I think this is a good point. I find people rabid adoration of GPL rather boring. Self interest isn't the worst thing in the world, esp, when as you say, the product you make creates even more wealth for the people who use it. You ought to have a little to send your kids to good schools.
Hey no one has really dug into this to understand why he's really saying this. Sun is working with the president of brazil (see his blog entry to him) to get opensource software (esp Solaris) as the foundation for the country.
So why is he bashing GPL? Pretty simple. With Sun's new CDDL license you DON'T have to opensource products which use that code. This means Brazil could make its own closed version of Solaris and sell it.
Its not clear that they would ever want to do this, but he is not saying this just to stir slashdot up, he is positioning his license for medium developed countries like Brazil. He is not talking third world here.
Yes I agree sometimes most of the work is doing the math on a smaller but complex file. Thanks for the specific info.
I just wonder how much further that reduces the market which is already limited. It seems like the highly graphicaly intensive oil and gas crowd and rendering crowd might not participate as easily in the remote grid shift.
What no one is mentioning is that these big cluster/grids that Sun/IBM are building to later sell over the network are dependant on the ratio between network speed and batch file sizes.
EXAMPLE: IBM is currently offering CPU/Hour service in Houston to oil and gas companies. Sounds great till you realize the multi-terabyte files that consume such a massive compute service are too big to be readily sent over the network. Instead they use vans to haul tape and disk over to IBM and then run the process on it.
What is the bandwith of a station wagon? Right now its faster than the internet on a 20 mile drive across Houston.
But even take it a step further and the ratio remains. What if I wanted to pay Sun/hr for CPUs while I worked on a big Maya render of 200 gigs. By the time I've sent that over cable modem have I gained a ton in performance time?
The problem I see is that we are making CPU massively parrallel but not networks. So will it EVER make sense to send a massive file to a commercial grid over a singular network connection.
Your right on here. Proximity style computing will have the "add a new block" abilities of grid with the low latency CPU/CPU communication of SMP. Also check out the new Niagra processor they are coming out with--the furtherance of the lots of cheap cores doing lots of threads, but all in one easy to work with box.
As Sun advances the threading and parrallism of cores/CPU and changes the game on bus latency and throughput with proximity it could get interesting.
Well because if a user doesn't tell Oracle that they are using Cent, given that the binary is exact to Red Hat Oracle will never know...thus they are a viable substitution
Thats why they don't want their name anywhere near Cent OS.
What your talking about is Red Hat Network which is a great value when your handling lots of servers.
I have tons of experience with Red Hat support. When you call in as a big time customer they are really pretty decent. When your not your on the road to India.
Just caused you've provisioned some big installations with RHN....
If what your saying is so true they wouldn't care about the Cent OS's of the world. I don't question your loyalty to them; I question the breadth of like minded admins.
You have to think bigger than just Red Hat's support. The reason companies pay so much is not just to get Red Hat's support, but BEA's or Oracles, or Websphere. Those guys don't technicaly support Cent OS.....
Its pretty deliberate. I don't think there is anything particularly evil about MS really--more that if you want to be a really dominant OS its the kind of behaviors you tend to have. Which says to me that RH will have to start doing everything it can to push alternatives to paying them out.
Brutal truth is that app certified OS markets always tip....god damnit.
That's excatly what I mean. Red Hat knows stuff like White Box or Cent are perfectly capable replacements. You could even call into say Oracle and their support guys would never know you had cent vs. Red Hat.
Red Hat's whole value-add and strategy is about using the business buerocracy to extract revenue. It sucks cause I used to feel they were all about the code.
Its actully pretty funny I think. When your product is as overpriced as Red Hat's you have to do everything you can to push out alternatives or the market will replace you.
Red Hat's value proposition isn't the quality of their support or they wouldn't care. Their value is the # of ISV's who write to their exact distribution. So the trick is simply to make sure no one else can nudge you with a hint that theirs is identical to RHs.
I really like the "Red Hed" from now on suggestion!
Its obvious that Oracle can do whatever they want in terms of their contracts and that users can eventually leave.
The problem is that it exerts an un-natural force on chip development. Sun is about to come out with an 32 core 4 cpu pizza box server next year. In terms of through put this may really be the way to go; but companies will be unable to select it because of the forces acting up stack.
It could lead to commercial applications acting as a disincentive to engineer the next break-through. If that happens the whole industry looses.
It's more about the earnings multiple. They only made 10 million dollars this last quarter, but are valued at over 3B.
The negativity has a lot to do with their earnings restatment. Whenever you have a sky-high earnings multiple you can't screw up or you'll make the market skittish.
If anything they've done a GREAT job of fighting back to something near their 90's bubble IPO value of 4B. Sadly however they are only worth that much because WallStreet understands the value of the lock-in they've created with ISV's on RHEL.
http://blogs.zdnet.com/index.php?p=364
If I was going to buy RH I'd also own Dell who is their biggest OEM. If you just want to own Red Hat prepare for a wild ride a high multiple stock brings with it.
Couldn't agree more. The old model was to give you a great distribution and work hard to win extra services revenue ect. The new model is "you want ISV's, you want RHEL, then pay up."
Its fine to have to pay for support on every version you install. Whats not fine is no longer offering the binaries for trials.
If I have to pay up I might as well use Solaris x86. I've been playing with d-trace a bunch and I'm thrilled with the results. Or if I'm going to just make vanilla secure web servers I'll use BSD. Red Hat aimed only for the high end with RHEL, so they are going to have to fight for it.
Red Hat isn't walking a tight rope. Clearly they have gone totally commercial in their interests. They are thrilled that all of the ISV's are certifying to RHEL.
http://blogs.zdnet.com/index.php?p=364
It's also not a question of if Red Hat is good for open-source. Clearly they are. They use their market capitalization to buy code (Sistina..) and move it into the OSS world. Their marketing is purely based on the goodness of the GPL and the OSS model. They champion it relentlessly.
The real question is one of distribution competition. Red Hat is cleary happy that its barriers against the other distributions are growing wildly. (Again pay attention to that zdnet link above)...well so what right?
The so what is the ultimate role we in the OSS world get to play. Are we going to be RHEL partner or pet? Will other distributions have a shot at the big-time, and at directing the Linux movement? If we are talking about using OSS applications it will remain a competition--but here is the catch. RHEL is grabbing all of the proprietary ISV certifications. They are not out to make money with the apache crowd anymore--they are hunting elephant with SAP, Oracle and the PeopleSofts of the world. They know if they can get all of the ISV's they will control most if not all of the real profits to be had in the Linux world, while their pets all scurry about helping them.
2 Final points.
1-Do the math on that quote about 4M per ISV. Red Hat claims 300 ISV's and 1k aps. If you multiply that cost to certify to RHEL out then ISVs have invested somewhere between 300M and 1.2B in their commitment to RHEL. Meaning it would take that much investment to move them all to Debian or some other competing distribution. Then consider SuSu was bought for only 200 million and you see how massive that barrier is. Scary.
2-They called Fedora the 'hobbyist' version. Give me a break! Red Hat spent the last 10 years convincing the world Linux was more than just an OS for hobbyists, and now they turn around and call their true OSS project just that. A RH sales-guy used the term several times in his pitch to me and I about lost it.
So yes they love OSS and the GPL, its what allows them to make a 90% margin. But do they love competing with other distributions and welcome an ongoing and balanced meritocracy? I can't believe that anymore.
Am I missing something here??? It doesn't look good.
Slashdot Mirror
I think this is a good point. I find people rabid adoration of GPL rather boring. Self interest isn't the worst thing in the world, esp, when as you say, the product you make creates even more wealth for the people who use it. You ought to have a little to send your kids to good schools.
Yeah people replace 5 year old Sun hardware with 400 boxes. DUH its been five years moores law anyone?
Hey no one has really dug into this to understand why he's really saying this. Sun is working with the president of brazil (see his blog entry to him) to get opensource software (esp Solaris) as the foundation for the country.
So why is he bashing GPL? Pretty simple. With Sun's new CDDL license you DON'T have to opensource products which use that code. This means Brazil could make its own closed version of Solaris and sell it.
Its not clear that they would ever want to do this, but he is not saying this just to stir slashdot up, he is positioning his license for medium developed countries like Brazil. He is not talking third world here.
Yes I agree sometimes most of the work is doing the math on a smaller but complex file. Thanks for the specific info.
I just wonder how much further that reduces the market which is already limited. It seems like the highly graphicaly intensive oil and gas crowd and rendering crowd might not participate as easily in the remote grid shift.
I'd actully love to be wrong!
What no one is mentioning is that these big cluster/grids that Sun/IBM are building to later sell over the network are dependant on the ratio between network speed and batch file sizes.
EXAMPLE: IBM is currently offering CPU/Hour service in Houston to oil and gas companies. Sounds great till you realize the multi-terabyte files that consume such a massive compute service are too big to be readily sent over the network. Instead they use vans to haul tape and disk over to IBM and then run the process on it.
What is the bandwith of a station wagon? Right now its faster than the internet on a 20 mile drive across Houston.
But even take it a step further and the ratio remains. What if I wanted to pay Sun/hr for CPUs while I worked on a big Maya render of 200 gigs. By the time I've sent that over cable modem have I gained a ton in performance time?
The problem I see is that we are making CPU massively parrallel but not networks. So will it EVER make sense to send a massive file to a commercial grid over a singular network connection.
Somone should do the math.
Your right on here. Proximity style computing will have the "add a new block" abilities of grid with the low latency CPU/CPU communication of SMP. Also check out the new Niagra processor they are coming out with--the furtherance of the lots of cheap cores doing lots of threads, but all in one easy to work with box.
As Sun advances the threading and parrallism of cores/CPU and changes the game on bus latency and throughput with proximity it could get interesting.
That's why I'd find it funny if they really went after Cent..I'm glad to hear they don't care.
FYI I'm a Solaris addict..for frame of reference and grain of salt.
Long live d-trace!
Well because if a user doesn't tell Oracle that they are using Cent, given that the binary is exact to Red Hat Oracle will never know...thus they are a viable substitution Thats why they don't want their name anywhere near Cent OS.
What your talking about is Red Hat Network which is a great value when your handling lots of servers.
I have tons of experience with Red Hat support. When you call in as a big time customer they are really pretty decent. When your not your on the road to India.
Just caused you've provisioned some big installations with RHN....
If what your saying is so true they wouldn't care about the Cent OS's of the world. I don't question your loyalty to them; I question the breadth of like minded admins.
You have to think bigger than just Red Hat's support. The reason companies pay so much is not just to get Red Hat's support, but BEA's or Oracles, or Websphere. Those guys don't technicaly support Cent OS.....
Its pretty deliberate. I don't think there is anything particularly evil about MS really--more that if you want to be a really dominant OS its the kind of behaviors you tend to have. Which says to me that RH will have to start doing everything it can to push alternatives to paying them out.
Brutal truth is that app certified OS markets always tip....god damnit.
That's excatly what I mean. Red Hat knows stuff like White Box or Cent are perfectly capable replacements. You could even call into say Oracle and their support guys would never know you had cent vs. Red Hat.
Red Hat's whole value-add and strategy is about using the business buerocracy to extract revenue. It sucks cause I used to feel they were all about the code.
Its actully pretty funny I think. When your product is as overpriced as Red Hat's you have to do everything you can to push out alternatives or the market will replace you.
Red Hat's value proposition isn't the quality of their support or they wouldn't care. Their value is the # of ISV's who write to their exact distribution. So the trick is simply to make sure no one else can nudge you with a hint that theirs is identical to RHs.
I really like the "Red Hed" from now on suggestion!
They charge per core. Its really pissing users off.
Its obvious that Oracle can do whatever they want in terms of their contracts and that users can eventually leave.
The problem is that it exerts an un-natural force on chip development. Sun is about to come out with an 32 core 4 cpu pizza box server next year. In terms of through put this may really be the way to go; but companies will be unable to select it because of the forces acting up stack.
It could lead to commercial applications acting as a disincentive to engineer the next break-through. If that happens the whole industry looses.
...yes I know. Late night forgot to move to text only...kinda cancelled any effect of it. I could barely read it.
It's more about the earnings multiple. They only made 10 million dollars this last quarter, but are valued at over 3B.
The negativity has a lot to do with their earnings restatment. Whenever you have a sky-high earnings multiple you can't screw up or you'll make the market skittish.
If anything they've done a GREAT job of fighting back to something near their 90's bubble IPO value of 4B. Sadly however they are only worth that much because WallStreet understands the value of the lock-in they've created with ISV's on RHEL.
http://blogs.zdnet.com/index.php?p=364
If I was going to buy RH I'd also own Dell who is their biggest OEM. If you just want to own Red Hat prepare for a wild ride a high multiple stock brings with it.
Couldn't agree more. The old model was to give you a great distribution and work hard to win extra services revenue ect. The new model is "you want ISV's, you want RHEL, then pay up."
Its fine to have to pay for support on every version you install. Whats not fine is no longer offering the binaries for trials.
If I have to pay up I might as well use Solaris x86. I've been playing with d-trace a bunch and I'm thrilled with the results. Or if I'm going to just make vanilla secure web servers I'll use BSD. Red Hat aimed only for the high end with RHEL, so they are going to have to fight for it.
Red Hat isn't walking a tight rope. Clearly they have gone totally commercial in their interests. They are thrilled that all of the ISV's are certifying to RHEL. http://blogs.zdnet.com/index.php?p=364 It's also not a question of if Red Hat is good for open-source. Clearly they are. They use their market capitalization to buy code (Sistina..) and move it into the OSS world. Their marketing is purely based on the goodness of the GPL and the OSS model. They champion it relentlessly. The real question is one of distribution competition. Red Hat is cleary happy that its barriers against the other distributions are growing wildly. (Again pay attention to that zdnet link above) ...well so what right?
The so what is the ultimate role we in the OSS world get to play. Are we going to be RHEL partner or pet? Will other distributions have a shot at the big-time, and at directing the Linux movement? If we are talking about using OSS applications it will remain a competition--but here is the catch. RHEL is grabbing all of the proprietary ISV certifications. They are not out to make money with the apache crowd anymore--they are hunting elephant with SAP, Oracle and the PeopleSofts of the world. They know if they can get all of the ISV's they will control most if not all of the real profits to be had in the Linux world, while their pets all scurry about helping them.
2 Final points.
1-Do the math on that quote about 4M per ISV. Red Hat claims 300 ISV's and 1k aps. If you multiply that cost to certify to RHEL out then ISVs have invested somewhere between 300M and 1.2B in their commitment to RHEL. Meaning it would take that much investment to move them all to Debian or some other competing distribution. Then consider SuSu was bought for only 200 million and you see how massive that barrier is. Scary.
2-They called Fedora the 'hobbyist' version. Give me a break! Red Hat spent the last 10 years convincing the world Linux was more than just an OS for hobbyists, and now they turn around and call their true OSS project just that. A RH sales-guy used the term several times in his pitch to me and I about lost it.
So yes they love OSS and the GPL, its what allows them to make a 90% margin. But do they love competing with other distributions and welcome an ongoing and balanced meritocracy? I can't believe that anymore.
Am I missing something here??? It doesn't look good.