2003.
Linux administrators just couldn't resist the upgrades and performance
enhancements made available through kernel upgrades. Unfortunately,
this does have an impact on downtime.
> >On average, individual enterprise Windows, Linux, and Unix servers
> >experienced 3 to 5 failures per server per year in 2005, generating 10 to
> >19.5 hours of annual downtime for each server.
How many servers were measured, over what period of time? What
downtime was excluded from measurements? How much of this downtime was
for kernel upgrades? How much was for application upgrades? How much
were actual "crashes"? How many were "routine reboots"?
The general observations that all Operating systems were much more
reliable was very accurate. Windows has gone from Weekly reboots with
NT 4.0 to Monthly reboots for Windows 2000 to Bimonthly or quarterly
reboots for Windows 2003. That is testiment to the enhancements to
Windows 2003. Ironically, many of these, such as MTS, MSMQ, enhanced
IPC, and improved isolation between threads (apartment threading
similar to UNIX fork() ), and faster task switching has made it
possible to eliminate such high risk programming patterns as running
multiple threads in a single process/memory image (similar to posix
threads in Linux/Unix).
> >But standard Red Hat Enterprise Linux, and Linux distributions from
> >"niche" open source vendors, are offline more and longer than either
> >Windows or Unix competitors, the survey said. The reason: the scarcity of
> >Linux and open source documentation.
This is just an absurd obvervation. It even sounds a bit misleading.
Yes, there are numerous "niche" distributions, and some of these new
distributions aren't always optimimally configured - usually because
they have been tweaked to be a "generic" as possible. I just find it
curious that rather than singling out Red Hat and stating it's
performance, the paragraph is accurate ONLY by lumping Red Hat
Enterprise Linux with the other 300 Linux distributions.
> >The Yankee Group made a point of stressing that the survey was not
> >sponsored or supported by any server OS maker.
The lack of any details, combined with the peculair wording of the
final paragraph tends to eliminate any legitimate claim to
impartiality. The author may have had a personal bias, which may also
be legitimate. But this "Survey" certainly lacks the level of
information and detail to be considered a legitimately impartial and
objective survey.
> Laura DiDio = fat bitch
Really uncalled for.
http://www.yankeegroup.com/public/research/author_page.jsp?ID=2E01C474DD294963
On the lower half of the page are a number of articles she has written,
but you have to be a subscriber ($$$$).
> Yankee Group = Microsoft FUD puppet
Here's the board
http://www.yankeegroup.com/public/about/ab_boa_bios.jsp
Lot's of nobody's working for obscure companies. A few from Alta
Let's look at Alta Communications
http://www.altacomm.com/
An investment corporation.
Perhaps they are looking to sell off some Microsoft and buy into some
Red Hat :D
|
|
