I have never experianced any issues from using quad-ethernet in the past.
1. Neither did I. It is not a question of single, dual quad but of network technique - putting all ports on a single network segment.
2. Check carefully the ip_output and stuff and see when linux actually uses all ethernet ports of a quad ether connected to a single network.
3. Check with Microsoft docs.
4. From my point of view linux is right. Relying on abnormal arp behaviour as well as techniques incompatible with most switches is plain wrong. If you actually want to support more than one uplink to a switch there are legitimate ways to do it (vlan tagging, etc).
RedHat and Dell made a press release stating that the Dell server used in the test was a "Red Hat Ready" configuration.
Marketing, marketing, marketing...
This test is not about "bang for bucks" it's about sheer performance.
Put a GigE on the server if you want performance. No quad ether gets even close... Been there, seen it, using it at the moment... Long live Gigabit Networking...
Why the hell are we discussing a benchmark ran on a hardware config designed especially for NT:
1. MindCraft once again used a quad ether (but skipped anouncing it) and the infamous "EtherStripping" break your switch stuff. 2. Mindcraft once again used the Dell machine which has a RAID running better under NT than under Linux
The benchmark is faulty by design:
1. If you want these speeds you use a Gig Ether on the server in full duplex mode not a questionable technique that actually breaks lots of real networks. 2. If you want real OS becnhmarking you use an architecture that is equivalently supported by bothe OSes.
Overall:
I have tested Linux with GigE (it can almost pull physical speed on machines much cheaper than the Mindcraft Dell monster) and NT has been officially tested by most GigE manufacturers. The results used to be available at the packetengines site butit looks like they were dropped when moving the site to alcatel. Anybody a link please? I would not quote them so nobody blames me for flamebaiting...
It will be rather interesting if someone finally does this benchmark on a sanely designed network (no etherstripping BS) and with proper hardware.
To conclude I expected better from RH than accepting a doomed bench (on hardware and in a network setup where they cannot win).
I am running Netscape on Debian 2.1 upgraded to kernel 2.2.10 and it crashes about once a month (after I removed the braindead broken locale preloads coming with debian).
They have very strict restrictions on using wireless comms because the resistance of most medical eqipment to noise is about zilch...
The doctors will love it, but it will be not enough... All medical vendors will have to finally produce reasonably resilient equipment and this is least likely to happen
Jerry did have some good points there. As an ex-OS2 fan I have to admit it. He missed some points as well:
1. Merlin had "as released" networking far superior to Win95. It did have some problems working in peer only networks (without a dedicated server around) but these were not as bad as described. It was much faster in all network transfers, it had a very good TCP stack and had support for all network file systems including NFS.
2. What killed OS2 were stupid vendor agreements. A typical example is the NE2000 cards (most popular OEM network card at the time) working only with Novell support. So as a result an otherwise good system was intentionally crippled to fit something signed by somebody.
3. And Merlin as I recall did run non-microsoft Win32 apps very well. If you new how to install them you could happily run them. It actually even run most apps from Office 6. It did have some problems with MS apps but I would prefer to leave the why and how to the reader of the judge's findings.
But overall the points are on 2 paragraphs out of forgot how many.
So why the heck doesn't he write the same thing and shut up about it? A quick perl or even shell script involving find, diff, sum (md5sum) should easily suffice and could probably be knocked up in under 5 minutes flat.
And the result will be that it will have at least one symlink exploit, will crash the box on some directory structures or do something else as dumb as it gets.
Writing scripts like that is not a 5 min job (unless you want to provide a nice root-comporomise backdoor).
There is a person who does it from time to time. Known as the "Solar Designer". Unfortunately his (and Andrew Tridgell) security pacthes do not get into the mainstream kernel and this really sucks...
No suitable URL's unfortunately but reading the mailing list archives may help. There you can note that the folks at OpenBSD have been through fixing all of our favorite YAREs (YetAnotherRootExploit) and have gone into details like file descriptor leaks (ever head of those;-). And so on.
And I love their attitude (to Paul Vixie software especially). Running named chrooted and as a non-root user;-). And there are many other small peeks and pokes here and there that make it much more bulletproof even at factory defaults compared to RedHat for example.
That is besides support for every sensible auth method/technique under the sun (one time passwords, encryption everywhere, cypherchained blowfish for storing passwords, etc).
100tx network with isa cards (if you can find them... did they even make them?)
They did but not for long;-)
Overall this means something most people are missing - faster and cheaper mainboards. At the moment all the ISA wiring is putting quite a bit of design strain when you try to wire a good MB. With the ISA (and hopefully the bloody PCI/ISA bridge) gone the mainboard prices will go DOWN and speeds on local buses UP.
I don't see how taking out a power plant by cracking their network is any different than taking it out with a well-placed missile. At least no one is killed this way.
I doubt that "no one is killed" if you take out by any means one of the mission crytical systems in the "Tcarna Voda" Nuclear Fascility. More likely a few thousand killed at least.
Besides the rant thanks god they had some "outdated information" on buildings in belgrade, not on power plants...
The teacher should be fired and blacklisted to never ever work with kids again. Reasons:
Rating a badly written adolescent fantasy with a "perfect" rating
The most important part of a student teacher relationship is trust. A teacher who does not know/understand this a shame for the profession and must not be allowed to teach Overall if there will be less teachers like this one there will be less columbines...
This is a very interesting term - release of linux... If you think about it... The goood thing about it is that it never actually gets released... never finished... always something left to play with...
Why the hell do I remember that in US the "Unberable Lightness of Being" got an X rating and a few R ratings were put quite often here and there so that people for sure do not see "questionable material" like the "House of the Spirits", "Go and See", etc.
Not mentioning the cases when artists selfsensor themselves like with South Park for a US release.
Anyway let's do not talk about the bill of rights until US elects a woman for a president.
UNDERSCORE IS NOT A VALID CHARACTER IN A HOST OR DOMAIN NAME. Whoever used it did not read the RFCs
Unfortunately most nameserver software allows it and Windoze people love it. Anyway if you see it you are most likely dealing with people who never read (and use GUI for administration).
Yes, from what I know novocaine locally. And beware some really stupid drug tests may test you positive for a few months after that. But whatever, human stupidity does not have upper limits.
The answer to your question is in your mail. You have installed something else besides exchange and exchange. If you install only exchange your boot times will increase.
All you need is to install Exchange Only (without SQL, IIS, exploiter and miscorosft mail service). Then after 6 months of average office email (30 employees) you will have boot times exceeding 1 hour on a P-II 333.
I guess I don't get it. It seems like they are creating something similar to "DDE" in the Windows environment to support their application embedding operations.
At first DDE is not windows but OS2 technology predating windows by far
At second DDE was and is network aware (thanks god that scripting kiddiez do not know how to use it under M$)
At third there is only one step from DDE to OpenDoc which is not bad as technology
Why not use a good ORB and have the added benefit of network communications?
Because it stinks by design as a general purpose exchange mechanism
It will never be 100% compatible across platforms (there will always be thy black sheep , call it M$, call it something else)
It will always be slower then a properly written platform specific exchange mechanism
It will always lag behind in utilizing all the capabilities of a particular OS or desktop environment
You forget that it is likely that junk mail shall follow. And discovering that you read "the hanky panky site" by your parents, and oversealos neigbour, etc can sometime be a very unpleasant experience (that is besides all the junk in your mailboxen - both electronic and snail).
I am actually glad that these practices are explicitly prohibited in europe by the data protection act.
I'd rather swap the Wintel for an alpha for the difference. An alpha in UK costs around 1500 pounds which makes it cheaper then this overclocking stuff..
You have opened a 286-er or 386-er. Up to somewhere around 93 AMD was manufacturing under licence. After that it modified the 286 core and the 386 core to improve performance (by about 5-10% over intel) and boosted the frequencies compared to Intel. Then the war began as we know it.
Actually AMD has so far bested Intel sooner or later on every compatible category, the problem being that by that time Intel was delivering the "new bigger and greater". Though quite often it was worse than the AMD top of the line for the old design. Compare a P5 at 60 with an AMD X5 at 166 for example;-)
Every time intel actually took on a war with AMD for a compatible product it lost. Examples are numerous: 386 vs am386, 486 vs Enahcedam486 and X5, P5 vs K5-PR series, MMX vs K6, P6 (PPro, PII, PIII) vs k6-2/K6-3. So I guess it will NOT win this time. It will have to deploy the new latest and greatest (namely IA64) in order to win. And then the cycle will start a new. The interesting part being that now the timing GAP between them is much much shorter.
Trailing? This is a highly questionable statement. It has been trailing in ads and M$ style kick the baby marketing strategy, but hardly in CPU's. Let us see (sorry for the ugly format but Taco is censoring table tags):
Note that these are Top of the Line CPUs, not what was available at the same time. The idea is where does AMD get when it wants to develop a concept, not where it stands at the moment.
286 ranking: Harris - 25MHz with additional prefetch and optimized core; AMD - 20MHz with additional prefetch and optimized core; Intel with 16 MHz barely...
386(SX and DX) ranking 40 Mhz AMD with optimized mul, Intel trailing at 33 Mhz with a worse core. Cyrix was mostly doing coprocessors at the time and there was basically no match there.
486 clones 166 Mhz AMD with X5 - 64 (128) instructions prefetch, write-back L1 cache, etc. Trailed far by Intel with 100 Mhz DX4 which even did not have a proper write back cache. That was the peak point with UMC, Cyrix, ITT, TI wrestling for the branch.
Pentium and clonesAMD K5 166Mhz, followed by Intel P5 133, Followed by Cyrix.
Pentium MMX and clonesAgain K6 266 Mhz trailed by Pentium MMX and mobile pentium.
Optimized P6 like coresAgain K6-3 450 Mhz trailed by PIII.
Athlon - Here intel does not have an answer yet.
A note - so far Intel has used better marketing and came out with products before AMD. So the fact, that AMD blew it out of the water in every CPU category sooner or later was never taken into account. Now AMD came out with Athlon before Intel. The game started to be interesting...
Adn an additional last comment. Intel can raize their frequency to terahertz if they want but with their current bus it will be still slower then an Athlon...
I think we will see big changes in the way academic feedback occurs in the next decades. It is already happening in computer science and mathematics: ideas at an early stage are disseminated in academic mailing lists, getting a quite different kind of feedback before being submitted to classical journal review
Nope. There is one problem with the academic community. As long as you let the mediocracy in you can no longer get rid of it. Hence, check the policies of many mathematical and other scientific journals. Quite a lot of them explicitly restrict any prior publishing on mailing lists, web, discussion forums, etc. This is not amazing because when mediocracy has already prevailed there is no place for healthy critisism.
I am not saying that the open reviews shall not prevail, I am just saying that your time estimate is rather over optimistic. Some "academic" individuals with rather die instead of allowing an open review. So until they (and possibly one or two generations of the successors they have installed) go on a pension open reviews as a wide practice in the scientific communnity do not stand a chance.
Slashdot Mirror
I have never experianced any issues from using quad-ethernet in the past.
... Been there, seen it, using it at the moment... Long live Gigabit Networking...
1. Neither did I. It is not a question of single, dual quad but of network technique - putting all
ports on a single network segment.
2. Check carefully the ip_output and stuff and see when linux actually uses all ethernet ports of a quad ether connected to a single network.
3. Check with Microsoft docs.
4. From my point of view linux is right. Relying on abnormal arp behaviour as well as techniques
incompatible with most switches is plain wrong. If you actually want to support more than one uplink to a switch there are legitimate ways to do it (vlan tagging, etc).
RedHat and Dell made a press release stating that the Dell server used in the test was a "Red Hat Ready" configuration.
Marketing, marketing, marketing...
This test is not about "bang for bucks" it's about sheer performance.
Put a GigE on the server if you want performance. No quad ether gets even close
Why the hell are we discussing a benchmark ran on a hardware config designed especially for NT:
1. MindCraft once again used a quad ether (but skipped anouncing it) and the infamous "EtherStripping" break your switch stuff.
2. Mindcraft once again used the Dell machine which has a RAID running better under NT than under Linux
The benchmark is faulty by design:
1. If you want these speeds you use a Gig Ether on the server in full duplex mode not a questionable technique that actually breaks lots of real networks.
2. If you want real OS becnhmarking you use an architecture that is equivalently supported by bothe OSes.
Overall:
I have tested Linux with GigE (it can almost pull physical speed on machines much cheaper than the Mindcraft Dell monster) and NT has been officially tested by most GigE manufacturers. The results used to be available at the packetengines site butit looks like they were dropped when moving the site to alcatel. Anybody a link please? I would not quote them so nobody blames me for flamebaiting...
It will be rather interesting if someone finally does this benchmark on a sanely designed network (no etherstripping BS) and with proper hardware.
To conclude I expected better from RH than accepting a doomed bench (on hardware and in a network setup where they cannot win).
What system are you running?
I am running Netscape on Debian 2.1 upgraded to kernel 2.2.10 and it crashes about once a month (after I removed the braindead broken locale preloads coming with debian).
Nope. It will harldy get there...
They have very strict restrictions on using wireless comms because the resistance of most medical eqipment to noise is about zilch...
The doctors will love it, but it will be not enough... All medical vendors will have to finally produce reasonably resilient equipment and this is least likely to happen
Jerry did have some good points there. As an ex-OS2 fan I have to admit it. He missed some points as well:
1. Merlin had "as released" networking far superior to Win95. It did have some problems working in peer only networks (without a dedicated server around) but these were not as bad as described. It was much faster in all network transfers, it had a very good TCP stack and had support for all network file systems including NFS.
2. What killed OS2 were stupid vendor agreements. A typical example is the NE2000 cards (most popular OEM network card at the time) working only with Novell support. So as a result an otherwise good system was intentionally crippled to fit something signed by somebody.
3. And Merlin as I recall did run non-microsoft Win32 apps very well. If you new how to install them you could happily run them. It actually even run most apps from Office 6. It did have some problems with MS apps but I would prefer to leave the why and how to the reader of the judge's findings.
But overall the points are on 2 paragraphs out of forgot how many.
In btw: Why is the icon featuring Chuck. Chuck is the FreeBSD mascot. The OpenBSD one does not wear running shoes...
should easily suffice and could probably be knocked up in under 5 minutes flat.
And the result will be that it will have at least one symlink exploit, will crash the box on some directory structures or do something else as dumb as it gets.
Writing scripts like that is not a 5 min job (unless you want to provide a nice root-comporomise backdoor).
There is a person who does it from time to time. Known as the "Solar Designer". Unfortunately his (and Andrew Tridgell) security pacthes do not get into the mainstream kernel and this really sucks...
No suitable URL's unfortunately but reading the mailing list archives may help. There you can note that the folks at OpenBSD have been through fixing all of our favorite YAREs (YetAnotherRootExploit) and have gone into details like file descriptor leaks (ever head of those ;-). And so on.
;-). And there are many other small peeks and pokes here and there that make it much more bulletproof even at factory defaults compared to RedHat for example.
And I love their attitude (to Paul Vixie software especially). Running named chrooted and as a non-root user
That is besides support for every sensible auth method/technique under the sun (one time passwords, encryption everywhere, cypherchained blowfish for storing passwords, etc).
They did but not for long ;-)
Overall this means something most people are missing - faster and cheaper mainboards. At the moment all the ISA wiring is putting quite a bit of design strain when you try to wire a good MB. With the ISA (and hopefully the bloody PCI/ISA bridge) gone the mainboard prices will go DOWN and speeds on local buses UP.
I don't see how taking out a power plant by cracking their network is any different than taking it out with a well-placed missile. At least no
one is killed this way.
I doubt that "no one is killed" if you take out by any means one of the mission crytical systems in the "Tcarna Voda" Nuclear Fascility. More likely a few thousand killed at least.
Besides the rant thanks god they had some "outdated information" on buildings in belgrade, not on power plants...
Rating a badly written adolescent fantasy with a "perfect" rating
The most important part of a student teacher relationship is trust. A teacher who does not know/understand this a shame for the profession and must not be allowed to teach
Overall if there will be less teachers like this one there will be less columbines...
This is a very interesting term - release of linux... If you think about it ... The goood thing about it is that it never actually gets released ... never finished ... always something left to play with ...
Why the hell do I remember that in US the "Unberable Lightness of Being" got an X rating and a few R ratings were put quite often here and there so that people for sure do not see "questionable material" like the "House of the Spirits", "Go and See", etc.
Not mentioning the cases when artists selfsensor themselves like with South Park for a US release.
Anyway let's do not talk about the bill of rights until US elects a woman for a president.
UNDERSCORE IS NOT A VALID CHARACTER IN A HOST OR DOMAIN NAME. Whoever used it did not read the RFCs
Unfortunately most nameserver software allows it and Windoze people love it. Anyway if you see it you are most likely dealing with people who never read (and use GUI for administration).
Yes, from what I know novocaine locally. And beware some really stupid drug tests may test you positive for a few months after that. But whatever, human stupidity does not have upper limits.
The answer to your question is in your mail. You have installed something else besides exchange and exchange. If you install only exchange your boot times will increase.
All you need is to install Exchange Only (without SQL, IIS, exploiter and miscorosft mail service). Then after 6 months of average office email (30 employees) you will have boot times exceeding 1 hour on a P-II 333.
Same for shutdown.
Seen it, been there.
- At first DDE is not windows but OS2 technology predating windows by far
- At second DDE was and is network aware (thanks god that scripting kiddiez do not know how to use it under M$)
- At third there is only one step from DDE to OpenDoc which is not bad as technology
Why not use a good ORB and have the added benefit of network communications?Because it stinks by design as a general purpose exchange mechanism
You forget that it is likely that junk mail shall follow. And discovering that you read "the hanky panky site" by your parents, and oversealos neigbour, etc can sometime be a very unpleasant experience (that is besides all the junk in your mailboxen - both electronic and snail).
I am actually glad that these practices are explicitly prohibited in europe by the data protection act.
I'd rather swap the Wintel for an alpha for the difference. An alpha in UK costs around 1500 pounds which makes it cheaper then this overclocking stuff..
Very very very good point!!!!
You have opened a 286-er or 386-er. Up to somewhere around 93 AMD was manufacturing under licence. After that it modified the 286 core and the 386 core to improve performance (by about 5-10% over intel) and boosted the frequencies compared to Intel. Then the war began as we know it.
;-)
Actually AMD has so far bested Intel sooner or later on every compatible category, the problem being that by that time Intel was delivering the "new bigger and greater". Though quite often it was worse than the AMD top of the line for the old design. Compare a P5 at 60 with an AMD X5 at 166 for example
Every time intel actually took on a war with AMD for a compatible product it lost. Examples are numerous: 386 vs am386, 486 vs Enahcedam486 and X5, P5 vs K5-PR series, MMX vs K6, P6 (PPro, PII, PIII) vs k6-2/K6-3. So I guess it will NOT win this time. It will have to deploy the new latest and greatest (namely IA64) in order to win. And then the cycle will start a new. The interesting part being that now the timing GAP between them is much much shorter.
Note that these are Top of the Line CPUs, not what was available at the same time. The idea is where does AMD get when it wants to develop a concept, not where it stands at the moment.
A note - so far Intel has used better marketing and came out with products before AMD. So the fact, that AMD blew it out of the water in every CPU category sooner or later was never taken into account. Now AMD came out with Athlon before Intel. The game started to be interesting...
Adn an additional last comment. Intel can raize their frequency to terahertz if they want but with their current bus it will be still slower then an Athlon...
mathematics: ideas at an early stage are disseminated in academic mailing lists, getting a quite different kind of feedback before being
submitted to classical journal review
Nope. There is one problem with the academic community. As long as you let the mediocracy in you can no longer get rid of it. Hence, check the policies of many mathematical and other scientific journals. Quite a lot of them explicitly restrict any prior publishing on mailing lists, web, discussion forums, etc. This is not amazing because when mediocracy has already prevailed there is no place for healthy critisism.
I am not saying that the open reviews shall not prevail, I am just saying that your time estimate is rather over optimistic. Some "academic" individuals with rather die instead of allowing an open review. So until they (and possibly one or two generations of the successors they have installed) go on a pension open reviews as a wide practice in the scientific communnity do not stand a chance.