You want to spot the real pro in the machine room he (or she) has a vacuum cleaner in one hand and a Brother p-touch in the other.
I honestly beleive there is a direct linear relationship between the efficiency and uptime of a shop and how anal they are about labelling stuff. I want to open a front door of a rack and see every server's hostname and every removable media device clearly labelled. I know YOU know that that CD-ROM drive is drive D: on the frodo server but I ain't got time to try to figure that out. Even more importantly I similarly want EVERY cable in the back of the rack to have some kind of useful label on both ends (unless it is less than a foot long then just one end is OK).
Agasi is swiftly replacing Scott McNeally and Uncle Fester for the Ken Olsen Memorial Trophy for IT industry executive most likely to make an idiot of himself in public.
They need to ditch him fast and send Henning Kagermann with him and bring back Hasso Plattner, if he would come back.
The telescope would be built by a coalition of 15 Canadian universities and the engineering firm AMEC, the same company that built the keck telescope. Nowhere does it say that it will be installed in Canada. In fact the expectation is either Chile or Hawaii.
The telescope will be made of many (thousands) small mirrors each of which have to be controlled with an accuracy of approx. 1/14000 the thickness of a human hair. (According to interview on CBC Radio One aired earlier today.)
Need to work in your listening skills, 780 mirrors, not thousands. You also got the accuracy figure wrong but I don't recall exactly what the correct value was but the denominator started with a 2.
As for abnormality, hey, there's no free lunch, you want to cram 14 2Way servers and SAN switchs and GBE switchs into 7U of rack you gotta give something up. You want something that behaves like a "normal" computer then get 14 1U pizza boxes and break out the cable ties.
Some things REQUIRE solaris to run or for the company/vendor to support.
True, although less so than it used to be, however I think you'll find that most of these dinosaurs will react poorly when you say "Solaris 10 on x86"? and you'll be stuck with SPARC iron.
This announcement covers running Solaris 10 for x86 on Xeon EM64T or AMD blades (HS20 and LS20) it does NOT mean that Solaris will run on the JS20 PowerPC blades.
What Google does barely deserves the label clustering.
Actually that's not really fair, the problem is the term clustering has become overloaded. What Google does is would be more completely described as "shared nothing" distributed computing. They use cheap as chips iron beacuse nobody cares if a transaction fails, because no data is lost, the end user just pushes refresh. Similarily the various grid compute "clusters" (SETI, Folding@Home etc.) can recover from a lost unit of work by sending it out for reprocessing after a timeout (or IIRC SETI doesn't wait, every unit of work is sent out multiply and the results that do come back are compared).
If on the other hand you are dealing with applications that actually save data, silly little things like, oh, electronic funds transfers or credit card charges, that's a whole different class of problem.
The current 2 Way (HS20) Xeon DP EM64T based blades use SCSI drives (not hot swapable). So you can put 14 2 way servers in a 7U chassis. In addition to just density this also helps on saving other infrastructure for network cabling, KVM and power.
The 4 way blades (HS40) are still junk, old IA-32 only Xeon MP chips, IDE drives etc. If you have an application that must scale above 2 Way then don't use a blade for that application.
Not quite the same. While Intel hyperthreading is in broad terms a similar idea the implementation is different.
As I understand it SMT duplicates all of the stuff before the guts of the processor so there are two complete pipelines etc. whereas Intel has a single pipeline so when the processor switches from one thread to another there is overhead involved in refilling the pipeline from the second thread that does not exist in the POWER implementation.
You need a few things to be able to do this. One is what IBM calls the hypervisor, this is essentially privledged instructions on the chip that are used to control access to underlying system resources. These instructuions are isolated so that the operating system cannot execute them. Two is a service processor, this is a separate, separately powered special purpose processor that CAN access the hypervisor instructions it also does things like control power etc.. Additionally you need to give each PCI card it's own bus so that a card cannot issue bus level events that will disrupt other cards.
You could build a x86 server with a service processor (in fact IBM already does) and bus per slot PCI and still not be able to do this beacuse Intel hasn't built the virtualization instructions into the x86 architecture.
does the car run if the On Star stuff is disconnected or disabled
Yes it does, mine died on me last year, unfortunately the failure mode was such that the darn thing kept honking the horn and taking over the radio so I had to take it in. The dealer refused to just power it off even though I no longer pay for the service and insisted on replacing it (for free on warranty) the explaination for this was that lack of the OnStar unit could somehow cause the airbags to not function correctly. I consider that to be utter bullcrap. The OnStar box is well hidden way up behind the dash on the passenger side and I frankly lack the energy and expertise to locate and disconnet the thing myself.
their product gave CRM systems the name "CRM" in the first place
Huh? CRM as a term was widely in use back when SAP was a one product company. IBM used it to refer to their internal tools and for about 15 minutes in the 90s they had this company called Corepoint that was theoretically gonna sell the stuff.
Who is actually number one is widely debated, SAP in private will now admit that some of their early claims about the number of CRM customers they had were included mostly customers with the SD component of R/3 licensed as they had basically no installs of the actual standalone CRM module. These days they actually have real CRM customers.
they announced that not only would they continue to support PeopleSoft, but they would release a new version (in about three years) that would allow for a direct upgrade from PeopleSoft to a combined Oracle/PeopleSoft product.
Hands up anyone in the room who thinks Oracle will actually deliver on that? Oh they'll continue to support the PeopleSoft install based because the love the maintenance revenue stream but Fusion will either be massively late or the "direct upgrade" will prove to be a grade A pain in the nether parts, probably both of those things.
So, it makes no difference to SAP which database sits underneath their product.
True technically SAP runs on top of several databases including Oracle, in other ways the relationship is much less smooth, these days if you sell for SAP you have two standing orders 1) getting new customer licenses is nice, coverted PeopleSoft customers are better and 2) whatever you do if at all possible screw Oracle while you do it, if you can screw Microsoft too even better but the main thing is screw Oracle.
They have connectors which join SAP to the database in a transparent fashion.
Not actually but that's a close enough approximation for most purposes. If you could actually see the C language kernel code you'd be suprised how much code specific to each database there is, as it is there is also a fair bit of this in ABAP code too.
I imagine the phone lines between Armonk and Walldorf and Redmond and Walldorf are pretty busy now.
Now that this penny has dropped IBM has got to be running the calculus on how much they can afford to tick off Oracle by buying SAP. As things are today IBM does much more business with SAP than they do with Oracle so I'm guessing there's about a 50% chance they will enter the game now.
As part of the so called "Mainframe Charter" IBM has been doing a bunch of stuff to revitalize the community around mainframes. Starting point for reading is:
Give me a groupware calendar that handles multiple timezones please!
I know this will induce shudders, but this is something that Lotus Notes does handle very well. Can't personally say a lot else nice about Notes but time zone handling in the Calendar is excellent.
Actually those boxes have POWER5 chips not PPC. Similar but no the same.
The only thing that IBM sells with PPC chips is the JS20 Blade for the Bladecenter (not counting things with PPC as embedded processors).
There was video tape taken by one of the TV networks quoting Cretien saying that the Liberals would "get rid of" the GST.
Link please? I think you are misremembering how Ms. Copps got caught, which was exactly that, now I know it's hard to tell the two of them apart but Sheila at least speaks one of our official languages.
It was not revenue neutral. There were huge cost overruns in bringing in the GST, on the order of large tens of millions of dollars. Yet even taking the cost overruns into account the GST brought in hundreds of millions of dollars more than "expected" in the first eight months.
IIRC the GST program cost over $2B to implement, but I forget how much of that was cost overruns.
I think that Chretien's election promise to scrap the GST was one of the crassest political ploys I've ever witnessed
The only problem with that was that he never promised to scrap the GST, Sheila Copps did but she was pretty much alone on that, what the red book said was "replace the GST" which is a very different thing and actually a good idea if the provinces would have played ball but Little Mikey and King Ralph wanted to score political points rather than improve the tax regime so it didn't much happen.
The problem with the GST was not that it was a cash grab, it wasn't, it was at inception revenue neutral with the MST it replaced and as you point out transprency is a good thing too.
The problems with the GST were a) it was expensive to implement for both the government who spends gun registry dwarfing amounts of money setting up and running the system and for businesses to whom a VAT is a royal pain in the tookas, especially small businesses and b) it is too open for fraud and abuse, periodically people still get caught creating fake transactions with used cars and the like to bilk the feds of GST rebates, many more are never caught.I still think someone ought to fix the GST.
In much of the/. reading world and more importanly according to the/. system clock and I was always told that April Fool's pranks played after noon work in reverse, they make the teller the April Fool.
Slashdot Mirror
You want to spot the real pro in the machine room he (or she) has a vacuum cleaner in one hand and a Brother p-touch in the other. I honestly beleive there is a direct linear relationship between the efficiency and uptime of a shop and how anal they are about labelling stuff. I want to open a front door of a rack and see every server's hostname and every removable media device clearly labelled. I know YOU know that that CD-ROM drive is drive D: on the frodo server but I ain't got time to try to figure that out. Even more importantly I similarly want EVERY cable in the back of the rack to have some kind of useful label on both ends (unless it is less than a foot long then just one end is OK).
Agasi is swiftly replacing Scott McNeally and Uncle Fester for the Ken Olsen Memorial Trophy for IT industry executive most likely to make an idiot of himself in public. They need to ditch him fast and send Henning Kagermann with him and bring back Hasso Plattner, if he would come back.
The telescope would be built by a coalition of 15 Canadian universities and the engineering firm AMEC, the same company that built the keck telescope. Nowhere does it say that it will be installed in Canada. In fact the expectation is either Chile or Hawaii.
Need to work in your listening skills, 780 mirrors, not thousands. You also got the accuracy figure wrong but I don't recall exactly what the correct value was but the denominator started with a 2.
Patience grasshopper http://www.eweek.com/article2/0,1895,1879130,00.as p
As for abnormality, hey, there's no free lunch, you want to cram 14 2Way servers and SAN switchs and GBE switchs into 7U of rack you gotta give something up. You want something that behaves like a "normal" computer then get 14 1U pizza boxes and break out the cable ties.
True, although less so than it used to be, however I think you'll find that most of these dinosaurs will react poorly when you say "Solaris 10 on x86"? and you'll be stuck with SPARC iron.
This announcement covers running Solaris 10 for x86 on Xeon EM64T or AMD blades (HS20 and LS20) it does NOT mean that Solaris will run on the JS20 PowerPC blades.
Actually that's not really fair, the problem is the term clustering has become overloaded. What Google does is would be more completely described as "shared nothing" distributed computing. They use cheap as chips iron beacuse nobody cares if a transaction fails, because no data is lost, the end user just pushes refresh. Similarily the various grid compute "clusters" (SETI, Folding@Home etc.) can recover from a lost unit of work by sending it out for reprocessing after a timeout (or IIRC SETI doesn't wait, every unit of work is sent out multiply and the results that do come back are compared).
If on the other hand you are dealing with applications that actually save data, silly little things like, oh, electronic funds transfers or credit card charges, that's a whole different class of problem.
The current 2 Way (HS20) Xeon DP EM64T based blades use SCSI drives (not hot swapable). So you can put 14 2 way servers in a 7U chassis. In addition to just density this also helps on saving other infrastructure for network cabling, KVM and power.
The 4 way blades (HS40) are still junk, old IA-32 only Xeon MP chips, IDE drives etc. If you have an application that must scale above 2 Way then don't use a blade for that application.
That's the way it worked with POWER4 and before. POWER5 and i5OS use the microcode hypervisor and there is no host O/S anymore.
Not quite the same. While Intel hyperthreading is in broad terms a similar idea the implementation is different.
As I understand it SMT duplicates all of the stuff before the guts of the processor so there are two complete pipelines etc. whereas Intel has a single pipeline so when the processor switches from one thread to another there is overhead involved in refilling the pipeline from the second thread that does not exist in the POWER implementation.
You need a few things to be able to do this. One is what IBM calls the hypervisor, this is essentially privledged instructions on the chip that are used to control access to underlying system resources. These instructuions are isolated so that the operating system cannot execute them. Two is a service processor, this is a separate, separately powered special purpose processor that CAN access the hypervisor instructions it also does things like control power etc.. Additionally you need to give each PCI card it's own bus so that a card cannot issue bus level events that will disrupt other cards. You could build a x86 server with a service processor (in fact IBM already does) and bus per slot PCI and still not be able to do this beacuse Intel hasn't built the virtualization instructions into the x86 architecture.
Yes it does, mine died on me last year, unfortunately the failure mode was such that the darn thing kept honking the horn and taking over the radio so I had to take it in. The dealer refused to just power it off even though I no longer pay for the service and insisted on replacing it (for free on warranty) the explaination for this was that lack of the OnStar unit could somehow cause the airbags to not function correctly. I consider that to be utter bullcrap. The OnStar box is well hidden way up behind the dash on the passenger side and I frankly lack the energy and expertise to locate and disconnet the thing myself.
Huh? CRM as a term was widely in use back when SAP was a one product company. IBM used it to refer to their internal tools and for about 15 minutes in the 90s they had this company called Corepoint that was theoretically gonna sell the stuff.
Who is actually number one is widely debated, SAP in private will now admit that some of their early claims about the number of CRM customers they had were included mostly customers with the SD component of R/3 licensed as they had basically no installs of the actual standalone CRM module. These days they actually have real CRM customers.
Hands up anyone in the room who thinks Oracle will actually deliver on that? Oh they'll continue to support the PeopleSoft install based because the love the maintenance revenue stream but Fusion will either be massively late or the "direct upgrade" will prove to be a grade A pain in the nether parts, probably both of those things.
True technically SAP runs on top of several databases including Oracle, in other ways the relationship is much less smooth, these days if you sell for SAP you have two standing orders 1) getting new customer licenses is nice, coverted PeopleSoft customers are better and 2) whatever you do if at all possible screw Oracle while you do it, if you can screw Microsoft too even better but the main thing is screw Oracle.
They have connectors which join SAP to the database in a transparent fashion.
Not actually but that's a close enough approximation for most purposes. If you could actually see the C language kernel code you'd be suprised how much code specific to each database there is, as it is there is also a fair bit of this in ABAP code too.
I imagine the phone lines between Armonk and Walldorf and Redmond and Walldorf are pretty busy now. Now that this penny has dropped IBM has got to be running the calculus on how much they can afford to tick off Oracle by buying SAP. As things are today IBM does much more business with SAP than they do with Oracle so I'm guessing there's about a 50% chance they will enter the game now.
http://www-03.ibm.com/servers/eserver/zseries/newf aces/
Quebec has something like 25% of the population of Canada and produces 48% of the engineering and science graduates.
There are plenty effed up things about Le Belle Province but their approach to technology education isn't one of them.
Give me a groupware calendar that handles multiple timezones please! I know this will induce shudders, but this is something that Lotus Notes does handle very well. Can't personally say a lot else nice about Notes but time zone handling in the Calendar is excellent.
Actually those boxes have POWER5 chips not PPC. Similar but no the same. The only thing that IBM sells with PPC chips is the JS20 Blade for the Bladecenter (not counting things with PPC as embedded processors).
Link please? I think you are misremembering how Ms. Copps got caught, which was exactly that, now I know it's hard to tell the two of them apart but Sheila at least speaks one of our official languages.
It was not revenue neutral. There were huge cost overruns in bringing in the GST, on the order of large tens of millions of dollars. Yet even taking the cost overruns into account the GST brought in hundreds of millions of dollars more than "expected" in the first eight months.
IIRC the GST program cost over $2B to implement, but I forget how much of that was cost overruns.
MST revenues in the last full year it was in place (1989-90) were $17.7B, GST in the first full year it was in place (91-22) was $15.3B. See http://www.fin.gc.ca/frt/2004/frt04_2e.html%23Tabl e6
GST needs to be restructured/replaced because VAT's are a bad way to do consumption taxes.
The only problem with that was that he never promised to scrap the GST, Sheila Copps did but she was pretty much alone on that, what the red book said was "replace the GST" which is a very different thing and actually a good idea if the provinces would have played ball but Little Mikey and King Ralph wanted to score political points rather than improve the tax regime so it didn't much happen.
The problem with the GST was not that it was a cash grab, it wasn't, it was at inception revenue neutral with the MST it replaced and as you point out transprency is a good thing too.
The problems with the GST were a) it was expensive to implement for both the government who spends gun registry dwarfing amounts of money setting up and running the system and for businesses to whom a VAT is a royal pain in the tookas, especially small businesses and b) it is too open for fraud and abuse, periodically people still get caught creating fake transactions with used cars and the like to bilk the feds of GST rebates, many more are never caught.I still think someone ought to fix the GST.
In much of the /. reading world and more importanly according to the /. system clock and I was always told that April Fool's pranks played after noon work in reverse, they make the teller the April Fool.
I know not how my post became so botched, thank you for the correction...