If the pricing per seat per year is still $100, I don't see a whole trove of enterprises converting their entire camp over to this product.
That's probably because you don't know what you're talking about. The Red Hat equivalent (actually slightly less than
equal) is RH Enterprise Linux WS, which will run you $179-$299. An "enterprise" is going to buy their software and a support contract. It isn't going to be running on whatever random collection of Sourceforge goodies the BOFH decides to pull over today.
Crippled in what way? You can download the whole 5-disk
beta and do a full install if you want. If you just want to
try it out, the 1-disk LiveCD lets you do that without having to go to the effort of installing. Sounds like everyone wins,
so what are you bitching about?
Oh, BS. It's enormously unusual for a tech company to last 21 years, let alone have half of the founders still involved.
This is probably over a major senior management disagreement...What other subject would they have time to talk about at Sun HQ?
And more BS. Joy was a founder, but he's not involved in the day-to-day "management" of the company. Regardless of what Sun management talks about at "HQ", Joy has been in his Aspen lab churning out cool stuff like Java for, what, 10 years now? I don't have a clue why he's leaving, but I can't imagine it's because he's pissed that McNealy hasn't drunk the Linux koolaid.
The author claims that Joy is such a visionary that Sun's entire R&D strategy will have to change. Then in the next line he says that Joy is such a luddite that he'll never get involved with another tech company. Which is it? Visionary or luddite?
This is so plainly just somebody looking for a way to bitch about Sun that I can't believe anybody bothered to mod it up.
Just for kicks, I looked at The Wall (which another poster used as an example above). He said that it cost $36 at Best Buy, which was on par with Cheap CDs. Your online choice was cheaper at $31. Amazon was the cheapest at $28, and shipping was free.
Re:Typical Sun Quote
on
LWCE Wrapup
·
· Score: 2, Interesting
By referring to this as a "typical Sun opinion", it's pretty clear that you came to this with an pre-existing bias. I think you're reading more into this than is really there, and you're the one who is missing the point.
I think the point he is trying to make is that in the end, software is all just bits on a disk. There is nothing about the development process that makes open source software magically better. Good software always takes hard work, thorough testing, and talented developers.
It's pretty clear (to me at least) that he is really talking about the typical customer or end user - not a developer. For somebody who is looking to roll out 1000 "Mad Hatter" desktops to secretaries and/or phone support people, the "free beer" aspect is more important than the "free speech" aspect. This isn't an environment where they are going to be upgrading the OS, applications, or windowing system with the latest tinderbox bits every night. They want something that works well and doesn't cost too much. They don't care whether it was developed by a corporate engineering team, an open source project, or an infinite number of monkeys.
If you look at the whitepapers put out by Intel,
their systems experts and compiler experts were
unable to come up with any single benchmark where they got more than 20-30% speedup.
Hyperthreading doesn't give you two CPUs. It lets
two threads share the resources of a single CPU.
If those threads are trying to use the same
resources, you're going to have exactly the same
limits as a single-threaded CPU (actually a bit
worse). Hyperthreading is only a win if the two
threads are using different types of resources.
The only time you're going to see anything even approaching 2x is if you have one application (or thread) that hits memory on almost every instruction running on the same CPU as an application that runs entirely out of cache.
Is HP-UX safe, or are they next? Sun and SCO have both been pretty clear that Solaris is unaffected by all of this. I don't recall seeing any similar assurances about HP-UX.
Sun paid Novell ~$100M to get out of paying royalties and for the right to distribute derivative works in perpetuity.
IBM, HP, and everybody else paid Novell in the neighborhood of $10M just to get out of paying royalties.
So, Sun is in the clear, and has been since long before this lawsuit was a gleam in the eye of the SCO executive team. What I'd like to know is,
even if SCO does win the lawsuit, can Sun continue
to distribute Linux?
OK, let's assume for a second you actually can build a rackmount server with exactly the same specs for less. Can you build 200 of them?
Sun isn't getting into this market to sell individual systems for people to run in their bedroom / offices. They are in the market to sell these in multi-rack installations to run the web servers that are attached to the clusters of Sun Fire 4800 app servers, which are attached to the failover-capable pair of 15Ks running Oracle on the back end.
Now, lets see you whip up a 72-way 576GB machine with over 100 GB/s of local memory bandwidth.
So, you have 5000 songs on your iPod but you don't plan to pay for any music you put on it. Tell me again how the record companies are screwing over the artists?
Seriously, people have been announcing plans to replace X with something lighter weight for roughly 20 years now. Every time one of the projects gets far enough along to slap together a web site and a half-assed demo, you guys fall all over yourselves to promote it.
This may finally be the project that gets it right, and 10 years from now it will deliver something that is generally useful. Until then, it probably doesn't need to be on the front page of/..
The Sun Ray idea is very cool even without the whole "hotelling office" setup.
I can prep a presentation in my office, and then walk over to a conference room with my card and pop it up on the screen there instantly.
I can have a debug session running in my office. If something goes flakey with the hardware, I can bring the whole session into the lab without stopping and re-establishing everything.
If I run into a problem with a piece of code, I
can grab my card, walk over to the original author's office, and show it to him on his Sun Ray - without him having to do so much as open a new window.
I can move seamlessly back and forth between my office and the "Internet Cafe" in the next building. I can start writing an email over lunch, and finish it when I'm back at my desk.
You get the portability of a laptop (within the campus at least), but it fits in your shirt pocket.
The complaint is that Sun is raising its prices without adding any new value.
Which is complete nonsense. "Orion" isn't just Solaris. It's Solaris with an added directory server, portal server, identity server, web server , app server, calendar server, cluster management , and god knows what else.
Sheesh, I can't believe the stuff that gets modded up sometimes.
>Sun sells x86 hardware at an inflated price. I >could go with Dell or build my own cheapo PC box, >then install Solaris 9 or Linux on the cheap. Sun >makes nothing.
Which reinforces my point. Sun sells x86 hardware. Maybe you're not buying it, but somebody is. If somebody is so eager to buy Sun gear that they are willing to pay a premium, doesn't it seem reasonable to think they might want Sun's OS as well?
>> People who buy $1M database servers also need >> cheap web servers to run the front end. If they >> can run the same OS on both sets of machines, >> it makes the whole package more appealing.
> So what? Sun does not make any money from that.
If having a cheap Solaris front end available makes it more likely for a customer to buy an expensive Solaris back end, then of course they make money on it.
> You are basically condeming the low-end SPARC > machines e.g. netra X1 to a slow death.
The Netra machines are designed explicitly for the telco market. The availability of an x86 Solaris has nothing to do with them. Sun does have some general-purpose low-end SPARC machines that might be affected by this release. I dunno what that impact might be. I'm sure that occurred to Sun's marketing folks.
> And linux is close enough to solaris on the > command line interface.
That's an awfully narrow view of the world. They are certainly not close enough at the API level.
>> People who buy big SPARC machines need software >> to run on them. Solaris x86 makes a damn cheap >> way for ISVs to develop software. As long as you >> don't do anything stupid, it's a simple recompile >> from Solaris x86 to Solaris SPARC.
> Big deal I could cross compile from a Linux machine anyway.
Not if you wanted to make use of anything other than bog standard Unix system calls. Without the user environment available, you can't make use of anything in the Solaris libraries. At the very least you would need the API stubs to compile against, and then you couldn't do any testing whatsoever.
> My point is that sun makes no money from me if > they don't sell me hardware.
That's correct. And irrelevant. Your one person. That's not where they're going to make their money.
> Sun should want me develop code on sparc > machines only not intel because their core > business is hardware!
Exactly. But with Solaris x86, you can do all of your development and testing on the cheap machines you referred to above. Then "porting" to SPARC really is as simple as a recompile. So, Sun has just given you a cheap way to do your development _and_ make your software available on SPARC, which is the "volume" Solaris platform. More software availability means more hardware sales.
Everybody wins. You get a cheap development platform, SPARC users get more software, Sun sells more hardware. It's a virtuous circle, which doesn't rely on Sun getting any income whatsoever directly from x86 Solaris.
I don't get why sun is releasing solaris 9 to the intel platform. I thought they were supposed to be a hardware company?
Yeah, and?
Sun sells x86 hardware. Now it runs Solaris as well as Linux.
People who buy $1M database servers also need cheap web servers to run the front end. If they can run the same OS on both sets of machines, it makes the whole package more appealing.
People who buy big SPARC machines need software to run on them. Solaris x86 makes a damn cheap way for ISVs to develop software. As long as you don't do anything stupid, it's a simple recompile from Solaris x86 to Solaris SPARC.
Also, the inside scoop was that they already _had_ Solaris 9 for intel, but the higher-ups didn't want to release it.
I don't think this is any sort of news. It certainly wasn't some secret skunkworks project.
The x86 builds never stopped. Projects that putback into Solaris always had to ensure that they didn't break the x86 build. It was all the other stuff that goes into releasing software that was halted: support, marketing, documentation, huge QA resources, etc.
For commercial products, writing the code is the easy part.
Thyme and sage are not tricky. They grow easily from seed and and are incredibly tough. I have both growing in direct sun in fairly dry soil, and they are thriving. I live on the edge of Zones 4 & 5, and I've never had a problem with these two lasting the winter. I've tried bringing thyme inside each winter, and it always dies by mid-January. It's much easier to scrape the snow off the outdoor plants and harvest what I need.
I cut each of them almost to the ground after each winter, and they spring right back up. (They should probably be cut back in the fall, but then what would I do for sage when roasting turkey in the winter?)
For starters, Linux performance on the 390 is _terrible_. The 390 can deliver awesome I/O performance - but not to processes running under Linux on top of VM. I saw some specJBB numbers on 390/Linux floating around, and they were less than 10% of what you would expect given the hardware.
Linux on the 390 is a really cool story from a geek perspective. Try to find somebody actually running it in a production environment.
Slashdot Mirror
Good god man! You've created a Content Black Hole! You're dragging down the Scores of every post in the vicinity!
That's probably because you don't know what you're talking about. The Red Hat equivalent (actually slightly less than equal) is RH Enterprise Linux WS, which will run you $179-$299. An "enterprise" is going to buy their software and a support contract. It isn't going to be running on whatever random collection of Sourceforge goodies the BOFH decides to pull over today.
Oh, BS. It's enormously unusual for a tech company to last 21 years, let alone have half of the founders still involved.
This is probably over a major senior management disagreement...What other subject would they have time to talk about at Sun HQ?
And more BS. Joy was a founder, but he's not involved in the day-to-day "management" of the company. Regardless of what Sun management talks about at "HQ", Joy has been in his Aspen lab churning out cool stuff like Java for, what, 10 years now? I don't have a clue why he's leaving, but I can't imagine it's because he's pissed that McNealy hasn't drunk the Linux koolaid.
"Score: 4, Interesting" my ass.
The author claims that Joy is such a visionary that Sun's entire R&D strategy will have to change. Then in the next line he says that Joy is such a luddite that he'll never get involved with another tech company. Which is it? Visionary or luddite?
This is so plainly just somebody looking for a way to bitch about Sun that I can't believe anybody bothered to mod it up.
Just for kicks, I looked at The Wall (which another poster used as an example above). He said that it cost $36 at Best Buy, which was on par with Cheap CDs. Your online choice was cheaper at $31. Amazon was the cheapest at $28, and shipping was free.
I think the point he is trying to make is that in the end, software is all just bits on a disk. There is nothing about the development process that makes open source software magically better. Good software always takes hard work, thorough testing, and talented developers.
It's pretty clear (to me at least) that he is really talking about the typical customer or end user - not a developer. For somebody who is looking to roll out 1000 "Mad Hatter" desktops to secretaries and/or phone support people, the "free beer" aspect is more important than the "free speech" aspect. This isn't an environment where they are going to be upgrading the OS, applications, or windowing system with the latest tinderbox bits every night. They want something that works well and doesn't cost too much. They don't care whether it was developed by a corporate engineering team, an open source project, or an infinite number of monkeys.
One small correction: there is a Cisco VPN client for Linux. I'm using it right now.
:).
There is no client for BSD, which is the main reason I'm still using Linux. (at least until the FedEx guy shows up with my shiny new Powerbook
No no no. Everybody hates Solaris and Sun is going to be dead in 6 months.
People have been saying that for 10 years now, so it must be true.
If you look at the whitepapers put out by Intel, their systems experts and compiler experts were unable to come up with any single benchmark where they got more than 20-30% speedup.
Hyperthreading doesn't give you two CPUs. It lets two threads share the resources of a single CPU. If those threads are trying to use the same resources, you're going to have exactly the same limits as a single-threaded CPU (actually a bit worse). Hyperthreading is only a win if the two threads are using different types of resources.
The only time you're going to see anything even approaching 2x is if you have one application (or thread) that hits memory on almost every instruction running on the same CPU as an application that runs entirely out of cache.
Is HP-UX safe, or are they next? Sun and SCO have both been pretty clear that Solaris is unaffected by all of this. I don't recall seeing any similar assurances about HP-UX.
Sun paid Novell ~$100M to get out of paying royalties and for the right to distribute derivative works in perpetuity.
IBM, HP, and everybody else paid Novell in the neighborhood of $10M just to get out of paying royalties.
So, Sun is in the clear, and has been since long before this lawsuit was a gleam in the eye of the SCO executive team. What I'd like to know is, even if SCO does win the lawsuit, can Sun continue to distribute Linux?
OK, let's assume for a second you actually can build a rackmount server with exactly the same specs for less. Can you build 200 of them?
Sun isn't getting into this market to sell individual systems for people to run in their bedroom / offices. They are in the market to sell these in multi-rack installations to run the web servers that are attached to the clusters of Sun Fire 4800 app servers, which are attached to the failover-capable pair of 15Ks running Oracle on the back end.
Now, lets see you whip up a 72-way 576GB machine with over 100 GB/s of local memory bandwidth.
So, you have 5000 songs on your iPod but you don't plan to pay for any music you put on it. Tell me again how the record companies are screwing over the artists?
Yet Another X Alternative? Must be Monday.
/..
Seriously, people have been announcing plans to replace X with something lighter weight for roughly 20 years now. Every time one of the projects gets far enough along to slap together a web site and a half-assed demo, you guys fall all over yourselves to promote it.
This may finally be the project that gets it right, and 10 years from now it will deliver something that is generally useful. Until then, it probably doesn't need to be on the front page of
Mad Hatter is a part of it.
And _that_ was modded up?
I can prep a presentation in my office, and then walk over to a conference room with my card and pop it up on the screen there instantly.
I can have a debug session running in my office. If something goes flakey with the hardware, I can bring the whole session into the lab without stopping and re-establishing everything.
If I run into a problem with a piece of code, I can grab my card, walk over to the original author's office, and show it to him on his Sun Ray - without him having to do so much as open a new window.
I can move seamlessly back and forth between my office and the "Internet Cafe" in the next building. I can start writing an email over lunch, and finish it when I'm back at my desk.
You get the portability of a laptop (within the campus at least), but it fits in your shirt pocket.
Which is complete nonsense. "Orion" isn't just Solaris. It's Solaris with an added directory server, portal server, identity server, web server , app server, calendar server, cluster management , and god knows what else.
Sheesh, I can't believe the stuff that gets modded up sometimes.
>Sun sells x86 hardware at an inflated price. I
>could go with Dell or build my own cheapo PC box,
>then install Solaris 9 or Linux on the cheap. Sun
>makes nothing.
Which reinforces my point. Sun sells x86 hardware. Maybe you're not buying it, but somebody is. If somebody is so eager to buy Sun gear that they are willing to pay a premium, doesn't it seem reasonable to think they might want Sun's OS as well?
>> People who buy $1M database servers also need
>> cheap web servers to run the front end. If they
>> can run the same OS on both sets of machines,
>> it makes the whole package more appealing.
> So what? Sun does not make any money from that.
If having a cheap Solaris front end available makes it more likely for a customer to buy an expensive Solaris back end, then of course they make money on it.
> You are basically condeming the low-end SPARC
> machines e.g. netra X1 to a slow death.
The Netra machines are designed explicitly for the telco market. The availability of an x86 Solaris has nothing to do with them. Sun does have some general-purpose low-end SPARC machines that might be affected by this release. I dunno what that impact might be. I'm sure that occurred to Sun's marketing folks.
> And linux is close enough to solaris on the
> command line interface.
That's an awfully narrow view of the world. They are certainly not close enough at the API level.
>> People who buy big SPARC machines need software
>> to run on them. Solaris x86 makes a damn cheap
>> way for ISVs to develop software. As long as you
>> don't do anything stupid, it's a simple recompile
>> from Solaris x86 to Solaris SPARC.
> Big deal I could cross compile from a Linux machine anyway.
Not if you wanted to make use of anything other than bog standard Unix system calls. Without the user environment available, you can't make use of anything in the Solaris libraries. At the very least you would need the API stubs to compile against, and then you couldn't do any testing whatsoever.
> My point is that sun makes no money from me if
> they don't sell me hardware.
That's correct. And irrelevant. Your one person. That's not where they're going to make their money.
> Sun should want me develop code on sparc
> machines only not intel because their core
> business is hardware!
Exactly. But with Solaris x86, you can do all of your development and testing on the cheap machines you referred to above. Then "porting" to SPARC really is as simple as a recompile. So, Sun has just given you a cheap way to do your development _and_ make your software available on SPARC, which is the "volume" Solaris platform. More software availability means more hardware sales.
Everybody wins. You get a cheap development platform, SPARC users get more software, Sun sells more hardware. It's a virtuous circle, which doesn't rely on Sun getting any income whatsoever directly from x86 Solaris.
Yeah, and?
Sun sells x86 hardware. Now it runs Solaris as well as Linux.
People who buy $1M database servers also need cheap web servers to run the front end. If they can run the same OS on both sets of machines, it makes the whole package more appealing.
People who buy big SPARC machines need software to run on them. Solaris x86 makes a damn cheap way for ISVs to develop software. As long as you don't do anything stupid, it's a simple recompile from Solaris x86 to Solaris SPARC.
I don't think this is any sort of news. It certainly wasn't some secret skunkworks project. The x86 builds never stopped. Projects that putback into Solaris always had to ensure that they didn't break the x86 build. It was all the other stuff that goes into releasing software that was halted: support, marketing, documentation, huge QA resources, etc.
For commercial products, writing the code is the easy part.
If Sun had used KDE, the desktop would tie back to a German group.
Even for Slashdot, that's an amazingly dumb statement. And demonstrably wrong: Staroffice is almost entirely developed in Germany.
If the web site is now busier than ever, that's a pretty ineffective chilling effect.
Thyme and sage are not tricky. They grow easily from seed and and are incredibly tough. I have both growing in direct sun in fairly dry soil, and they are thriving. I live on the edge of Zones 4 & 5, and I've never had a problem with these two lasting the winter. I've tried bringing thyme inside each winter, and it always dies by mid-January. It's much easier to scrape the snow off the outdoor plants and harvest what I need.
I cut each of them almost to the ground after each winter, and they spring right back up. (They should probably be cut back in the fall, but then what would I do for sage when roasting turkey in the winter?)
For starters, Linux performance on the 390 is _terrible_. The 390 can deliver awesome I/O performance - but not to processes running under Linux on top of VM. I saw some specJBB numbers on 390/Linux floating around, and they were less than 10% of what you would expect given the hardware.
Linux on the 390 is a really cool story from a geek perspective. Try to find somebody actually running it in a production environment.