Because they are... on a theoretical standpoint. Take a DVD drive that was designed long before there were recordable formats and see which it reads better. Odds are, the +R format will be better supported because it closely matches the format of a pressed DVD-ROM.
Everything they tested was new, modern toys. Did they bother to mention how much of that stuff was explicitly designed to read DVD-R/RW discs? No. They didn't test compatibility; they tested supportablity.
And wherein is the problem with statistics... one false assumption leads to a shuttle exploding. All machines don't have to be operational for the entire system to be "up" -- the entire office doesn't come to a halt because one person's computer is broken.
Let M be the number of operating machines: M=N - normal system M<N - degraded system M<<N - critical system; where that line falls depends on specifics. M=0 - "down" system
As such, the more machines added to the server farm, the less likely the entire system will be unavailable.
Ancillary costs... you don't need an expensive switch; you need two cheap switches and two nics per server -- that's two 100$ switches and an additional 10$ nic per server. Yes, power costs money... a 300W PS drawing its maximum power (which it better not be doing) at the industry average $0.07/kWh is $184.09 per year on average. KVM's are cheap. IP address space from your local ISP is also fairly cheap -- and there's really no need to give every machine a public IP address. Every router on the market supports NAT -- some of them support server load balancing (cisco "SLB"). Setting up a linux box to do server load balancing ("linux virtual server") is pretty easy and getting easier every day. Outside rack space is the only measurablly expensive part, but it's generally cheaper than plain office space.
Re:Faster is not the only concern for many busines
on
Can .NET Really Scale?
·
· Score: 1
Actually, I'd have to say this is not true. There is such a laughablly low demand for co-lo (find me a colo that doesn't have a bunch of empty space) that you can get as much space as you need for whatever you're willing to pay for it.
Sure, approaching schmucks like Exodus and Rackspace.com will cost you tremendously. But there are other (less advertised) facilities that have been sitting completely empty for years now.
And this is assuming there's a real need to colo everything. Several dozen mid-tower sized desktop "servers" will fit perfectly in the space of one office cubicle. The bandwidth requirements haven't been provided, but I'd assume a T1 or two would be sufficient.
There are several reasons... first, how can they prove you bought a cisco router? Any schmuck could then download it -- it doesn't matter that it doesn't run under windows. Second, there's next to no way to ensure you download the correct version and feature level that you are entitled to. With access to the software center, I can download any IOS image for any device with any feature-set. I can download the "kitchen sink" feature image to a router with no entitlement for those features (i.e. firewall feature set is very much not a free thing.) Of course, Cisco does log downloads, but I've only rarely been contacted about an image download (for a quality survey.)
Most people can simply get the image themselves. Others (read: less likely to be targeted immediately) can ask Cisco for the updated version and Cisco will get back with them -- of course, Cisco is under some load today...
What hardware and IOS were you using? I saw the same "output queue" filling problem with a 7401 a year ago. And it took bloody forEVER to get them to fix it. (it was a frame-relay packet queing priority error -- lmi status would get pushed into the head of the packet queue, transmit, and not be removed... 21 minutes later, the interface dies.)
(That's why my 7401's are running the never-released 12.2.9S. And will not be changing them even at gun point.)
What's with all the nuts blindly grabbing new IOS images and reloading routers? Do they not know how much new shit they may be breaking in the process.
I have yet to update an IOS version and not have some new inventively screwed up. (bugs fixed in one release, undone in another and so forth.) Cisco makes good stuff, but their quality has fallen through the floor in recent years.
Somehow, I cannot see people paying 10,000$ for Ciscoworks just to upgrade all their routers at once. And judging from the way most of CW "works" I don't f***kin' trust it to muck with flash and reboot routers.
To be accurate, it's a fire code issue. And one for which I'm very surprised there haven't been any fines. (Yes, the facilities are inspected regularly.)
Or they think too far... Where I work, should there be a power failure (and this would require a major event to keep the power out for more than a few minutes -- the building is classified as an EOC) when you are not on the first or second floor, you're totally screwed. All the stairs are behind gated doors. Without power from the security systems, the doors will not open. And very few people have an actual, physical key to open the doors.
I've worked here for 2+ years. I've been cursing the security system since the first day the power went out -- it doesn't take much to make that thing lose what little brain it has. (Plus, it was in a closet without a lock.)
Generally, I'd have to agree. But this is one of the cases to prove the point... the windows based solution is proving to be much less of a headache. (in fact, almost zero headache.
Re:Not interested in being acquired?
on
Darl McBride Interview
·
· Score: 2, Interesting
It's the same old expensive, complicated licensed, shit. Even today's "modern" versions don't stand up against linux v0.98 from the mid '90s.
I've thrown around the name "SCO" as a bad joke for ~15 years (or more.) It really is laughable to see McBride's comments. With 20 years exposure to the computing world, I've seen exactly 2, yes, 2, SCO systems in active deployment (and one in the test lab at NCR -- the backend for the cash registers.) One was a Lucent VoIP gateway and the other is an ACD manager for a Meridian phone switch. (the former never went anywhere and the later has been replaced with a windows box.)
ATI does have mobility drivers, however, they do not provide them to end users as the mobility chips are not made available to end users -- they are for system (read: laptop) manufacturers. You're supposed to bug your laptop maker for new ("less bug ridden") drivers.
In my experience, all of ATI's drivers are a pile of poo. If you ever find one that works for you, don't f**king change it. (And make thousands of copies of the installer so you don't lose it when ATI deletes it.)
Well, there's cheap, and then there's cheap... And then there's the people that do things like you:-)
If you have a business need for all those things, then you certainly have a business large enough to pay for device(s) to do all those things. The more VPNs you need, the more CPU will be required to handle them. The brain inside a Linksys is, well, very stupid. 10/8 doesn't count. Linksys routers have had syslog capability for years. And IDS is well beyond the scope of any "mini router" (and capability of a 386) and is certainly more complicated than it's worth to roll your own. Most of the "junk" selected to be a "mini router" would add new meaning to suck if used as a DNS or Web cache.
It's much faster and easier to hand over 60$ for a device that can do ~80% of the things I'd like to get done than it would be to spend weeks building and tuning a custom solution that no one else knows how to manage (and for which there is no one else to turn for assistance.) Yes, I used to be a "I'll do it myself" person, but there aren't enough hours in the year to do everything myself.
ISDN is not line powered, but it's still attached to 911 service. That said, there's a difference between "life line" and "911" service. And if your house is on fire, the first thing you should do is leave; call for help from somewhere else.
Indeed. JFS, NUMA, and RCU did not exist in '85 when the original code was licensed. And it's fairly obvious there was no reciprocal agreement or Unixware/OpenServer would've had those features long ago.
Now, crazy licenses like that do exist. I have an NDA from USR hanging on the back of my front door (someday I'll frame it) that boils down to "we'll give you the source code as long as you hand over every change you make. For free." I laughed my ass off at that -- and obviously never signed it.
Perhaps, but they cannot march into your office and demand that you delete all the files, et. al. that were created from that software (read: derived works)
Even after the license has expired, SCO cannot demand IBM revoke every copy of AIX in the world simply because SCO doesn't own what IBM developed. They can demand their own code be returned or destroyed which would leave gaps in the AIX source, but there's nothing at all to stop IBM from continuing to support the existing AIX installed base. IBM may be prohibited from selling any more AIX, but the existing copies will continue to exist. (royalties and licensing fees were already paid for on them.)
There's a little legal thing called ex post facto... No legal team is going to allow a licensing agreement where each party does not retain ownership of their respective code, or has provisions for retroactive dissolution of the agreement. IBM has always had good lawyers.
No, it means a lot of people THINK SCO is gonna make a lot of money out of this. Anyone who understands the issues generally agrees SCO has no case. IBM very certainly knows they have no case. As such, IBM will not settle or buy SCO. This whole mess is years from reaching a court and more years from any actual outcome. SCO does not have the resources to fight a protracted, decade long court battle. IBM has resources; they have products and customers and market, unlike SCO.
Actually, I was just thinking this should be setting off warnings at the SEC. It may be that SCO is perpetrating a "fraud" to inflate their stock price. However, until the mythical 80 lines of code appear in court for all to see, there simply isn't a single molecule of usable evidence: exactly who copied what from whom? I'd like to know. I'd also like to see them back up their "enterprise" claims. (and the ancient egyptians were too technologically stupid to build the pyramids; it had to be aliens.)
There's a growing mound of questions for which there are no answers. SCO has been very crafty and manipulative. No one qualified to make any informed judgement will agree to the NDA. The only people who have signed the NDA and seen what SCO wants them to see are "technical" idiots -- they could've been shown the code equivalent of cave art drawn in red crayon and not known any better. They are certainly unaware of the volume and extent of UNIX(tm) code that has been used to teach people to program.
SCO's lawyers need to read the GPL again. (or the monkeys talking to the press need to listen to those lawyers.) If you modify code covered by the GPL, your changes fall under the GPL as well, esp. if you publish your changes.
ISDN is simply expensive (unless you live in TN.) It's not at all difficult. There are a few extra steps beyond "plug in line", but really, how hard is it to enter the switch type and SPIDs provided by the telco? (If you can get ISDN, you can get IDSL... where I work, we have IDSL lines in SC terminated on a DSLAM back in Raleigh, NC. As long as you are withing 50,000ft of an ISDN capable CO, the line can be terminated anywhere.)
Satellite is a pain in the ass for what you get out of it. You cannot install it yourself (legally) -- and most of the people paid to install it aren't very good at it.
I have ISDN. Yes, it's twice the cost of a cable modem and ~1.5x the cost of IDSL. However, I've had this line longer than either cable modems or DSL have existed. Aside from the CO crashing once, I've had no problems with it.
2.5 is nowhere near ready to be called 2.6. There's just too much stuff that's been changed that hasn't made it to all corners of the kernel. And the kernel maintainers have started taking a Microsoft-like stance of publicly stating they'll remove some stuff that doesn't work ("how many people still need support for Adaptec 1542CF's anyway?")
I don't buy it. The plane is constructed with the same wiring as everything else. If this were the case, there would be a great deal of things on the ground screwed up by cell phones by the nature of the tiny wires/traces within them.
Signals inside waveguides are isolated from the junk outside the guide just as the junk outside is isolated from the signal inside. As such, one would have to place their cell phone inside the waveguide to mess anything up.
I'm sure politics is to blame for the lack of microwave ovens. A microwave oven (a broken/leaky one at least) produces several hundred thousand times the RF of a cell phone -- the result would be just as bad (or worse) for the humans as the electronics.
That space is referred to as "the swamp" (and for good reason.) That space should still be globally routable. My BGP tables have thousands of/24's in there (even from Sprint who publicly state/20 is as small as they go.) And SAVVIS is sending things as small as/29's -- their inbound filter will allow us to send them/32's!
ARIN has specific guidelines for returning address space and renumbering. Basically, they give you the space you can actually prove you need with some renumbering grace period afterwhich your original allocation is revoked.
Slashdot Mirror
Because they are... on a theoretical standpoint. Take a DVD drive that was designed long before there were recordable formats and see which it reads better. Odds are, the +R format will be better supported because it closely matches the format of a pressed DVD-ROM.
Everything they tested was new, modern toys. Did they bother to mention how much of that stuff was explicitly designed to read DVD-R/RW discs? No. They didn't test compatibility; they tested supportablity.
And wherein is the problem with statistics... one false assumption leads to a shuttle exploding. All machines don't have to be operational for the entire system to be "up" -- the entire office doesn't come to a halt because one person's computer is broken.
Let M be the number of operating machines:
M=N - normal system
M<N - degraded system
M<<N - critical system; where that line falls depends on specifics.
M=0 - "down" system
As such, the more machines added to the server farm, the less likely the entire system will be unavailable.
Ancillary costs... you don't need an expensive switch; you need two cheap switches and two nics per server -- that's two 100$ switches and an additional 10$ nic per server. Yes, power costs money... a 300W PS drawing its maximum power (which it better not be doing) at the industry average $0.07/kWh is $184.09 per year on average. KVM's are cheap. IP address space from your local ISP is also fairly cheap -- and there's really no need to give every machine a public IP address. Every router on the market supports NAT -- some of them support server load balancing (cisco "SLB"). Setting up a linux box to do server load balancing ("linux virtual server") is pretty easy and getting easier every day. Outside rack space is the only measurablly expensive part, but it's generally cheaper than plain office space.
Actually, I'd have to say this is not true. There is such a laughablly low demand for co-lo (find me a colo that doesn't have a bunch of empty space) that you can get as much space as you need for whatever you're willing to pay for it.
Sure, approaching schmucks like Exodus and Rackspace.com will cost you tremendously. But there are other (less advertised) facilities that have been sitting completely empty for years now.
And this is assuming there's a real need to colo everything. Several dozen mid-tower sized desktop "servers" will fit perfectly in the space of one office cubicle. The bandwidth requirements haven't been provided, but I'd assume a T1 or two would be sufficient.
There are several reasons... first, how can they prove you bought a cisco router? Any schmuck could then download it -- it doesn't matter that it doesn't run under windows. Second, there's next to no way to ensure you download the correct version and feature level that you are entitled to. With access to the software center, I can download any IOS image for any device with any feature-set. I can download the "kitchen sink" feature image to a router with no entitlement for those features (i.e. firewall feature set is very much not a free thing.) Of course, Cisco does log downloads, but I've only rarely been contacted about an image download (for a quality survey.)
Most people can simply get the image themselves. Others (read: less likely to be targeted immediately) can ask Cisco for the updated version and Cisco will get back with them -- of course, Cisco is under some load today...
What hardware and IOS were you using? I saw the same "output queue" filling problem with a 7401 a year ago. And it took bloody forEVER to get them to fix it. (it was a frame-relay packet queing priority error -- lmi status would get pushed into the head of the packet queue, transmit, and not be removed... 21 minutes later, the interface dies.)
(That's why my 7401's are running the never-released 12.2.9S. And will not be changing them even at gun point.)
What's with all the nuts blindly grabbing new IOS images and reloading routers? Do they not know how much new shit they may be breaking in the process.
I have yet to update an IOS version and not have some new inventively screwed up. (bugs fixed in one release, undone in another and so forth.) Cisco makes good stuff, but their quality has fallen through the floor in recent years.
Somehow, I cannot see people paying 10,000$ for Ciscoworks just to upgrade all their routers at once. And judging from the way most of CW "works" I don't f***kin' trust it to muck with flash and reboot routers.
To be accurate, it's a fire code issue. And one for which I'm very surprised there haven't been any fines. (Yes, the facilities are inspected regularly.)
Or they think too far... Where I work, should there be a power failure (and this would require a major event to keep the power out for more than a few minutes -- the building is classified as an EOC) when you are not on the first or second floor, you're totally screwed. All the stairs are behind gated doors. Without power from the security systems, the doors will not open. And very few people have an actual, physical key to open the doors.
I've worked here for 2+ years. I've been cursing the security system since the first day the power went out -- it doesn't take much to make that thing lose what little brain it has. (Plus, it was in a closet without a lock.)
Generally, I'd have to agree. But this is one of the cases to prove the point... the windows based solution is proving to be much less of a headache. (in fact, almost zero headache.
It's the same old expensive, complicated licensed, shit. Even today's "modern" versions don't stand up against linux v0.98 from the mid '90s.
I've thrown around the name "SCO" as a bad joke for ~15 years (or more.) It really is laughable to see McBride's comments. With 20 years exposure to the computing world, I've seen exactly 2, yes, 2, SCO systems in active deployment (and one in the test lab at NCR -- the backend for the cash registers.) One was a Lucent VoIP gateway and the other is an ACD manager for a Meridian phone switch. (the former never went anywhere and the later has been replaced with a windows box.)
ATI does have mobility drivers, however, they do not provide them to end users as the mobility chips are not made available to end users -- they are for system (read: laptop) manufacturers. You're supposed to bug your laptop maker for new ("less bug ridden") drivers.
In my experience, all of ATI's drivers are a pile of poo. If you ever find one that works for you, don't f**king change it. (And make thousands of copies of the installer so you don't lose it when ATI deletes it.)
Well, there's cheap, and then there's cheap... And then there's the people that do things like you :-)
If you have a business need for all those things, then you certainly have a business large enough to pay for device(s) to do all those things. The more VPNs you need, the more CPU will be required to handle them. The brain inside a Linksys is, well, very stupid. 10/8 doesn't count. Linksys routers have had syslog capability for years. And IDS is well beyond the scope of any "mini router" (and capability of a 386) and is certainly more complicated than it's worth to roll your own. Most of the "junk" selected to be a "mini router" would add new meaning to suck if used as a DNS or Web cache.
It's much faster and easier to hand over 60$ for a device that can do ~80% of the things I'd like to get done than it would be to spend weeks building and tuning a custom solution that no one else knows how to manage (and for which there is no one else to turn for assistance.) Yes, I used to be a "I'll do it myself" person, but there aren't enough hours in the year to do everything myself.
(The term is "line powered".)
ISDN is not line powered, but it's still attached to 911 service. That said, there's a difference between "life line" and "911" service. And if your house is on fire, the first thing you should do is leave; call for help from somewhere else.
Indeed. JFS, NUMA, and RCU did not exist in '85 when the original code was licensed. And it's fairly obvious there was no reciprocal agreement or Unixware/OpenServer would've had those features long ago.
Now, crazy licenses like that do exist. I have an NDA from USR hanging on the back of my front door (someday I'll frame it) that boils down to "we'll give you the source code as long as you hand over every change you make. For free." I laughed my ass off at that -- and obviously never signed it.
Perhaps, but they cannot march into your office and demand that you delete all the files, et. al. that were created from that software (read: derived works)
... No legal team is going to allow a licensing agreement where each party does not retain ownership of their respective code, or has provisions for retroactive dissolution of the agreement. IBM has always had good lawyers.
Even after the license has expired, SCO cannot demand IBM revoke every copy of AIX in the world simply because SCO doesn't own what IBM developed. They can demand their own code be returned or destroyed which would leave gaps in the AIX source, but there's nothing at all to stop IBM from continuing to support the existing AIX installed base. IBM may be prohibited from selling any more AIX, but the existing copies will continue to exist. (royalties and licensing fees were already paid for on them.)
There's a little legal thing called ex post facto
If you could find work elsewhere, you'd already be there. I know a number of people who don't have jobs and haven't had for a long time (>1year.)
No, it means a lot of people THINK SCO is gonna make a lot of money out of this. Anyone who understands the issues generally agrees SCO has no case. IBM very certainly knows they have no case. As such, IBM will not settle or buy SCO. This whole mess is years from reaching a court and more years from any actual outcome. SCO does not have the resources to fight a protracted, decade long court battle. IBM has resources; they have products and customers and market, unlike SCO.
Actually, I was just thinking this should be setting off warnings at the SEC. It may be that SCO is perpetrating a "fraud" to inflate their stock price. However, until the mythical 80 lines of code appear in court for all to see, there simply isn't a single molecule of usable evidence: exactly who copied what from whom? I'd like to know. I'd also like to see them back up their "enterprise" claims. (and the ancient egyptians were too technologically stupid to build the pyramids; it had to be aliens.)
There's a growing mound of questions for which there are no answers. SCO has been very crafty and manipulative. No one qualified to make any informed judgement will agree to the NDA. The only people who have signed the NDA and seen what SCO wants them to see are "technical" idiots -- they could've been shown the code equivalent of cave art drawn in red crayon and not known any better. They are certainly unaware of the volume and extent of UNIX(tm) code that has been used to teach people to program.
SCO's lawyers need to read the GPL again. (or the monkeys talking to the press need to listen to those lawyers.) If you modify code covered by the GPL, your changes fall under the GPL as well, esp. if you publish your changes.
(insert GPL gehad here)
ISDN is simply expensive (unless you live in TN.) It's not at all difficult. There are a few extra steps beyond "plug in line", but really, how hard is it to enter the switch type and SPIDs provided by the telco? (If you can get ISDN, you can get IDSL... where I work, we have IDSL lines in SC terminated on a DSLAM back in Raleigh, NC. As long as you are withing 50,000ft of an ISDN capable CO, the line can be terminated anywhere.)
Satellite is a pain in the ass for what you get out of it. You cannot install it yourself (legally) -- and most of the people paid to install it aren't very good at it.
I have ISDN. Yes, it's twice the cost of a cable modem and ~1.5x the cost of IDSL. However, I've had this line longer than either cable modems or DSL have existed. Aside from the CO crashing once, I've had no problems with it.
That's ingenious! Let someone "backhoe" that fiber. Muhuhaha.
2.5 is nowhere near ready to be called 2.6. There's just too much stuff that's been changed that hasn't made it to all corners of the kernel. And the kernel maintainers have started taking a Microsoft-like stance of publicly stating they'll remove some stuff that doesn't work ("how many people still need support for Adaptec 1542CF's anyway?")
I don't buy it. The plane is constructed with the same wiring as everything else. If this were the case, there would be a great deal of things on the ground screwed up by cell phones by the nature of the tiny wires/traces within them.
Signals inside waveguides are isolated from the junk outside the guide just as the junk outside is isolated from the signal inside. As such, one would have to place their cell phone inside the waveguide to mess anything up.
I'm sure politics is to blame for the lack of microwave ovens. A microwave oven (a broken/leaky one at least) produces several hundred thousand times the RF of a cell phone -- the result would be just as bad (or worse) for the humans as the electronics.
That space is referred to as "the swamp" (and for good reason.) That space should still be globally routable. My BGP tables have thousands of /24's in there (even from Sprint who publicly state /20 is as small as they go.) And SAVVIS is sending things as small as /29's -- their inbound filter will allow us to send them /32's!
ARIN has specific guidelines for returning address space and renumbering. Basically, they give you the space you can actually prove you need with some renumbering grace period afterwhich your original allocation is revoked.