Dell needed a 64bit platform if only for marketing purposes and with their once exclusive deal with Intel that ment Itanium because there was nothing else.
In reality it wasn't much of an option for them because Itanium has done best in HPC and high end SMP boxes. HPC is a market that Dell could address but they struggled against IBM and HP who were able to wrap Itanium and x86 offering with services and software support to impliment big HPC clusters. Large SMP is a market that Dell could not realistically address because there is no commodity interconnect for >4 module Itanium/EMT-64 leaving them with the option of cutting a deal with Unisys or one of the other >4 way Itanium vendors.
This left them with the 1-4 way commercial server market and x86-64's performance in both its AMD and Intel guises increasingly squeezed Itanium and both are AMD and Intel are now 64bit. Dells best hope to build relatively big commodity based servers now rests with Intel and AMD x86-64 products because both of them will offer 4 core modules for 4 socket servers much sooner than Itanium as they both offered 2 core modules for 4 socket servers earlier than the 2 core Itaniums.
Both IBM and Sun clearly thnink that there is a market for >4 module x86-64 boxes, IBM produce the X3 range and Sun produces the X4600. HP probably have a positioning problem with Itanium and Dell doesn't have >4 module interconnect.
The chip rumour mill has CloverTown (4 way 5160 MCM) running slower FSB than 1333 MHz with cache snoop traffic going over FSB between the 2 Woodcrest modules that make up the CloverTown MCM.
Sun has just released a faster US IV+ (1.8 Ghz) that will probably appear in the 490, 890 and above.
If the article is correct then there will be no replacement for the US IIIi
The T2 will replace the T1 but this will be a system upgrade rather than a module change. Sun tends to provide module upgrades for the mid range and high end SPARC units as well as for the AMD based boxes (generally).
There will be a whole new server platform in 2008 based on the Rock CPU called Supernova this will be a complete replacement for the US IV based platforms which were due to be replaced by US V. Sun normally builds a mid to high end server platform that supports two generations of US processors. The current servers supported US III/USIII+/US IV and now US IV+.
Assuming Supernova arrives in 2008 Sun will have a low end platform based on T2 possible with 1 and 2 module T2's and then Supernova based on Rock above that.
In fact you could buy a large X4600 for less money than a V440. A X4600 with 4 x Dual Core 885 Opterons and 32 GB of RAM costs $39K, a 4 way V440 with 32GB of RAM costs $40K. The X4600 gives you well over 2x the performance, its more expandible and it costs less.
This all depends on which SAPRC chip you are refering to. T1 for the right workloads is pretty much on it own and you generally need 4 cores to catch it. US IV+ is less of a stellar performer but then people buy very big Sun's for a lot of reasons only one of which might be single CPU throughput.
You are refering to a different family of processors.
The UltraSPARC III/IV/IV+ use the Gigaplane Interconnect which is designed to support >4 Modules, the UltraSPARC IIIi uses JBus which is designed to support 4 CPU's. The two are not compatible so you cannot drop a US IV+ into a IIIi slot.
You seem to have forgotten, PAM, XFN (ever editted nsswitch.conf) the slab memory allocator. You also seem to be labouring under the impression that I published an exhaustive list. I didn't.
Looks like someone has been doing a bit of Patent spring cleaning at IBM. What an eclectic mix, I particularly liked the Tamper Proof Screw patent lets hope it shows up sometime soon in the latest Linux kernel.
OO is in fact the largest single OpenSource project but lets ignore it completely because it just serves to illustrate how far behind IBM actually is.
Sun has donated the following Netbeans, JXTA, Chunks of GNOME, NFS V3/RPC, Glassfish, Grid Engine, Chunks of Tomcat, Chunks of Mozilla, Chuncks of Apache, Cubic Spline, XFN, PAM, ZFS, DTRACE to name a small part of their largess
The last time anyone counted Sun had donated more code to the OpenSource community than any other organisation other than UC Berkeley and that count was before OpenSolaris.
In addition ro actual code donations Sun has the distinction of being the key technology company responsible for creating open-standards with clearly defined specs which allowed FOSS projects to develop working FOSS alternatives to commercial products. The slab memoy allocator, XML, most of CORBA and a whole host of boring but crucial components that make modern OS's possible. Long before HP and IBM caught the OpenSource bug Sun was busy publishing SPEC's documenting how interfaces worked etc and doing the boring stuff that made Linux possible.
Your posting gives the impression that IBM is a huge donator to OpenSource projects and Sun with the exception of OpenSolaris which you are unsure about isn't.
Let me soothe your concerns, in fact Sun without OpenSolaris dwarfs IBM in terms of OpenSource contributions, as has been pointed out on a number of occasions more code in RedHat was donated by Sun than any other commercial company IBM and RedHat included. This excludes Sun's donations such as OpenOffice and it also excludes a huge amount of IP donated by Sun in the form of properly documented standards Patents and interfaces that most of the other commercial donators to OpenSource had to be dragged kicking and screaming to.
Woodcrest is for 2 socket only systems. Giving a maximum of 4 cores per system
Clovertown wil be a 4 core module(s) its actually 2 Woodcrest modules in a single package, where that single package presents 2 separate units to the Frontside Bus.
Intel has improved the bus performance of Woodcrest by supporting two Frontside Busses one for each Woodcrest module. This new chipset is also designed to support Clovertown but at a lower FSB clock speed. Core to Core cache snooping traffic is confined to a Woodcrest module and does not traverse the FSB unless you have 2 modules. With Clovertown this is true for each dual core Woodcrest module that makes up Clovertown but cache snooping between the two Woodcrest modules does traverse the FSB.
Unlike AMD Intel still uses external memory controllers which each CPU has to access via FSB so unlike Opteron Woodcrest and Clovertown do not have the benefit of very fast access to local memory and the total per core bandwidth for say an 8 core Clovertown will be lower than the agregate bandwidth available to an 8 core Opteron.
Try buying a similar set of components off the shelf from a major systems vendor such as HP or IBM.
IBM and HP want ~14-15K for a 2U 2x 2.6Ghz dual core Opteron 285 server fully loaded with memory and buying the 500 GB SATA II drives bare from them will cost you close to $700 per drive. The HP DL585 which is a similar size to Thumper though it cannot hold more than 4 drives comes in at $20K with 2 x dual core Opteron 285's 16 GB of RAM and 2 internal drives.
So for ~$47K you would have yourself a 2 CPU 2U server and 46 disks sitting on a table or for $52K you could have a 2 CPU 4U server with 44 drives sitting on the table.
Sun's prices would seem to be competitive when compared with other major systems vendors and that is probably what matters to the bulk of their customers.
Why would Sun want to Open Source Java, for that matter why on earth would you want them to ?
One key factor in Java's success has been the high degree of compatibility between Java implimentations on the range of platforms it runs on and the stability of the Java API/ABI's over a long time span. Compare and contrast this with Linux the only FOSS project that matches Java in size and complexity and you start to realise why Sun have been right not to give in to the OSS zealots and OpenSource Java.
As a Linux user (SuSe, Fedora and RedHat) I have discovered that the only guaranteed compatibility between different Linux distributions is at the source code level, RPM hell is a constant issue and even a minor kernel upgrade can have a knock on effect that ripples throughout the entire platform.
The idea that the same model which delivers this level of complexity should be applied to Java is bewildering and one can only conclude that anyone suggesting this wants Java to fail.
The reality is that Java is free as in it does not cost anything and the process by which new Java standards are defined is open enough to allow it to develop in a way that benefits Java users, Java developers and ISV's sadly this cannot be said for FOSS in its Linux incarnation.
Your point about Sun is wide of the mark. Sun sucessfully sued Microsoft for nearly over a billion dollars, they forced MS to open up some of their interfaces, hardly an alliance.
At the same time Sun is the commercial company who has supported the FOSS community longest and hardest. Sun's insistence on documented interfaces, agreed standards at a time when HP and IBM would have let you sleep with their sister rather than disclose an interface spec gave the FOSS community a standards framework in to build components against.
Sun's source and IP donations which dwarf all of the other commercial players have also provided real working technology that has filled in much of the gaps in the FOSS platform stack.
The technology war map shows a united front of Linux/BSD companies fighting MS as well as Mozilla, Apache etc what it doesn't point out is that RedHat, Apache, Mozilla etc are hugely dependant on Sun as a supplier of source/IP than any other commercial company. Without Sun's donations its highly unlikely that Linux for example would be in a possition to compete with MS at all.
Global warming could stop the North Atlantic Conveyor from running, this would mean that despite a mean increase in temperature worldwide that the UK, Northern Europe and the US North East would have much colder and much longer winters and a lower mean temperature while further south without cooling effect of the northerly movement of large quantities of warm water the mean temperature would rise dramatically. This would dramatically reduce the habitable land mass and also dramtically reduce the amount of land available for productive agriculture.
What ever the reason for increased temperatures, manmade or natural Bush would be smart to jump on the global warming bandwagon and use it as an excuse to wean the US of its dependence on a carbon based economy.
Smart because no one expects oil prices to decrease over the next decade, instead the expectation is that they will increase as demand continues to outstrip supply. The US's governments refusal to use global warming as a lever to get US industry to reduce its dependence on carbon bodes badly for the US economies competitiveness compared with more forward thinking economies with more aggressive targets for carbon reduction.
At the same time the US stands a good chance of missing out on the new markets associated with renewables, Sweden for example is the largest supplier of wind turbines worldwide partly because it has a large internal market. Germany is the largest supplier of biomass heating systems and again there is a large internal market.
The US seems to have opted to avoid short term pain in favour of long term economic decline as US industry becomes increasingly uncompetitive compared with more efficient users of carbon.
Slightly baffled as to why you would have put yourself through the pain of installing Linux on a T2000 when it comes with a perfectly good/better/much better (you choose) OS pre installed called Solaris.
DBMS don't require FPU performance since they don't issue floating point instructions. The app server market is also dominated by integer workloads, think Java and J2EE app servers as an example.
The T1 looks like an exceptionally effective Java/J2EE platform from the slew of great benchmark results Sun has published for the paltform. It is also no slouch as a DBMS platform as is SAP results show. It does lack single threaded performance so its going to be better as an OLTP platform for DBMS's than for high end reporting.
It also makes a good mailserver/messaging platform as its Notes performance demonstrates.
Correct, T2 is expected to be lower power or equivalent to the T1, part of this is because T2 will be built in a 65nm process as opposed to the 90nm process used to fabricate T1's.
The changes in T2 are 2 pipelines per core, up from 1. 8 threads per core, up from 4. FPU per core up from 1 per module. Faster memory subsystem, additional hw support for encryption and nework offload. On chip cache is expected to remain the same.
Sun's origional Motorola 68K based workstations had optional FPU's as did the first "desktop" SPARC workstation the 4/110. Sun workstations or servers equipped with a VME bus also had access to an optional Weitek FPU unit.
Even more exotic was the TAAC-1 which was a wide instruction word processor which could be used for FFT's, imaging etc.
One correction the TII (Niagara II) will be the first heavily multi-threaded SPARC CPU with one FPU per core, it is due out next year with rock being due out in 2008.
As an ex Sun employee I can assure you that Sun has always been well aware of the threat that IBM posed. Having being partly responsible for disposing of DEC Sun was always rubbing directly up against IBM after that.
I agree that Sun could have OpenSourced Solaris earlier, but there where practical difficulties in doing this such as who owned the rights to all the Solaris components etc which meant that it was always going to take a long time.
Sun clearly misjudged the OpenSource communities obsession with marketing over substance. Sun quite rightly regarded itself and still does as one of the mainstays of the OpenSource community, even before OpenSolaris Sun's donations to the OpenSource community dwarfed IBM and all the other players except FSF. What they did not expect was the extent to which the OpenSource community was prepared to swallow other vendors OpenSource marketing campaigns hook line and sinker. One of the best examples of this being the IBM and Sun patent donations.
IBM were lauded and Sun were flamed, this despite the fact that Sun donated patents which were not due to be renewed and which were related to OS's etc something you could not say for the IBM "donation".
Solaris 10 is available for Sun for free download you can use it for anything the only stipulation is that you have to pay for support.
This is actually more leniant than RedHat for example where you cannot use RedHat without a support contract and instead you have to rely on using Fedora.
If you do require full support as you would if you ran RedHat Enterprise Linux then you obviously pay Sun and Sun charges less per system than RedHat for what they argue is better support.
So if you are a large organisation deploying Solaris x86 you can use Solaris 10 for free on your development and test boxes and only pay Sun for support on the production environment, and there you pay less than you did for RedHat. If as in some environments half of your servers are dev/test systems then this can equate to a very large saving.
Slashdot Mirror
Dell needed a 64bit platform if only for marketing purposes and with their once exclusive deal with Intel that ment Itanium because there was nothing else.
In reality it wasn't much of an option for them because Itanium has done best in HPC and high end SMP boxes. HPC is a market that Dell could address but they struggled against IBM and HP who were able to wrap Itanium and x86 offering with services and software support to impliment big HPC clusters. Large SMP is a market that Dell could not realistically address because there is no commodity interconnect for >4 module Itanium/EMT-64 leaving them with the option of cutting a deal with Unisys or one of the other >4 way Itanium vendors.
This left them with the 1-4 way commercial server market and x86-64's performance in both its AMD and Intel guises increasingly squeezed Itanium and both are AMD and Intel are now 64bit. Dells best hope to build relatively big commodity based servers now rests with Intel and AMD x86-64 products because both of them will offer 4 core modules for 4 socket servers much sooner than Itanium as they both offered 2 core modules for 4 socket servers earlier than the 2 core Itaniums.
Both IBM and Sun clearly thnink that there is a market for >4 module x86-64 boxes, IBM produce the X3 range and Sun produces the X4600. HP probably have a positioning problem with Itanium and Dell doesn't have >4 module interconnect.
The chip rumour mill has CloverTown (4 way 5160 MCM) running slower FSB than 1333 MHz with cache snoop traffic going over FSB between the 2 Woodcrest modules that make up the CloverTown MCM.
Sun has just released a faster US IV+ (1.8 Ghz) that will probably appear in the 490, 890 and above.
If the article is correct then there will be no replacement for the US IIIi
The T2 will replace the T1 but this will be a system upgrade rather than a module change. Sun tends to provide module upgrades for the mid range and high end SPARC units as well as for the AMD based boxes (generally).
There will be a whole new server platform in 2008 based on the Rock CPU called Supernova this will be a complete replacement for the US IV based platforms which were due to be replaced by US V. Sun normally builds a mid to high end server platform that supports two generations of US processors. The current servers supported US III/USIII+/US IV and now US IV+.
Assuming Supernova arrives in 2008 Sun will have a low end platform based on T2 possible with 1 and 2 module T2's and then Supernova based on Rock above that.
In fact you could buy a large X4600 for less money than a V440. A X4600 with 4 x Dual Core 885 Opterons and 32 GB of RAM costs $39K, a 4 way V440 with 32GB of RAM costs $40K. The X4600 gives you well over 2x the performance, its more expandible and it costs less.
Not quite, the V210 and X2100, the X2200 and the T1000 do not have dual hot swapable PSU's.
They are all 1U boxes and most 1U boxes don't have dual hot pluggable PSU's.
This all depends on which SAPRC chip you are refering to. T1 for the right workloads is pretty much on it own and you generally need 4 cores to catch it. US IV+ is less of a stellar performer but then people buy very big Sun's for a lot of reasons only one of which might be single CPU throughput.
You are refering to a different family of processors.
The UltraSPARC III/IV/IV+ use the Gigaplane Interconnect which is designed to support >4 Modules, the UltraSPARC IIIi uses JBus which is designed to support 4 CPU's. The two are not compatible so you cannot drop a US IV+ into a IIIi slot.
You seem to have forgotten, PAM, XFN (ever editted nsswitch.conf) the slab memory allocator. You also seem to be labouring under the impression that I published an exhaustive list. I didn't.
Looks like someone has been doing a bit of Patent spring cleaning at IBM. What an eclectic mix, I particularly liked the Tamper Proof Screw patent lets hope it shows up sometime soon in the latest Linux kernel.
Not even close.
OO is in fact the largest single OpenSource project but lets ignore it completely because it just serves to illustrate how far behind IBM actually is.
Sun has donated the following Netbeans, JXTA, Chunks of GNOME, NFS V3/RPC, Glassfish, Grid Engine, Chunks of Tomcat, Chunks of Mozilla, Chuncks of Apache, Cubic Spline, XFN, PAM, ZFS, DTRACE to name a small part of their largess
The last time anyone counted Sun had donated more code to the OpenSource community than any other organisation other than UC Berkeley and that count was before OpenSolaris.
In addition ro actual code donations Sun has the distinction of being the key technology company responsible for creating open-standards with clearly defined specs which allowed FOSS projects to develop working FOSS alternatives to commercial products. The slab memoy allocator, XML, most of CORBA and a whole host of boring but crucial components that make modern OS's possible. Long before HP and IBM caught the OpenSource bug Sun was busy publishing SPEC's documenting how interfaces worked etc and doing the boring stuff that made Linux possible.
Your posting gives the impression that IBM is a huge donator to OpenSource projects and Sun with the exception of OpenSolaris which you are unsure about isn't.
Let me soothe your concerns, in fact Sun without OpenSolaris dwarfs IBM in terms of OpenSource contributions, as has been pointed out on a number of occasions more code in RedHat was donated by Sun than any other commercial company IBM and RedHat included. This excludes Sun's donations such as OpenOffice and it also excludes a huge amount of IP donated by Sun in the form of properly documented standards Patents and interfaces that most of the other commercial donators to OpenSource had to be dragged kicking and screaming to.
Woodcrest is for 2 socket only systems. Giving a maximum of 4 cores per system
Clovertown wil be a 4 core module(s) its actually 2 Woodcrest modules in a single package, where that single package presents 2 separate units to the Frontside Bus.
Intel has improved the bus performance of Woodcrest by supporting two Frontside Busses one for each Woodcrest module. This new chipset is also designed to support Clovertown but at a lower FSB clock speed. Core to Core cache snooping traffic is confined to a Woodcrest module and does not traverse the FSB unless you have 2 modules. With Clovertown this is true for each dual core Woodcrest module that makes up Clovertown but cache snooping between the two Woodcrest modules does traverse the FSB.
Unlike AMD Intel still uses external memory controllers which each CPU has to access via FSB so unlike Opteron Woodcrest and Clovertown do not have the benefit of very fast access to local memory and the total per core bandwidth for say an 8 core Clovertown will be lower than the agregate bandwidth available to an 8 core Opteron.
Try buying a similar set of components off the shelf from a major systems vendor such as HP or IBM.
IBM and HP want ~14-15K for a 2U 2x 2.6Ghz dual core Opteron 285 server fully loaded with memory and buying the 500 GB SATA II drives bare from them will cost you close to $700 per drive. The HP DL585 which is a similar size to Thumper though it cannot hold more than 4 drives comes in at $20K with 2 x dual core Opteron 285's 16 GB of RAM and 2 internal drives.
So for ~$47K you would have yourself a 2 CPU 2U server and 46 disks sitting on a table or for $52K you could have a 2 CPU 4U server with 44 drives sitting on the table.
Sun's prices would seem to be competitive when compared with other major systems vendors and that is probably what matters to the bulk of their customers.
Why would Sun want to Open Source Java, for that matter why on earth would you want them to ?
One key factor in Java's success has been the high degree of compatibility between Java implimentations on the range of platforms it runs on and the stability of the Java API/ABI's over a long time span. Compare and contrast this with Linux the only FOSS project that matches Java in size and complexity and you start to realise why Sun have been right not to give in to the OSS zealots and OpenSource Java.
As a Linux user (SuSe, Fedora and RedHat) I have discovered that the only guaranteed compatibility between different Linux distributions is at the source code level, RPM hell is a constant issue and even a minor kernel upgrade can have a knock on effect that ripples throughout the entire platform.
The idea that the same model which delivers this level of complexity should be applied to Java is bewildering and one can only conclude that anyone suggesting this wants Java to fail.
The reality is that Java is free as in it does not cost anything and the process by which new Java standards are defined is open enough to allow it to develop in a way that benefits Java users, Java developers and ISV's sadly this cannot be said for FOSS in its Linux incarnation.
Your point about Sun is wide of the mark. Sun sucessfully sued Microsoft for nearly over a billion dollars, they forced MS to open up some of their interfaces, hardly an alliance.
At the same time Sun is the commercial company who has supported the FOSS community longest and hardest. Sun's insistence on documented interfaces, agreed standards at a time when HP and IBM would have let you sleep with their sister rather than disclose an interface spec gave the FOSS community a standards framework in to build components against.
Sun's source and IP donations which dwarf all of the other commercial players have also provided real working technology that has filled in much of the gaps in the FOSS platform stack.
The technology war map shows a united front of Linux/BSD companies fighting MS as well as Mozilla, Apache etc what it doesn't point out is that RedHat, Apache, Mozilla etc are hugely dependant on Sun as a supplier of source/IP than any other commercial company. Without Sun's donations its highly unlikely that Linux for example would be in a possition to compete with MS at all.
Sadly you may be forced to deal with both.
Global warming could stop the North Atlantic Conveyor from running, this would mean that despite a mean increase in temperature worldwide that the UK, Northern Europe and the US North East would have much colder and much longer winters and a lower mean temperature while further south without cooling effect of the northerly movement of large quantities of warm water the mean temperature would rise dramatically. This would dramatically reduce the habitable land mass and also dramtically reduce the amount of land available for productive agriculture.
What ever the reason for increased temperatures, manmade or natural Bush would be smart to jump on the global warming bandwagon and use it as an excuse to wean the US of its dependence on a carbon based economy.
Smart because no one expects oil prices to decrease over the next decade, instead the expectation is that they will increase as demand continues to outstrip supply. The US's governments refusal to use global warming as a lever to get US industry to reduce its dependence on carbon bodes badly for the US economies competitiveness compared with more forward thinking economies with more aggressive targets for carbon reduction.
At the same time the US stands a good chance of missing out on the new markets associated with renewables, Sweden for example is the largest supplier of wind turbines worldwide partly because it has a large internal market. Germany is the largest supplier of biomass heating systems and again there is a large internal market.
The US seems to have opted to avoid short term pain in favour of long term economic decline as US industry becomes increasingly uncompetitive compared with more efficient users of carbon.
Slightly baffled as to why you would have put yourself through the pain of installing Linux on a T2000 when it comes with a perfectly good/better/much better (you choose) OS pre installed called Solaris.
Not quite sure how that got modded informative.
DBMS don't require FPU performance since they don't issue floating point instructions. The app server market is also dominated by integer workloads, think Java and J2EE app servers as an example.
The T1 looks like an exceptionally effective Java/J2EE platform from the slew of great benchmark results Sun has published for the paltform. It is also no slouch as a DBMS platform as is SAP results show. It does lack single threaded performance so its going to be better as an OLTP platform for DBMS's than for high end reporting.
It also makes a good mailserver/messaging platform as its Notes performance demonstrates.
Correct, T2 is expected to be lower power or equivalent to the T1, part of this is because T2 will be built in a 65nm process as opposed to the 90nm process used to fabricate T1's.
The changes in T2 are 2 pipelines per core, up from 1. 8 threads per core, up from 4. FPU per core up from 1 per module. Faster memory subsystem, additional hw support for encryption and nework offload. On chip cache is expected to remain the same.
The weitek unit was an FPU. Like the IBM 8087 you had to compile applications specifically to use the weitek unit.
Sun's origional Motorola 68K based workstations had optional FPU's as did the first "desktop" SPARC workstation the 4/110. Sun workstations or servers equipped with a VME bus also had access to an optional Weitek FPU unit.
Even more exotic was the TAAC-1 which was a wide instruction word processor which could be used for FFT's, imaging etc.
One correction the TII (Niagara II) will be the first heavily multi-threaded SPARC CPU with one FPU per core, it is due out next year with rock being due out in 2008.
Sun actually made a trading loss of $39 million dollars. The rest is exceptional charges.
Not brilliant but not dreadfull either.
As an ex Sun employee I can assure you that Sun has always been well aware of the threat that IBM posed. Having being partly responsible for disposing of DEC Sun was always rubbing directly up against IBM after that.
I agree that Sun could have OpenSourced Solaris earlier, but there where practical difficulties in doing this such as who owned the rights to all the Solaris components etc which meant that it was always going to take a long time.
Sun clearly misjudged the OpenSource communities obsession with marketing over substance.
Sun quite rightly regarded itself and still does as one of the mainstays of the OpenSource community, even before OpenSolaris Sun's donations to the OpenSource community dwarfed IBM and all the other players except FSF. What they did not expect was the extent to which the OpenSource community was prepared to swallow other vendors OpenSource marketing campaigns hook line and sinker. One of the best examples of this being the IBM and Sun patent donations.
IBM were lauded and Sun were flamed, this despite the fact that Sun donated patents which were not due to be renewed and which were related to OS's etc something you could not say for the IBM "donation".
Solaris 10 is available for Sun for free download you can use it for anything the only stipulation is that you have to pay for support.
This is actually more leniant than RedHat for example where you cannot use RedHat without a support contract and instead you have to rely on using Fedora.
If you do require full support as you would if you ran RedHat Enterprise Linux then you obviously pay Sun and Sun charges less per system than RedHat for what they argue is better support.
So if you are a large organisation deploying Solaris x86 you can use Solaris 10 for free on your development and test boxes and only pay Sun for support on the production environment, and there you pay less than you did for RedHat. If as in some environments half of your servers are dev/test systems then this can equate to a very large saving.