Linux-Advocacy Digest #928, Volume #30 Sat, 16 Dec 00 09:13:03 EST
Contents:
Re: Uptimes ("Otto")
Re: Caifornia power shortage... (JFW)
Re: Caifornia power shortage... (JFW)
Re: Sun Microsystems and the end of Open Source (Gary Hallock)
Re: Sun Microsystems and the end of Open Source (Pan)
Re: Sun Microsystems and the end of Open Source (Pan)
Re: Sun Microsystems and the end of Open Source ("Chad Myers")
Re: Whistler review. ("Chad Myers")
----------------------------------------------------------------------------
Reply-To: "Otto" <[EMAIL PROTECTED]>
From: "Otto" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: Uptimes
Date: Sat, 16 Dec 2000 13:51:09 GMT
"R.E.Ballard ( Rex Ballard )" <[EMAIL PROTECTED]> wrote in message
news:91bht2$ii4$[EMAIL PROTECTED]...
: > I'd argue the cost of commercial NT based web site.
: > For less than half a million dollars,
: > you could have the necessary hardware and software.
:
: My numbers included staffing, consulting, and development
: costs for a typical enterprise specific environment. Sure,
: you could hang the Win2K and SQL Server on a few Compaqs or
: Netfinities, but you'd still have to make the content relevant,
: stress test your applications, tune the applications, and integrate
: to the rest of the enterprise.
The actual development and testing the web applications are pretty much the
same, regardless of the platform being used. Since there isn't much
difference, if any in hardware cost, then your numbers are simply false.
Remember, we are not talking about mom and pop web sites where you can use
some old 486 machines for web servers and be happy.
: > People tend to forget that OEM servers come
: > with 10 CALs for each servers.
:
: But if you are running e-mail, discussion groups, other "back-office"
: functions, or protracted SQL Server sessions, the license counts can
: escalate WELL beyond the 10 CALs in just a few minutes (Back-office
: funcions are based on the maximum number of uniquely identified users
: connected during the peak 1/2 hour period (not even 1/2 the peak-hour
: load).
Why in the world one would need back-office functions for a commercial web
site?
: If you are providing a web interface to a back-office server for
: messages, the licenses become unrealistic. Many companies have
: flipped from Back-office to UNIX/Pop3 simply because the cost of
: the UNIX server is a fraction of the cost of licenses, especially
: if the UNIX system is FreeBSD or Linux.
And many commercial entities simply don't use back-office for web interface.
:
: If you are connected directly to the internet, and you
: have a business event that triggers a huge traffic event,
: your half-hour peak could involve over 1 million CALS spread
: across 100 servers (the Victoria Secret video).
No it would not. IIS 4.0 has no licensing requirement for internet
connections.
:
: > Your recurring cost is way out of whack.
: > Salaries are salaries, regardless of the OS.
:
: But the Linux and UNIX administrators typically support an average
: of 20 CPUs/FTE compared to 5 CPUs/FTE in complex server (multiple
: services on a single server) environment. Much of this has to do
: with the fact that:
: First, Linux and UNIX provide full support for remote management
: (X11 and SSH), scripting (shells), scheduling (cron). Microsoft has
: shown contempt for scripting, scheduling, and command line interfaces.
The very same remote capabilities allow hackers to take over the *nix box.
It all depends on how competent your *nix admin is. I prefer a third party
software, ControlIT or PCAnywhere.
:
: Second, the servers can easily run multiple services on a single
: machine. Microsoft suffers from DLL hell. New features on
: Windows 2000 improve the situation somewhat, but the core strategy
: for getting high availability is to use redundant arrays of simple
: single-function servers. Even these servers must usually be SMP
: with RAID. And Applications must be recoded to COM+.
Is it MS which suffers from DLL hell, or is it the third party program's
installation routine which overwrite DLLs?
:
: Third, since the applications adhere to strict standards
: that don't change significantly over time (backward compatibility),
: there is less retraining. Microsoft considers frequently changing
: standards and backward incompatibility to be a critical strategic
: revenue source (you MUST buy upgrades).
Sure, you can use vi, but do you really want to? Define frequent and
"MUST"....
:
: Fourth, because Linux and FreeBSD include source code
: for critical infrastructure functions, there is a much
: better chance of preventing the recurrance of future problems.
: Microsoft never publishes critical infrastructure code (they
: offer free "examples" in their developer program), and they
: don't even acknowledge severe security holes until they have
: released a Patch (which usually breaks compatibilities).
The open/closed source has its own advantages/disadvantages. Problems will
resurface regardless of the platform.
: > And I had/have clients who had NDA with SUN, what's your point?
:
: I have had access to maintenance records and availability statistics
: on numerous systems over the last 10 years, and have a pretty good
: sense of the relative performance. The recent publication by Netcraft
: of "uptime charts" has shown that my availability numbers are pretty
: darn close.
I'd love to see NT's relative performance data from 1990 :). If you'd
compare NT 3.0 from 1993 to the Windows2000 of today, you could see some
obvious performance difference. NT is only 7 years old and it keeps getting
better.
: I'm not giving client names, I'm not giving actual statistical
: data, and I'm not publishing referrals. And I won't. People
: don't like it, but publishing this information wouldn't be
: appropriate. What I can say is that the data mentioned covers
: 3500 NT servers, 700 UNIX servers (Solaris and AIX), 100,000
: internal users and 1 million customers. As for the Mainframes
: (20) I only heard of two significant full failures in 2 years.
The data cited in such a matter is worthless, however well the intentions
are.
: > What is scalability has to do with availability (uptime)?
: > Clustering is for availability, does not improve scalability.
:
: This depends on how many servers you set up. If you can only
: configure two servers redundantly, then each has to be capable of
: handling the load of the partner, which means you have to buy
: twice as much hardware, software, resources... If you can
: spread the load evenly across 5 servers with failover spread
: amongst the remaining 4 machines in the case of a failure, you
: only need 20% more hardware.
Again, redundant configurations improve availability only and not
scalability. The apparent limitation of the platform does not change,
regardless how many servers one would install. The commercial web sites use
redundant configurations for availability no matter what the platform being
used. Having 4 or 5 web servers load balanced insures the five nines,
regardless of the platform being used.
Otto
------------------------------
From: JFW <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.ms-windows.nt.advocacy,comp.os.ms-windows.advocacy,comp.sys.mac.advocacy,comp.os.os2.advocacy,comp.unix.advocacy
Subject: Re: Caifornia power shortage...
Date: Sat, 16 Dec 2000 04:55:40 -0800
On Sat, 16 Dec 2000 00:15:32 GMT, Woofbert <[EMAIL PROTECTED]> wrote:
>In article <[EMAIL PROTECTED]>,
>[EMAIL PROTECTED] (The Ghost In The Machine) wrote:
>
[snippage]
>> Within these two methods, there are a number of ways -- plasma
>> bottling and laser implosion being the only two I'm familiar with at
>> all -- to induce the reaction.
>
>The 4 H => 1 He + E reaction requires amazing temperatures and
>pressures. IIRC, it's not even the reaction that happens in stars.
You're probably slapping your head about now for that statement, but I
gotta ask: What precisely is the form of hydrogen you envision being
common in interstellar space?
Of _course_ it's the 4 H => 1 He + E reaction. For Sol, anyways, for
now. The aberrant cases you mention like H2+H3 are meaningless in the
kind of pres/temp environments at the core of a star.
>The H2 + H3 => He + N + E reaction can happen at vastly lower
>temperatures and pressures.
Yes, but H2 and H3 cannot exist at _vastly_ higher temperatures and
pressures, combined with massive particle flux.
H2 + H3 => He + N + E has other problems, as well. It's difficult to
maintain H2 and H3 in the plasma chambers that are the current Tokomak
hopes. They work for single-shot thermonukes, but they're messy for
sustainable reactions.
>It is actually the reaction that powers
>H-bombs. (This is why the Germans were interested in heavy water during
>the war. Heavy water has H2 or even H3 instead of the usual H1.) Tritium
>is radioactive, but the neutrons that fly out of this reaction are very
>fast. What they hit becomes radioactive. The containment vessel will
>itself become radioactive after a while. The magnetic bottle can't
>contain the neutrons ... they are electrically neutral, after all.
The containment vessel is the least of the problem in sustaining H2/H3
reactions. You also rapidly get those stray high-energy neutrons
corrupting both the H2/H3 medium itself. Neither Deteurium nor
Tritium is remotely stable in a very high-energy neutron-rich
environment.
>> I for one wouldn't mind "breeder reactors". I'm a little annoyed at
>> people who do -- although there are issues regarding handling of
>> plutonium (the most poisonous substance on earth; a gram, if spread
>> into the air, would kill every human thereon).
<as an aside, but this was so egregious and nobody else mentioned it,
I felt I had to do so, to prevent further mindless fear-mongering>
Please recheck your sources, you're off by a few dozen orders of
magnitude on this one. Plutonium's nasty, but a single atom will not
kill a human, nor even a few hundred atoms. Do the math, figure out
how many Pu atoms exist in a gram of reactive Pu.
Dioxin, last I checked, held the winner as most poisonous substance on
Earth. And it's dispersable, unlike Pu, which tends to burn.
[snip]
>> There are also issues with those who would nab some of the plutonium
>> and make bombs with it.
>
>It's actually to our advantage to let people think this is easy. They'll
>try and get killed trying to steal it, mill it, or deploy it. Meanwhile
>they're forgetting that it's less risky to buy or steal fertilizer and
>kerosene, and the resulting bomb *will* work.
Further, there's the intrinsic joke of thinking that breeders are
somehow an "efficient" source of Pu. Let's be real here, if you've
got the cash, the fissionable material is there for the buying in the
Russo-Asian marketplace.
Problem is, as they're discovering, cheap nukes _aren't_ SMALL nukes.
And if you don't know what you're doing, trying to build a nuke tends
to kill you long before you get anywhere. And trying to get a big
nuke anywhere meaningful won't happen.
jfw
------------------------------
From: JFW <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.ms-windows.nt.advocacy,comp.os.ms-windows.advocacy,comp.sys.mac.advocacy,comp.os.os2.advocacy,comp.unix.advocacy
Subject: Re: Caifornia power shortage...
Date: Sat, 16 Dec 2000 05:02:24 -0800
On Sat, 16 Dec 2000 02:42:40 GMT, "Chad Myers"
<[EMAIL PROTECTED]> wrote:
>
>"Woofbert" <[EMAIL PROTECTED]> wrote in message
>news:[EMAIL PROTECTED]...
>> In article <[EMAIL PROTECTED]>,
>> [EMAIL PROTECTED] (The Ghost In The Machine) wrote:
>>
>> > In comp.os.linux.advocacy, Craig Gullixson <[EMAIL PROTECTED]>
>> > wrote on 13 Dec 2000 15:06:01 GMT <91838p$1i0e$[EMAIL PROTECTED]>:
>> > >Chad Myers wrote:
>> > >
>> > >Ah, the fusion thing. It has been expected that cheep fusion power
>> > >will be available in a 10 to 20 year timeframe since at least the
>> > >late '50s.
>>
>> Yes, I recall that sometime in the '80s some government agency scheduled
>> a major breakthrough in fusion technology for the year 2001.
>
>Well, the Tokomak in Japan has reached marginal success. I remember reading
>near the end of last year that they had managed to sustain a reaction for
>some definite amount of time which proved the theory realistic. A breakthrough
>next year wouldn't suprise me.
Okay, "marginal success" is VERY LOOSE in this case. Personally, I
think in 5-10 years, now that they've solved some of the critical
materials issues, and research containment issues, they might have a
reactor to test for a decade or two to prove it's safe.
I think functional fusion is still at least 20 years off. We might
push it into service before then, but if so, I suspect we're gonna
find crap that makes us DEARLY regret it.
>> The 4 H => 1 He + E reaction requires amazing temperatures and
>> pressures. IIRC, it's not even the reaction that happens in stars.
>
>Isn't this the reaction that the Tokomak uses? It uses a super-heated
>(several billion degrees, IIRC) to create a mass of plasma.
I've already addressed that that IS the reaction in stars, so I'll set
that aside.
It is also the reaction the Tokomak uses, IIRC (H2 and H3, again, not
being particularly survivable in the plasma bottle environment in
question).
jfw
------------------------------
Date: Sat, 16 Dec 2000 08:50:18 -0500
From: Gary Hallock <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Sun Microsystems and the end of Open Source
"Chad C. Mulligan" wrote:
>
> Au contraire, Gary, it is you who is confused. No movement for two years on
> Mozilla, the same for the 2.4 Kernel, the same for released versions of Star
> Office.
>
>
No development in Star Office in two years???? Had you bothered reading about
the release of Star Office as Open Source, you would realize that Sun made a
conscious decision to NOT release the 5.x version, but go straight to the 6.x
version that was currently in development. That was 2 months ago, not two
years.
http://linuxtoday.com/news_story.php3?ltsn=2000-10-13-002-21-NW-DT-SW
As for the 2.4 kernel, there has been lots of movement. The test12 version
just became available a few days ago:
ftp://ftp.kernel.org/pub/linux/kernel/v2.4/linux-2.4.0-test12.tar.bz2
The 2.4 kernel hasn't even been in development for 2 years.
And there has been a lot of movement in Mozilla as well. As usual, Chad, you
have no idea what you are talking about.
Gary
------------------------------
From: Pan <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Sun Microsystems and the end of Open Source
Date: Sat, 16 Dec 2000 05:55:51 -0800
Reply-To: [EMAIL PROTECTED]
"Chad C. Mulligan" wrote:
>
> Au contraire, Gary, it is you who is confused. No movement for two years on
> Mozilla,
I guess that beta version that I recently installed on my computer with
Komodo is an illusion?
> the same for the 2.4 Kernel,
...2.401 snapshot is installed on one of my pc's along with Mandrake 7.2
and running just fine, thanks.
> the same for released versions of Star
> Office.
Still refusing to admit you were wrong about that?
I keep waiting for the factual level of your posts to go up Chad. Am I
simply wainting in vain? Can your need to FUD linux be so strong that
you would feel a need to continually and repeatedlay lie about the state
of these projects, or are you just hopelessly out of touch with reality?
--
[EMAIL PROTECTED]
http://salvador.venice.ca.us
------------------------------
From: Pan <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Sun Microsystems and the end of Open Source
Date: Sat, 16 Dec 2000 06:01:44 -0800
Reply-To: [EMAIL PROTECTED]
"Chad C. Mulligan" wrote:
> So all the work Star Division did was just thrown in the trash?
Who said anything is being thrown away? See Star Portal for details.
> There isn't any progress or there would be a new certified release.
No, if there weren't any new progress, they would quit doing new
builds. They have now done 3 builds. 605, 609, and 613. Your argument
is tantamount to saying that there is no progress during the 9 months
gestation period during pregnancy just because the child hasn't been
born yet. It's a stupid argument. You should know better.
> This is just like Mozilla being on the verge of beta release for more than two
> years now. To coin a phrase "Where's the beef?"
For open office, it's a 6.13 alpha build. For the linux kernel, it is a
2.4x snapshot beta that has been released with some linux distris
including mandrake. I'm not sure what build mozilla is on.
--
[EMAIL PROTECTED]
http://salvador.venice.ca.us
------------------------------
From: "Chad Myers" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Sun Microsystems and the end of Open Source
Date: Sat, 16 Dec 2000 13:40:50 GMT
"Chad C. Mulligan" <[EMAIL PROTECTED]> wrote in message
news:HLE_5.27584$[EMAIL PROTECTED]...
>
> "Gary Hallock" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > "Chad C. Mulligan" wrote:
> >
> > > "
> > >
> > > Trends follow trends. Mozilla, Star Office, Kernel 2.4....... See the
> > > pattern.
> >
> > Nope, sorry, I see no trend, pattern, or connection. What in the world
> does
> > Sun releasing the source code for Star Office have to do with the 2.4
> kernel?
> > You seem to be quite confused.
> >
>
> Au contraire, Gary, it is you who is confused. No movement for two years on
> Mozilla, the same for the 2.4 Kernel, the same for released versions of Star
> Office.
No! No! No! Chad, remember Linus himself said to expect 2.4 by the end of the
year, remember! It's coming, I know it is! He promised! Linus never lies!
<grin>
What is it with large Open Source projects that just seem to stagnate? It
seems the OSS model doesn't scale well.
-Chad
------------------------------
From: "Chad Myers" <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy,comp.sys.mac.advocacy
Subject: Re: Whistler review.
Date: Sat, 16 Dec 2000 13:42:32 GMT
"Chad C. Mulligan" <[EMAIL PROTECTED]> wrote in message
news:pNC_5.27283$[EMAIL PROTECTED]...
>
> "Ketil Z Malde" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > "Chad C. Mulligan" <[EMAIL PROTECTED]> writes:
> >
> > >> A printer driver can crash a Windows box?
> >
> > > Actually No.
> >
> > And conclusive proof of that resides...where? Only in your vivid and
> > deranged imagination? Thought so.
> >
>
> Printer drivers run in user space therefore cannot crash a system, that and
> 7 years experience with the operating system an never getting a blue screen
> from the system caused by a printer driver.
Printer drivers are, however, display devices of sorts and therefore make
calls to the GDI and such and could possibly mangle an API or two and wreak
havoc.
-Chad
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list by posting to comp.os.linux.advocacy.
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Advocacy Digest
******************************
