You have good points, especially with the Mainframe. However, the DEC/VMS claim only stands on it's own for binary compatibility. You really cannot throw an OS/Platform into this discussion when that OS/Platform is on its way out the door.
First, who in their right mind would purchase a multi million dollar machine (in your example the E15K) and not know if it will do the job? Not only do you find out if it does the job, you make sure that it will do the job for a long long time. That said, Sun has something called iForce Centers. These are big labs with tons of Sun products and tons of Sun experts where customers spending millions (which is the type of custom refered to in this thread, but certainly others can go to iForce Centers as well) can come and put their application to the test. You can setup any environment with any amount of the hardware and work with the Sun experts to make sure your application runs like a champ. They will throw everything at your app and use everything you want to make your app work. Then at the end of the process they give you an itemized list of hardware, software and man hours, not for the test, but for assembling it for yourself. Of course it is very likely at this point you will give it back to them with a big fat check. But you do not have to.
So you see, it will do the job on the E15K and you will know this before you spend any money.
As for the Altix. Would I spend millions on hardware from a company that is all but dead? No. Do I know *anybody* who is using an Itanium based setup (Linux or otherwise)? No. Do I think Itanium is going nowhere? Yes. Has Itanium based SGI Altix systems been around long enough for me to spend millions on one? No.
We are talking about the enterprise here. Not the big closet the end of the hall, next to your little sister's room. With enterpise computing you take no chances; you make very scientific decisions. Unfortunately your answer is no better than those I blasted who compare the E15K to a whole bunch of Dells in a beowulf setup; not because the Altix can or cannot match the E15K (it might be able to) but because you have no fucking experience in what you are talking about.
Glad to be aboard. I have long wondered what everybody else is on. You should have seen my jaw drop when I went to a local Sun event (mind you I had already been using Sun equipment and had been very happy with it) when I learned just how long they support their OS's. I also learned that if you certify you app on Solaris and a newer version of Solaris is released and ends up breaking your app, they will either fix Solaris or they will pay you to fix your app on Solaris. Absolutely Incredible!!! Then the guy telling us all this said he has a small app that he wrote 12 years ago, lost the source code 5 years ago and that he runs on Solaris every day.
No major vendor even comes close to bringing this level of value, longevity and investment insurance.
My favorite Slashdot posts run along the line of "Why would my company purchase a 106 processor E15K when we can just create a Linux Cluster out of Dell PC's?" This lack of understanding of the fundamental difference between a horizontal cluster utilizing parallel programming and a single server image scaled to 106 processors and half a terabyte of RAM, just baffles me. Unfortunately this kind of logic is far too common here. Don't get me wrong, there are people on Slashdot with tons of knowledge and lots of experience. Unfortunately it *seems* that 80% of the posters refer to servers that are likely sitting in their bedroom next to their nightstand.
People who question Sun typically do not know shit about Sun and make comparisons analogous to comparing a Ford and a BMW. Unfortunately Sun is largly to blame for this in their lack of marketing some of the incredible facts I have mentioned in my posts. But those who know, know.
Those companies willing to spend a few million do not care as much about speed as they do about their application being able to run on future versions of a vendor's processor with no recompile (and garanteed). They want to make sure that the OS they are buying with their million dollar setup is supported by the vendor for at least 10 years. They want to make sure that when the next version of the processor the vendor designs comes out that they can put it in their existing box; replacing the present processors or along side of them (without having to bring the box down). If raw throughput was Sun's only goal they could make Sparcs as fast as anybody else. But binary compatibility, open architecture, mix and match and endless support cycles for the OS is what makes million dollar companies say no to raw speed and yes to Sun. Oh and Sun's machines have incredible throughput and perform very well in real world scenarios.
"There is no surprise either that heavy weights such as IBM, Dell, SUN and even HP -- who pretty much designed Itanium -- put some of their eggs in their AMD busket."
Good post, except for the above quote. IBM has one (eServer 325) machine that is based on the Opteron. This machine is being marketed to high performance computing environments and not really as a general purpose machine. Not sure if it was The Register or The inquirer, but one of the sites had quotes from a top IBM executive on their lack of plans for more Opteron based models (i.e. a 4 way and some other form factors).
Dell has outright said they are not supporting the Opteron and were holding out to see what happens in the market (which could mean they were holding out for Intel).
HP's Opteron systems are a rumor and have yet to be substantiated.
That leaves Sun. Sun is the only major vendor with both an available Opteron system and another (4 way, 6 hdd) on the way. They are also the only major vendor that has made long term commitments to Opteron on a variety of setups/form factors, including workstations and 8 way servers. Sun has also promised to deliver 32/64 bit Linux (third party) and 32/64 bit Solaris on Opteron.
In fact when Sun announced all of there support, that should have tipped us all off that the Intel annoucement was coming. When would Sun do what everybody else is doing, unless everybody else was actually gonna be doing something else? Don't be surprised two years from now when Sun is still the only major vendor of Opteron systems.
Man you are way off regarding automatic update and Linux. I use apt-rpm and can set it up to do many kinds of automatic updating. I can deploy updates from one machine to the entires company either automagically or manually. I *never* have to visit a *NIX box. In fact if you do *have* to go visit a *NIX box for anything but a complete hardware failure then you are a shitty admin. Windows boxes are the only boxes that I ever had to go visit. What you laud MS for being able to do I have been able to do in *NIX for years, with or without deployment tools such as apt. *NIX environments are fully automatable out of the box and have always been that way. Don't come here telling us about some new Microsoft automation feature and expect any *NIX admin worth a shit to even be mildly impressed.
In fact I am calling for a ban of any comments that have to do anything with Microsoft automation tools. Especially any which must be payed for. Mentioning such things is like going to the Detroit Auto show and talking about how your favorite brand of cars now have windshield wipers.
Actually this it is really simple to see what is going on here. Microsoft has be shovelling bullshit for years that Linux is hurting UNIX more than it is hurting Windows. They must have thought about that statement (assuming it is true) and deduced that Linux is doing a better job at getting to a part of the market that Windows has never touched. How can this trend be changed? By taking away what Microsoft perceives is Linux's edge when competing for big UNIX customers. So Microsoft thinks Linux has a compatibility edge (true but not the whole story). SFU is MS's means of combating the "Linux hurts UNIX more than Windows phenomenon." If anybody thinks this is MS's way of working better with Linux they are nuts. This move is soley to combat Linux.
Let's face it, you cannot make the statment about Linux hurting UNIX more without at some point realizing that it means Linux is doing better at grabbing former UNIX market share.
"Sun will also create a Solaris flavor of the Java desktop system."
I certainly see this being possible. Linux is a non essential part of the Java Desktop System and could easily be replaced with Solaris or one of the BSD's. In fact Sun is using the same interface internally and they are using it with SunRay's running on Sparc/Solaris. Running the JDS on Linux right now is very good for public relations and very economical (all of it was developed by someone else).
I do love Sun and I do tend to post when people who don't know jack shit about them try to expound on what Sun is and is not. Forgive me for only speaking out on a subject that I know. As for your post, you did not respond to a damn thing I said.
Do you honestly think that the 2.6 kernel is going to put Linux anywhere near Solaris in scalability. Will 2.6 run on a 106 processor machine without any futzing (and the first fuck who mentions clustering, beowulf or mosix should bend over and eat his own shit). Will 2.6 do domaining/containers/zones? Will it doe dynamic reconfiguration? If Linux can scale to 32X I will be happy, but to think it is going to be on par with Solaris (Sparc mind you) you are crazy.
You are a total idiot. If Sun is trying to be a Swiss Army knife then what are HP and IBM? If anything Sun this is Sun putting their big toe in the water. They are still a Sparc/Solaris/Java company that happens to also have a few solution for the X86 world. If you think the the Java Desktop System is about Linux (I know you did not mention it; this is for anybody who reads this) then you are sadly mistaken. Anybody who think Linux is an integral part of JDS is a fuckwit. Sun will replace Linux with Solaris and there will be no difference noticed to the user. It is a desktop, an interface, a set of applications. The OS does not matter. In fact OS's are not gonna matter a shit in about 5 years.
Look retard, Open Source software does not do everything. Sometimes there are no other *viable* alternatives for people than what they are paying for. Ask the 90% of desktop computer users using MS Windows and MS Office. Yeah there is OpenThis and OpenThat, but moving people over to it/they and having them be happy is another situation altogether. Trust me, I love Linux, open source and anything that is not MS, but my users don't. Been there, done that.
Therein lies the problem, everybody still assumes that Sun hardware is so much more expensive that Intel based distributors. The assumption is that a Sun box cannot be had for anything less thant $20K and for the same amount I can buy a whole army of Dell servers (which gets you nowhere when you are running Oracle or SAP, etc).
"ACL's are not the only situation where Micrsoft has forged ahead..."
If you consider what MS has for their ACL's as "forged ahead" then you are mistaken. There is nothing intuative or easy about Microsoft's maze of file permissions and layers of confusion.
The problem is that Sparc/Solaris is overkill for commodity tasks such as basic web servers. There's no reason to spend the extra money.
You mean this kind of extra money:
V100 $995
V120 $1995
V210 $2995
V240 $3495
Netra 120 $3395
5 Servers not counting the blade serves, the Vx servers or the Cobalt stuff starting for under $4K. 2 Single Processor, 2 Dual processor and 1 NEBS single processor. It is not longer a valid argument that Sun is just too expensive.
SUn makes a shitload more money from there $20-100k sun servers. They are losing to wintel and lintel.
What does this mean? Are you saying this is where Sun is making money and this is where Intel is threatening them? If so do me a favor and find me a Dell server that will cost me more than $50K (single server, it is possible, but that would be if you install every option). Find me a Dell server with more than 4 processors. Find me an Intel server that does not need some special hacked OS with more than 8 processors. Sun considers anything from 1 to 8 processors Entry Level Servers. Do you think Intel thinks an 8 way box is Entry Level? Yes the very low end of the range you gave is certainly lost for Sun, but $50K - $100K is a totally different story.
Sun keeps laying off and laying off...
They have had 2 layoffs where they have gone from 40,000 enployees to about 35,000 employees. You make it sound like they are SGI or something. This is still a company that does 12 Billion a year in revenue. They have about 6 billion in the bank. They are not going anywhere anytime soon.
Solaris on intel is considered dead...
This is also crap. Sun is now pushing Solaris x86 out the door on server preinstalled!!! To my knowledge they have never done this before. I would guess that there are more people using x86 Solaris now than ever before. With the SCO issue at hand I can see Solaris X86 being a very appealing choice. This is not to say that Solaris x86 is taking the market by storm, but by its own numbers it is far from dead.
java is free so they tet no money from it.
This is the most misunderstood thing in the technology industry. Sun makes plenty of money off of Java. Sun had a choice:
Restrict Java to Sun/Solaris forcing those who want to use it to adopt their hardware/OS.
Allow Java to run on all hardware/OS's, but charge a boatload for it.
Give Java away, allowing it to be easily adoptable/learned and reap the benefits when people begin to run more and more Java applications in their enterprise.
The first option is totally out. Nobody would have been forced to go all Sun just for the sake of Java. The second option would have probably generated some decent sales, but it would never have generated the market that exists today. The third option create a dominating market, of which Sun gets a share of in server sales, services, etc. This was definitely the best option. I am not sure why the media and those in the Tech Industry cannot understand it.
They are a business. Did you expect them to not use a golden opportunity to profit from a less than ideal situation? They spent the money to make sure they were on solid legal ground and now they are glad they did. It amazes me that people bash Sun when they struggle to do anything right for their bottom line and then get mad when Sun uses perfectly sound leverage to try to gain market share with a tactic that may hit home to the beloved operating system.
Slashdot Mirror
You have good points, especially with the Mainframe. However, the DEC/VMS claim only stands on it's own for binary compatibility. You really cannot throw an OS/Platform into this discussion when that OS/Platform is on its way out the door.
"Don't be surprised two years from now when Sun is still the only major vendor of Opteron systems."
Seeing how HP just announced their two way Opteron box and future boxes (4/8 way) I guess this prediction is shot :).
Another uninformed person.
First, who in their right mind would purchase a multi million dollar machine (in your example the E15K) and not know if it will do the job? Not only do you find out if it does the job, you make sure that it will do the job for a long long time. That said, Sun has something called iForce Centers. These are big labs with tons of Sun products and tons of Sun experts where customers spending millions (which is the type of custom refered to in this thread, but certainly others can go to iForce Centers as well) can come and put their application to the test. You can setup any environment with any amount of the hardware and work with the Sun experts to make sure your application runs like a champ. They will throw everything at your app and use everything you want to make your app work. Then at the end of the process they give you an itemized list of hardware, software and man hours, not for the test, but for assembling it for yourself. Of course it is very likely at this point you will give it back to them with a big fat check. But you do not have to.
So you see, it will do the job on the E15K and you will know this before you spend any money.
As for the Altix. Would I spend millions on hardware from a company that is all but dead? No. Do I know *anybody* who is using an Itanium based setup (Linux or otherwise)? No. Do I think Itanium is going nowhere? Yes. Has Itanium based SGI Altix systems been around long enough for me to spend millions on one? No.
We are talking about the enterprise here. Not the big closet the end of the hall, next to your little sister's room. With enterpise computing you take no chances; you make very scientific decisions. Unfortunately your answer is no better than those I blasted who compare the E15K to a whole bunch of Dells in a beowulf setup; not because the Altix can or cannot match the E15K (it might be able to) but because you have no fucking experience in what you are talking about.
My username did not give it away? :)
Glad to be aboard. I have long wondered what everybody else is on. You should have seen my jaw drop when I went to a local Sun event (mind you I had already been using Sun equipment and had been very happy with it) when I learned just how long they support their OS's. I also learned that if you certify you app on Solaris and a newer version of Solaris is released and ends up breaking your app, they will either fix Solaris or they will pay you to fix your app on Solaris. Absolutely Incredible!!! Then the guy telling us all this said he has a small app that he wrote 12 years ago, lost the source code 5 years ago and that he runs on Solaris every day.
No major vendor even comes close to bringing this level of value, longevity and investment insurance.
My favorite Slashdot posts run along the line of "Why would my company purchase a 106 processor E15K when we can just create a Linux Cluster out of Dell PC's?" This lack of understanding of the fundamental difference between a horizontal cluster utilizing parallel programming and a single server image scaled to 106 processors and half a terabyte of RAM, just baffles me. Unfortunately this kind of logic is far too common here. Don't get me wrong, there are people on Slashdot with tons of knowledge and lots of experience. Unfortunately it *seems* that 80% of the posters refer to servers that are likely sitting in their bedroom next to their nightstand.
People who question Sun typically do not know shit about Sun and make comparisons analogous to comparing a Ford and a BMW. Unfortunately Sun is largly to blame for this in their lack of marketing some of the incredible facts I have mentioned in my posts. But those who know, know.
Those companies willing to spend a few million do not care as much about speed as they do about their application being able to run on future versions of a vendor's processor with no recompile (and garanteed). They want to make sure that the OS they are buying with their million dollar setup is supported by the vendor for at least 10 years. They want to make sure that when the next version of the processor the vendor designs comes out that they can put it in their existing box; replacing the present processors or along side of them (without having to bring the box down). If raw throughput was Sun's only goal they could make Sparcs as fast as anybody else. But binary compatibility, open architecture, mix and match and endless support cycles for the OS is what makes million dollar companies say no to raw speed and yes to Sun. Oh and Sun's machines have incredible throughput and perform very well in real world scenarios.
"There is no surprise either that heavy weights such as IBM, Dell, SUN and even HP -- who pretty much designed Itanium -- put some of their eggs in their AMD busket."
Good post, except for the above quote. IBM has one (eServer 325) machine that is based on the Opteron. This machine is being marketed to high performance computing environments and not really as a general purpose machine. Not sure if it was The Register or The inquirer, but one of the sites had quotes from a top IBM executive on their lack of plans for more Opteron based models (i.e. a 4 way and some other form factors).
Dell has outright said they are not supporting the Opteron and were holding out to see what happens in the market (which could mean they were holding out for Intel).
HP's Opteron systems are a rumor and have yet to be substantiated.
That leaves Sun. Sun is the only major vendor with both an available Opteron system and another (4 way, 6 hdd) on the way. They are also the only major vendor that has made long term commitments to Opteron on a variety of setups/form factors, including workstations and 8 way servers. Sun has also promised to deliver 32/64 bit Linux (third party) and 32/64 bit Solaris on Opteron.
In fact when Sun announced all of there support, that should have tipped us all off that the Intel annoucement was coming. When would Sun do what everybody else is doing, unless everybody else was actually gonna be doing something else? Don't be surprised two years from now when Sun is still the only major vendor of Opteron systems.
Man you are way off regarding automatic update and Linux. I use apt-rpm and can set it up to do many kinds of automatic updating. I can deploy updates from one machine to the entires company either automagically or manually. I *never* have to visit a *NIX box. In fact if you do *have* to go visit a *NIX box for anything but a complete hardware failure then you are a shitty admin. Windows boxes are the only boxes that I ever had to go visit. What you laud MS for being able to do I have been able to do in *NIX for years, with or without deployment tools such as apt. *NIX environments are fully automatable out of the box and have always been that way. Don't come here telling us about some new Microsoft automation feature and expect any *NIX admin worth a shit to even be mildly impressed.
In fact I am calling for a ban of any comments that have to do anything with Microsoft automation tools. Especially any which must be payed for. Mentioning such things is like going to the Detroit Auto show and talking about how your favorite brand of cars now have windshield wipers.
Actually this it is really simple to see what is going on here. Microsoft has be shovelling bullshit for years that Linux is hurting UNIX more than it is hurting Windows. They must have thought about that statement (assuming it is true) and deduced that Linux is doing a better job at getting to a part of the market that Windows has never touched. How can this trend be changed? By taking away what Microsoft perceives is Linux's edge when competing for big UNIX customers. So Microsoft thinks Linux has a compatibility edge (true but not the whole story). SFU is MS's means of combating the "Linux hurts UNIX more than Windows phenomenon." If anybody thinks this is MS's way of working better with Linux they are nuts. This move is soley to combat Linux.
Let's face it, you cannot make the statment about Linux hurting UNIX more without at some point realizing that it means Linux is doing better at grabbing former UNIX market share.
"Sun will also create a Solaris flavor of the Java desktop system."
I certainly see this being possible. Linux is a non essential part of the Java Desktop System and could easily be replaced with Solaris or one of the BSD's. In fact Sun is using the same interface internally and they are using it with SunRay's running on Sparc/Solaris. Running the JDS on Linux right now is very good for public relations and very economical (all of it was developed by someone else).
I do love Sun and I do tend to post when people who don't know jack shit about them try to expound on what Sun is and is not. Forgive me for only speaking out on a subject that I know. As for your post, you did not respond to a damn thing I said.
Do you honestly think that the 2.6 kernel is going to put Linux anywhere near Solaris in scalability. Will 2.6 run on a 106 processor machine without any futzing (and the first fuck who mentions clustering, beowulf or mosix should bend over and eat his own shit). Will 2.6 do domaining/containers/zones? Will it doe dynamic reconfiguration? If Linux can scale to 32X I will be happy, but to think it is going to be on par with Solaris (Sparc mind you) you are crazy.
You are a total idiot. If Sun is trying to be a Swiss Army knife then what are HP and IBM? If anything Sun this is Sun putting their big toe in the water. They are still a Sparc/Solaris/Java company that happens to also have a few solution for the X86 world. If you think the the Java Desktop System is about Linux (I know you did not mention it; this is for anybody who reads this) then you are sadly mistaken. Anybody who think Linux is an integral part of JDS is a fuckwit. Sun will replace Linux with Solaris and there will be no difference noticed to the user. It is a desktop, an interface, a set of applications. The OS does not matter. In fact OS's are not gonna matter a shit in about 5 years.
I cannot look at two monitors in one glance, so what is your point?
And it is why the Utra Sparcs have had 8MB of L2 Cache for a long time.
Look retard, Open Source software does not do everything. Sometimes there are no other *viable* alternatives for people than what they are paying for. Ask the 90% of desktop computer users using MS Windows and MS Office. Yeah there is OpenThis and OpenThat, but moving people over to it/they and having them be happy is another situation altogether. Trust me, I love Linux, open source and anything that is not MS, but my users don't. Been there, done that.
Therein lies the problem, everybody still assumes that Sun hardware is so much more expensive that Intel based distributors. The assumption is that a Sun box cannot be had for anything less thant $20K and for the same amount I can buy a whole army of Dell servers (which gets you nowhere when you are running Oracle or SAP, etc).
You need to come to South Florida. Ferrari seems to be doing very well from where I am sitting. I see probably 10 a week and never see Maserati's.
Why don't you check when Sun submitted that benchmark and how old that system is. They have written off TPC-C a long time ago.
Can you say Troll?
"ACL's are not the only situation where Micrsoft has forged ahead..."
If you consider what MS has for their ACL's as "forged ahead" then you are mistaken. There is nothing intuative or easy about Microsoft's maze of file permissions and layers of confusion.
We started moving into x86/Linux for high end commercial servers the moment the economy started to go south and we had to cut costs.
Who is "[w]e" and what do you consider high end?
The problem is that Sparc/Solaris is overkill for commodity tasks such as basic web servers. There's no reason to spend the extra money.
You mean this kind of extra money:
5 Servers not counting the blade serves, the Vx servers or the Cobalt stuff starting for under $4K. 2 Single Processor, 2 Dual processor and 1 NEBS single processor. It is not longer a valid argument that Sun is just too expensive.
SUn makes a shitload more money from there $20-100k sun servers. They are losing to wintel and lintel.
What does this mean? Are you saying this is where Sun is making money and this is where Intel is threatening them? If so do me a favor and find me a Dell server that will cost me more than $50K (single server, it is possible, but that would be if you install every option). Find me a Dell server with more than 4 processors. Find me an Intel server that does not need some special hacked OS with more than 8 processors. Sun considers anything from 1 to 8 processors Entry Level Servers. Do you think Intel thinks an 8 way box is Entry Level? Yes the very low end of the range you gave is certainly lost for Sun, but $50K - $100K is a totally different story.
Sun keeps laying off and laying off...
They have had 2 layoffs where they have gone from 40,000 enployees to about 35,000 employees. You make it sound like they are SGI or something. This is still a company that does 12 Billion a year in revenue. They have about 6 billion in the bank. They are not going anywhere anytime soon.Solaris on intel is considered dead...
This is also crap. Sun is now pushing Solaris x86 out the door on server preinstalled!!! To my knowledge they have never done this before. I would guess that there are more people using x86 Solaris now than ever before. With the SCO issue at hand I can see Solaris X86 being a very appealing choice. This is not to say that Solaris x86 is taking the market by storm, but by its own numbers it is far from dead.
java is free so they tet no money from it.
This is the most misunderstood thing in the technology industry. Sun makes plenty of money off of Java. Sun had a choice:
The first option is totally out. Nobody would have been forced to go all Sun just for the sake of Java. The second option would have probably generated some decent sales, but it would never have generated the market that exists today. The third option create a dominating market, of which Sun gets a share of in server sales, services, etc. This was definitely the best option. I am not sure why the media and those in the Tech Industry cannot understand it.
You proved his point right there. *Almost* is something that someone buying Sun does not want to consider. Almost is not good enough.
They are a business. Did you expect them to not use a golden opportunity to profit from a less than ideal situation? They spent the money to make sure they were on solid legal ground and now they are glad they did. It amazes me that people bash Sun when they struggle to do anything right for their bottom line and then get mad when Sun uses perfectly sound leverage to try to gain market share with a tactic that may hit home to the beloved operating system.