Sun may have been the last major server maker to embrace Linux, but Chief Executive Scott McNealy argues that the company will benefit more than its competitors from the open-source software. Update: A day after Sun Microsystems CEO Scott McNealy said open-source software is his company’s friend, a prominent advocate of the collaborative programming philosophy, Eric S. Raymond, has called upon the server maker to open the code of Java.
Sun CEO: Open source is our friend; ESR Replies
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
I think it should be clear to Sun that Java does not need to be held so close to their vest. The original premise of using Java as a point of leverage to sell servers has failed. I would offer that the R&D and development costs on Java are a money loser for Sun at this point, and they are clinging to control of JAva as a means of keeping Sun relevant.
If Sun wants to cut costs they should open Java up and let the community start hacking on it. Keep the JCP – there probably should be a group deciding what is official Java. Perl has Larry Wall and Linux has Linus doing pretty much the same thing – deciding what is and what isn’t in the product they created.
My only thought is that they are holding out hoping IBM will acquire them to get to Java.
In any case I really don’t know if it makes much difference, I don’t know of anyone really holding off of Java because of its closed nature.
http://www.catb.org/~esr/writings/let-java-go.html
Because times and are changing. Software doesn’t last forever. Customers allover the globe are slowly replacing commercial Unix with Linux. Linux won’t last forever, will be have to be replaced at some point. How many people use 128-way processors? And how efficient are they compared to clustered Linux machines?
Times are changing, and I think SUN could cut tremendous cost by investing in Linux instead of Solaris.
Finally, I’d appreciate it if you were less insulting.
>How many people use 128-way processors?
Many people who have the money to ask Sun to update their software for the next 30 years. Sun has such contracts for 30 years in the future, yes, and these contracts do keep Solaris *relevant*.
Both Sun and ESR have said this before.
I talked to sun, they told me they would like to choose an open source license, they told me that they cannot find one and that it may be done in the future. They claim people like IBM and Microsoft will use it like they have done in the past and screw it up. They want to guide the development of java. Johnathon told me that Java is why they have a software business, its how they make money. So why in the HECK would they open source everything and screw over their business? ESR you have ben acting like a pyscho. NO offense I’ve talked to you before in support of you but sometimes you need to keep your mouth shut because your not a businessman you want everything to be free and nobody to have a job.
This is my first time hearing of 30year long contracts, especially in the IT/Computing industry. The industry changes too fast for any business to be locked in a 30 year contract with a single vendor.
Besides, contracts are often revised and renegotiated. I still stand by my point that SUN should channel most of their resources towards Linux. But of course SUN lacks focus and business acumen. Great engineers don’t automatically mean great business managers.
Linux as a technology is a lot more flexible to work with from a vendor perspective. Why for instance is SUN wasting much resources porting Solaris to X86 when they could easily package a customized linux solution for their customers?
Why spend so much R&D on a technology when you could share the cost with so many people and Big corporations around the world?
If SUN focuses on Linux then could provide customized solutions for clients on any major platform of their choice. Heck, they can even divest into the embbeded technology industry, a young promising and growing sphere in IT. Can they do that with Solaris? Do you see the reason why Solaris is an unbearable cost that can be easily eliminated with a customized and well designed Linux solution?
Businesses could careless about whether their mainframe has Linux or Solaris running on it, as long as it is serviceable and it performs the intended task flawlessly and of course it is cheap.
When the industry was young and naive, business have been known to spend up to $25 000 just to run a website. SUN still thinks clients like those exist today. I hope they have enough thirty year long contracts to milk to survive.
I challenge you to find ONE EXAMPLE of a thirty year contract signed by ANY company WITH ANY company in IT.
The average IT firm does not even exist that long. No one would try to forecast out that far.
WAITING FOR ONE EXAMPLE. ONE.
Solaris, its supported, scalable and it works GREAT on the SPARC platform which is extremely great. It’s supported stable and secure. Solaris vs. Linux in support and security, Id’ say solaris beats it.
I choose solaris freebsd and linux because i like to give customers choice, however, solaris is cheaper in the long run. You can buy a solaris OS and its supported you have 1 company to go to you dont have all these distros and you dont have to wait on your pacific distro to release updates and you dont have to worry about if your machine breaks you have to pay $/hour to fix it. It’s more cost efficient in the long run. However, When you are doing things that require the OS to be altered linux may be more suitable. It really depends on what you are doing, clearly, it depends per business. I’m not biased I use linux as desktops at home. I love it and I love the other OS’s as well. So, although I’m not a linux fanatic and i dont want software companies to die, I probably will get alot of mean mean mean insulting comments back and thats expected.
i dont see innovation in the open source field, at all. None. Tell me an innovation in open source. I still love open source though.
IBM Mainframes would be one example. Many out in the wild are quite old, still maintained, and still supported.
This has nothing to do with OSes. This has everything to do with SUN providing solutions for clients, revaluating their services and broadening their sources of revenue, while keeping costs down/minimal. I’m sorry, Solaris can scale 7 million processors, and SUN might still go down under, if they don’t reinvent themselves as players in the market.
As much as I would like to see Sun open Java, that’s not practical for the company. Giving away Java will basically make it faster for Sun to go bankrupt. Overall that’s the last strong thing Sun has.
It is said that a leading advocate for the open source is simply an idiot. No I am not kidding. Here is why, java is already free. It is not open source as Raymond wants it, but it is free, and you can easily develop for it. If you can go and burn Linux CDs, you can also burn and distribute java packages along with it. It is quite accessible to everybody. There is no way, millions of developers are going to switch to java just because Sun made it open as Raymond wants it to be open. I respect Raymond’s wishes, and in fact I would like to see the same thing, but please don’t bullshit us here.
Willett said. “I don’t think Python is going to take over Java. Java is the language to learn. There’s an infrastructure set up where people get trained in Java. People do, because they know they’ll get jobs that will pay them well.”
Heh heh. That’s pretty funny.
I learned Java a long while back but then stopped using it when I realized that Sun wasn’t really concerned about Java on the desktop. Would’a loved to see Sun support gcj + Swing… Anyhow, H2O under the bridge. I’m long-done with Java.
Just recently began learning Python. I *love* it so far. Simple, clean, consistent, plays well with C and C++. I’m now wondering why there’s all this concern over Mono and/or Java on GNU/Linux. Why bother — just use Python.
Really. Just use Python.
Does python offer me strong typing and rigorous, enforced contracts? If not its useless for me, or for any serious largescale development.
Good for prototyping though, and as a toy.
Businesses could careless about whether their mainframe has Linux or Solaris running on it, as long as it is serviceable and it performs the intended task flawlessly and of course it is cheap.
This is the most worngheaded view of how companies and enterprises work. Customers do care what software runs on thier systems. Did you know that there are still companies running solaris 2.5.1 and 2.6 in thier enterprises today and they don’t want to upgrade becuase of thier invesment in thier software systems.
If you don’t mind my asking how many companies Datacenters have setup and help run.
Solaris is one of the main reason people buy Sun systems. It’s rock solid and reliable and extremely mature. Linux lacks seriously in features enterprises really care about. Linux can’t even do crashdumps by default without patches from SGI. Why do you think SGI and IBM are so keen on adding such features as crash dumps, profiling and tracing frameworks to linux. Because linux is not ready for the enterprise in its current state. It might get there eventually.
IBM and SGI are into linux because thier own versions of unix aren’t as popular as Solaris. I don’t think AIX customers are going to like what IBM is doing, I think IBM knows that and they are not going to get rid of AIX just yet. SGI and IRIX are just about dead and linux on IA64 is carving a niche market for them.
I think it’s hard to take someone like ESR seriously when he makes ridiculously absurd insinuations about Sun scheming with SCO to down Linux. Talk about a huge blow to someone’s credibility.
I really can’t fathom some of the open-source zealotry out there. What a ghastly thing it is for a company to invest so much time, money and effort into a technology and — *gasp!* — have the audacity to keep “control” over that technology! The only thing worse is to turn that investment into revenue — the ultimate evil!
Look, I’m a software engineer by day. I develop with Java. Of all the programming languages I have experience with (probably not many), it’s the best language I’ve dealt with so far. Yeah, it’s not perfect. I don’t like everything. There are some quirks and bugs… but to be honest, I’m able to do A LOT of good stuff with a lot less effort through Java. If Java ever gets turned over to the open-source Linux-weenie el33t haxors, I might as well throw myself off a bridge.
As for Solaris/Linux… Sun has a bunch of *talented* engineers who are very devoted to Solaris development, and put A LOT of effort into Solaris. It’s their baby. Are we saying Sun should tell those engineers, “Look, you’re working Linux now, because we’re dropping Solaris like a rock… and if you don’t like it, you’re fired and we’ll replace you with some Linux guy off the street.” That’s Sun’s big “plan for success”? Doesn’t sound like a bright idea to me. A lot of this sounds more like “It would be good for Linux…”, yet packaged under “It would be good for Sun…”.
All JMO,
John
Whichever way it goes, I’d love to see the best damn Java language + APIs man can produce. I love it as it is, and I want to love it more.
I’d like to point out that Raymond’s reasons for open-sourcing Java does not resonate with me strongly. He’s going to have to be more convincing if he makes big statements like, “The choice is between control and ubiquity.” Ubiquity? Exactly how Eric? The licensing issues? Give me a break.
I can’t even begin to comment on the Perl and Python comment. Lol. For such a so-called “experienced” programmer, how can he compare apples and oranges to such a large audience and get away with it?
For christ sake and for the last time, this is not about OSes. I don’t care if Linux can’t do crashdumps, which I think it does. I care about performance though. And that’s all clients care about. That and stability.
The university I am at is replacing all our Unix labs(Mixture of IRIX HPUX Solaris) with Linux. Ask them why? “Better performance, much cheaper” is what will echo in your ears. Try telling them “but Linux doesn’t crashdump”.
Several of the colleges database servers are slowly being converted too. Same reasons as above. If my business can just get comparable performance for cheaper, I don’t care if your product has “intelligent-software-autobug-tracking-abstracted-virtual-objected-ana lysis-detection”, I’m chosing the cheaper alternative that gives as much performance and is as stable.
This is not a “Solaris Vs Linux” slug fest. This is reality. People are just not seeing the advantages of many commercial Unix over Free Linux anymore. Once again I don’t care for neither Solaris’ or Linux’ features.
Because the guy thinks that he is commanding millions of open source soldiers, thanks to the attention he gets by bashing Microsoft. You bash Microsoft, you get immediate attention from the IT media. If you have zero knowledge on security, but claim to have a report which shows that Microsoft windows is inherently insecure, news.com will make you the number one story. If you say presentation software is stupid, nobody will care, but if you say power point is stupid, you will become the number one news. That’s why Raymond thinks he is someone who has to be respected, but I don’t see him as a IT expert. The guy knows few tricks with perl, python etc… and thinks he is the best programmer out there and an IT expert. He might be good at programming, so am I, but I don’t reach the same conclusions.
Perl is a good language, it will always there. Python is also good, that’s what I heard, but none of them can be compared to java or c#. There is too much work behind those platforms. You have to have thousands of good programmers and experts to be able to compete with these two languages. In open source, you don’t have enough number of people with enough level of skill. You often see people without much vision and technical skill claiming to be better than Microsoft. I personally saw this over and over again. Many of these people do not even know what is java or c# is about.
anon wrote:
Does python offer me strong typing and rigorous, enforced contracts? If not its useless for me, or for any serious largescale development.
I can’t comment on that because, as I said, I’m only just learning Python now. However, my guess is that if Python does not offer what you want, and C++ does, you can compile whatever C++ modules you need and then access them from within your Python code (though I don’t yet know the details of how that’s done). C++ backend, Python front-end.
Regarding large-scale development, I’ve heard that Python scales well, but perhaps someone here could comment on how well it does with large projects, as compared to Java and Mono. That is, when running a large application, how much resources the PVM consumes compared to a typical JVM and whatever Mono uses.
MrPillows wrote:
I can’t even begin to comment on the Perl and Python comment. Lol.
Are you referring to ESR’s:
Sun’s insistence on continuing tight control of the Java code has damaged Sun’s long-term interests by throttling acceptance of the language in the open-source community, ceding the field (and probably the future) to scripting-language competitors like Python and Perl.
Please do comment. What’s nonsensical about that? (I’m no Java guru, so I’m genuinely curious.)
There is more to java or c# then you see. There are tons of development tools that support these languages. Particularly java is very well established. There are tons of papers that discuss tons of issues related with this language and the platform itself. There are tons of apis, libraries, jvm implementations, etc… for java.
Even if python had a much better syntax, without that level of support it is nothing compared to java. Perl is not a language that you can use for serious long-running development. I am quite an expert on perl. It syntax practically makes it useless. I heard that python is better, much better when it comes to syntax. Actually I have seen very interesting implementations, so probably I am going to learn it sooner or later. But taking on java requires python developers to develop something as big as Java. That’s probably not going to happen. You might be thinking only about syntax and few libraries, there are tons of more details there. Garbage collection, language specification, RMI, Java Spaces, etc… Just taking on garbage collection is not a simple task.
Finally python is not even a compiled language.
“This has nothing to do with OSes. This has everything to do with SUN providing solutions for clients, revaluating their services and broadening their sources of revenue, while keeping costs down/minimal. I’m sorry, Solaris can scale 7 million processors, and SUN might still go down under, if they don’t reinvent themselves as players in the market.”
I couldn’t agree More! SUN needs a serious management enema i cringe every time that idiot MCNealy speaks. It would be shame for such a cool company to fall to the way side becasue they cannot adapt to changing customer needs.
“Solaris is one of the main reason people buy Sun systems. It’s rock solid and reliable and extremely mature. Linux lacks seriously in features enterprises really care about. Linux can’t even do crashdumps by default without patches from SGI. Why do you think SGI and IBM are so keen on adding such features as crash dumps, profiling and tracing frameworks to linux. Because linux is not ready for the enterprise in its current state. It might get there eventually.”
Finally someone else that understands why linux is not ready for prime time in the enterprise. Attention Linux zealots: read it, learn it, live it…
I think linux would be ready for the enterprise if it were more securite, sorry, It’s not. Does Linux perform great? Yes. With Linux 2.6 there is a huge performance increase. It’s performance is no longer a problem its security. Furthermore, If I’m a huge company and I want to put millions of dollars onto a platform and have that platform hold sensetive data would I choose Linux or Solaris? Solaris- It’s stable, supported by one company, legal proof, etc. We’ve already seen that linux illegally ripped off *BSD. ESR prooved that. As I said before I’m not a GPL advocate as I love BSD. I’m ok with the GPL. GPL = free for all, must be provided for all, cannot close code, everyone can contribute. It’s the communist’s Ideal for GPL. Believing in software communism is not wrong, it destroys alot of jobs though. I never thought of the GPL that way before but IT’s true if you read it and understand it. ESR and the FSF and more go on bashing corporations who hire thousands of programmers. They dont have to worry about their jobs they can write news articles and get paid, they can go speak and say stupid things about some major company and claim he represents the entire open source movement. These people are radical. I, myself, is a major advocate of the BSD and MIT licenses, I like apache too 😉
compare java 5 years ago and compare its performance now, and no, dont use MSJVM. Talk about improvement. Cross-platform software hasn’t been so good. It’s also great for the enterprise and server side.
The university I am at is replacing all our Unix labs(Mixture of IRIX HPUX Solaris) with Linux. Ask them why? “Better performance, much cheaper” is what will echo in your ears. Try telling them “but Linux doesn’t crashdump”.
Universities are not serious data centers or enterprises. I used to maintain machines in my engineering department’s labs when I was in school too. We were replacing old HPs,DECs and Suns too (we also got newer Suns) and I was evangalizing linux then too. But University != Real world. In a university if the lab machine goes down a service work order is filed and students move to another machine. The mail server is down for a few minutes/hours no big deal.
Once our sysadmin made a stupid mistake in one of his back up scripts and destroyed about 200 student home directories and the last back up was a from a week before and he didn’t even get yelled at. Students lost important work. His backup script relied on a nfs mounted directory on a linux machine. The linux machine crashed and the back script destroyed most of the home directories.
Several of the colleges database servers are slowly being converted too. Same reasons as above. If my business can just get comparable performance for cheaper, I don’t care if your product has “intelligent-software-autobug-tracking-abstracted-virtual-objected-ana lysis-detection”, I’m chosing the cheaper alternative that gives as much performance and is as stable.
Good for your school. The last time I chatted with my friend who still goes to the univeristy I used to they had moved our trusty HPUX mail server to a dell running linux and it wasn’t up for more than 14 days at a time. So to handle the load they moved it to a 8 way dell and it now stays up for may be 35-70 days at a time. I rember the HPUX box was up for 200+ days and it was a single 240 mhz machine. Is the new machine faster? yes. Is it more stable? no. Does it work for the university? probably because it is cheaper Will this work for a big company doing millions of dollars of business? No.
Enterprise customers do care about debugging and self analysis. Have you seen the IBM self healing systems Ads??!!!
This is not a “Solaris Vs Linux” slug fest. This is reality. People are just not seeing the advantages of many commercial Unix over Free Linux anymore. Once again I don’t care for neither Solaris’ or Linux’ features.
No its not reality you are still in a University. When you graduate and come into the realworld and talk to real customers may be then you can tell me what they do and don’t want.
I ask you again how many datacenters have you setup and run successfully?
Universities are not serious data centers or enterprises
This is what happens when people have no idea what they are talking about. For really big data centers and enterprises, linux is not even an option. We have a very serious IT staff here. No student is involved with the type of the stuff you are talking about, unless of course it is research computer. We mostly use Unix machines of course, not because they are better though. Mostly people know these systems, and traditionally people had unix boxes on these systems. I don’t know much about windows systems either. I know a lot about Linux and Unix, but since I am a rational person I don’t make up stuff.
A university moving to windows shows something. Pay attention to why they made that choice, understand that. Don’t rush to bashing Microsoft. Whether you bash Windows or not will not change others’ minds.
Oh by the way, please don’t try to insult other people based on where they do their work. Our department had one of the best IT people out there. Colleges may be a different matter, but universities like berkely, stanford are like enterprises. There are tons of people working there. Managing their email, etc.. all take lots of effort.
But if you think enterprises show something, then look at the market share of apache on the enterprise companies. Yes IIS has a higher market share in the enterprise companies. Just a quick note where your logic may take you.
> Sun CEO: Open source is our friend
Of course it’s their friend. So many people that work for NULL for increasing the value in their moneybox. I would call all free volunteers ‘friends’ too as long they work without any salery.
No its not reality you are still in a University. When you graduate and come into the realworld and talk to real customers may be then you can tell me what they do and don’t want.
I ask you again how many datacenters have you setup and run successfully?
Don’t be silly. Who told you I was a student?
Look, I don’t even care because you are turning this into a “Solaris is a more secure more stable OS than Linux” debacle, and frankly, I think that’s glaring fanboyish. I really don’t have time for that.
If SUN wants to survive they have market trends to look at and follow. But Solaris will only get them so far. Linux will take them places. That’s my opinion, you have every right to disagree with it, just don’t start the Linux is unstable insecure crap.
I made a blanket statement about universities, I should retract it. Some big universities like Stanford are run like big data centers. The last I checked most for Stanfords Unix machines are sun boxes. You could code your assignments on any platform but they were graded on Solaris. This of course depends on the class you take.
Look I am not trying to claim one platform is better than another. I use Mac OS X, windows, solaris and linux.
But when some one comes out and tries to give out opinions like Sun should just adopt linux because it is what everyone wants. And performance is the only criteria and nobody really cares about RAS it’s all about price/performance. That just exudes inexperience and lack of real market data.
Java is what it is today because Sun kept tight holds on it. Look at what happened with J++ and microsoft. I don’t think it is prudent ot make that mistake again. The JCP works quiet well, IBM, BEA and many companies have thier own java implementations. Customers do have a choice on which implementation to use.
“Look, I don’t even care because you are turning this into a “Solaris is a more secure more stable OS than Linux” debacle, and frankly, I think that’s glaring fanboyish. I really don’t have time for that.”
Um, actully i got the impression that this was a real world mission critical data-center environment vs small time University environment. Having worked in a very large mission critical real world environment i can whole heartedly agree with Raptor. Would i consider moving some windows server functionality to a linux server? Perhaps. Would i replace the 16×16 V-Class HP server houseing the US’s largest cable companies 100GB finacial database? Not on your life.
Linux is not as stable as Solaris. That’s a very well known fact. That’s why IBM is helping out Linux.
If Sun makes Java open, if it promotes Linux, etc.. how the hell do you think Sun is going to survive.
I also don’t like lots of actions of this company, but please don’t be stupid and suggest that they should give away everything for free and just become a commodity company. You know they have no chance of surviving by selling their expensive hardware anymore. It is just like Apple. I also don’t see why they have to open java. So far java become java by being the way it is now.
While it probably isn’t OSI approved
The Java source
is at
http://wwws.sun.com/software/communitysource/j2se/java2/download.ht…
Um, actully i got the impression that this was a real world mission critical data-center environment vs small time University environment.
The University I work at has over 45000 students (Not including Staff). It conducts millions of transactions a month. How that is small time and less mission critical baffles me?
Would i consider moving some windows server functionality to a linux server? Perhaps. Would i replace the 16×16 V-Class HP server houseing the US’s largest cable companies 100GB finacial database? Not on your life.
Yes, yes, we’ve all heard that before. Linux is only good for proxies, firewall and web servers. Friend, you are living in the early 90s. When you here news of Linux replacing commercial Unix what do you think those clients were using those Unix for? Webserving?
Linux is not as stable as Solaris. That’s a very well known fact. That’s why IBM is helping out Linux.
Show me the facts. Then I’ll shut up.
I use a number of proprietary UNIX platforms in my field of work. Many of the system tools for these platforms are now written in java. As a result they are slow to the point of being useless.
Fortunately I use a lot of Linux systems, where java does not come with the distribution. System tools are typically written in C or python.
As far as I’m concerned, Sun can keep java to itself.
Yes, yes, we’ve all heard that before. Linux is only good for proxies, firewall and web servers. Friend, you are living in the early 90s. When you here news of Linux replacing commercial Unix what do you think those clients were using those Unix for? Webserving?
Well good for those companies that can make Linux solutions work for them. Don’t get me wrong i like Linux and i fell eventually i might get to be Enterprise ready primarly through companies like Redhat,HP and IBM. Currently i don’t feel linux is quite there yet and as an experienced SA i can’t currently recommend it for mission critical deployments.
The University I work at has over 45000 students (Not including Staff). It conducts millions of transactions a month. How that is small time and less mission critical baffles me?
Simple…When the example 100GB database goes down, hourly revenue losses are messured in 10’s if not 100’s of thousands of dollars. Mainly due to paying employees to sit on their thumbs becasue they can’t do their work. Thankfully due to the excellent service HP provides. Our unscheduled downtimes(which were very rare) were limited to on average 1-2 hours depending on the issue.
So what if students can’t access their email for an hour or two.
ESR is a good guy. He recognized that the central issue over this technology is ‘who controls it’. If a technology does not give the user control over the tool, than the technology is weak and is of poor quality. Our technology allows us to master nature, but if we can not have control over these tools, than the technology always ends up recomplicating, and repeating the original problem.
Open source users would be well off to discover tools like Perl and Python, and it is possible that we would gain from sharing knowledge with the BSD community. Ultimately these tools are going to be more useful than tools that we have no control over. The money in Java is an illustion. In countries like Canada and America, wealth is highly conentrated. Fifty percent of the population owns nintey five percent of the wealth. The top ten percent owns fifty percent of the wealth < http://www.policyalternatives.ca/publications/ragsandrichessummary…. >.
I’d like to know what ESR thinks we should do not that Microsoft owns XML.
There is more to java or c# then you see. There are tons of development tools that support these languages.
Well, for Python there’s an IDE called IDLE. I’m pretty sure it comes with its own debugger called pdb, and of course, there’s vim.
Particularly java is very well established. There are tons of papers that discuss tons of issues related with this language and the platform itself. There are tons of apis, libraries, jvm implementations, etc… for java.
Well, I hear that Python has a rather extensive library, just like Java.
Even if python had a much better syntax, without that level of support it is nothing compared to java.
Huh? Python is fully implemented. I’m not sure what you mean there.
Perl is not a language that you can use for serious long-running development. I am quite an expert on perl. It syntax practically makes it useless. I heard that python is better, much better when it comes to syntax. Actually I have seen very interesting implementations,
Jython?
so probably I am going to learn it sooner or later. But taking on java requires python developers to develop something as big as Java. That’s probably not going to happen.
I think it already has happened. The Python community is rather large, and there’s a lot of modules out there to make your work easy.
You might be thinking only about syntax and few libraries, there are tons of more details there. Garbage collection, language specification, RMI, Java Spaces, etc… Just taking on garbage collection is not a simple task.
Dunno about RMI and “Java spaces”, but Python is garbage-collected, and it’s already stable and implemented and ready for use.
Finally python is not even a compiled language.
Python compiles to an intermediate byte code just like Java. Then that byte code gets interpreted by a PVM (just like a JVM). Also, there’s some sort of JIT compiler for Python called Psyco that you might want to look into.
The University I work at has over 45000 students (Not including Staff). It conducts millions of transactions a month. How that is small time and less mission critical baffles me?
from http://www.clarion.edu (http://www.clarion.edu/admiss/fast_facts.shtml):
Clarion Campus (main): Established 1867, Clarion, PA; three miles north of I-80, Exit 64; 44 buildings; 99 acres.
Degree programs: 90+ academic degree programs at associate, bachelor’s, and master’s levels;
Enrollment: 6,000 students from 32 states and 30 countries in 1999;
Faculty/Student Ratio: 1:18;
Average Class Size: 22 to 26;
Students Receiving Financial Aid: About 75%.
So your schools enrollment went from 6000 – 45000 in 5 years. Why isn’t he website up to date? May be the other campus has more students.
Venango Campus: http://www.clarion.edu/academic/venango/index.htm
HTTP Error 404
404 Not Found
The Web server cannot find the file or script you asked for. Please check the URL to ensure that the path is correct.
Please contact the server’s administrator if this problem persists.
Looks like you linux webserver/database has a problem and your school is losing millions!!!
Ok I hate to have done this. I have always tried to be objective and reasonable in my discussions here on OS news. The only reason I got into this debate was to show this person that if he wants to paritcipate in a good doscussion he should stop trolling.
This discussion is about Java, Sun and open source. While this thread was not completely off topic it is waste of time. So please stop this Sun should do whatever I think to save thier business nonsense.
How many successful opensource companies have you and ESR run with huge profits better than Sun? How to run a business adivce coming from ESR is a joke he has never run a multi-billion dollar buisness.
What makes you think root works at Clarion? I’m not saying he doesn’t… but man… I’d hate to imagine someone might think I work at Comcast.
John
I am not sure. But I hazarded a guess. You usualy have to be on campus or dialed into the campus network for the ip logger on OS new to pick up the address. So it is highly likely that root either goes to clarion or is logged into its network.
I doubt the univserity also runs an ISP. Schools only allow students, faculty or staff to access thier networks.
Unless he is logged into someone elses account. if the university root goes to has 45000 students and is in PA it won’t be hard to track it down. But I am not going to spend the time to do that. I have seen a lot of linux fanboys and many of them pit roots profile. Serious professionals are usually more reasonable and don’t make sweeping generalizations like root does.
Well all of this is just a guess and hunch. All I am trying to do is have a good discussion here and root is not helping.
just shut up and quit embarassing yourself. it’s pretty obvious that raptor knows what he’s talking about and you don’t.
Raptor is a Sun fanboi who will bash Linux until he banned from here. Why bother arguing with someone who clearly doesn’t understand Linux or what it can do. He’ll argue that “Sun is for Enterprise” mantra until the day he dies. You and I know that Linux is a very reliable OS and that Sun has jumped the shark. I guess its no fun seeing an OS you have put so much time into fall, but hey that’s IT. We went through the same thing with Vines and Netware. Times change. Bashing the OS that is replacing your current fav isn’t going to stop the rise of Linux and migration away from Proprietary Sun technologies and Solaris. Sorry but that’s a fact and Sloris has already seen its glory days. I guess you hang out with the Amiga users though.
Those of us who know Sun and how they’ve treated Linux in the press do chuckle at Sun these days. The constant Linux bashing they used to due in the past just makes their recent attempts to embrace Linux all the more desperate. Its too late and nobody wants to see Sun succeed with Linux. Sorry Sun. You had your chance. Thanks for buying Star Office though. We Open Office users appreciate it.
I like them all. I think Sun has a good balance between them all. Solaris is for sun server. Nothing is better than that combination. For cheap deesktop there is linux and java desktop using x86. the hardware is even Windows certifcated, so if a customer don’t like ist, he can put linux off and switch to windows.
Abbout Java, I don’t understand why Sun should open source it. It doesn’t make java better, just look at some open source implementations of java 😉 I don’t mean that the people behind open source java projects lack skill. They all are smart, but they lack resource. At sun, ibm, hp there a alot of people doing java research everyday (and get payed for doing it 😉
>> Universities are not serious data centers or enterprises.
???????
Many universities have
– more users
– more data
– more apps
– more cool cutting edge hardware
than many of the businesses I have been in.
>> For really big data centers and enterprises, linux is not even an option.
Google and Yahoo run larger server farms than you have ever seen no matter where you work and who you are on free software and commodity hardware.
You can trump the size of their server farms, don’t bother with a counterexample
Google and Yahoo run larger server farms than you have ever seen no matter where you work and who you are on free software and commodity hardware.
Last I checked google was the only true fully linux shop of the two you mentioned. Yahoo many systems one of which is linux.
It’s funny you mention Google. Google is a search engine which indexes webpages stored on other enterprises systems. Most of googles data is backed up by other companies, those whose sites it indexes. Googles crawlers just index those documents all around the internet and can reindex them on a failure.
Yes google does maintain a cache of webpages it indexes. But if the node the cached page is on dies the it an refetch it from the source. Databases at enterrpises don’t have that luxury once your main database crashes the data is gone. You have to restore from backup and maybe a redundant system.
One can’t really draw parallels between google and conventional enterprises. Google does what it does well because they have R&D on building thier infrastructure and linux on comodity hardware fits well in thier architecture.
Google’s infrastructure is very unique and not all companies’ IT departments can be run like google. Linux can not meet all the compiuting needs of eenterprises as of today. Solaris can probably never meet the needs of the embeded market or even comodity desltops it would be foolish to claim it can. That is just how it is. Linux tries to do everything while it does reasonably on all of them it doesn’t do one thing fully well.
Take 2.6 for example it was manily a few enterprise features that went in desktop features are on hold till the next release. I rember reading robert loves and linus’ interviews on this. Linux can probably satisfy the needs of many markets but meeting all the needs of any given market it doesn’t do.
The motto always is the right tool for the right job.
No Google needs serious backsup. You can read the Google File System paper, just search on google. They basically rely on replication for higher level of reliability. The basic idea is redundancy. You have to replicate stuff to make sure that one or two single node failures will not cause loss of data. Also remember that, the data itself is really not too much. The text is around 4TB I believe.
I’m somewhat amazed at the weath of misinformation and FUD that’s being posted here wrt. to Linux, Java, Solaris, and Open Source.
I think it should be clear to Sun that Java does not need to be held so close to their vest. The original premise of using Java as a point of leverage to sell servers has failed. I would offer that the R&D and development costs on Java are a money loser for Sun at this point, and they are clinging to control of JAva as a means of keeping Sun relevant.
If Sun wants to cut costs they should open Java up and let the community start hacking on it.
Nothing in this statement even approximates the truth. Java is free to use and develop with, however, Sun keeps the source closed so that it will maintain control of Java. Java is a tremendous asset to Sun and suggesting that they should give it away is rediculous. IBM desperately wants to leverage control of Java away from Sun and opening Java’s source would certainly give it much more ability to do so. Microsoft also stands to benefit a lot from a fragmented Java market and would likely love nothing more than to see the same happen. Keeping the source closed is about maintaining control of a valuable asset. Suggesting that Java no longer provides revenue for Sun is equally rediculous. Sun is a company, not a charity. If Java really didn’t provide any important value to Sun then I’m sure that they’d “open Java up and let the community start hacking on it.”
Java was never a tool to leverage the sales of servers. Take a look at the places where the largest adoptions of Java have occurred. I think you’ll find this has little to do with selling more boxes. The whole point of Java was to provide a way for developers to write programs once and then run them on any platform. That’s not about server lock-in, that’s about giving customers the freedom to take investments in programs they’ve written and move them to any platform they choose.
Times are changing, and I think SUN could cut tremendous cost by investing in Linux instead of Solaris.
By cut tremendous cost, you mean lose tremendous revenue and alienate their customer install base, I assume? Sun has invested in Linux and has also made a commitment to deliver all of the software they port to x86 to both Linux and Solaris x86. Companies like Sun don’t just drop Solaris for Linux. Consider that they have a huge number of customers who expect a commitment to Solaris for the long term. And, frankly, with all of the money Sun has invested in Solaris it would be stupid to discard it. Consider that Linux is just now trying to address problems that Solaris has been solving for the last 5-7 years. Companies like RedHat and Google are able to run mission critical Linux because they put an *enormous* investment into modifying Linux to suit their needs. For most companies, that’s not feasibile.
Linux as a technology is a lot more flexible to work with from a vendor perspective. Why for instance is SUN wasting much resources porting Solaris to X86 when they could easily package a customized linux solution for their customers?
This is another statment that simply isn’t true. Nothing about Linux inherently lends itself to be any more flexible to vendors. In fact, Linux, from a vendor perspective, is really a difficult problem. Most ISVs certify on one, maybe two Linux platforms: RedHat and maybe something else. So, how does this imply any additonal flexibility from a vendor perspective. Customers will still have to buy RedHat, which, I should point out, is now more expensive than the JES at the enterprise level. Go figure.
As far as porting to x86 is concerned, Sun has had a x86 Solaris port since 2.7, the effort has already been spent. Now they’re trying to capitalize upon that. Which is very different than doing this from scratch. Further, they’re already in the process of creating Linux solutions. You should really take some time to find out what these companies do before you make blanket statements that simply aren’t true.
I got capped at 8,000 characters and had to split this up.
When the industry was young and naive, business have been known to spend up to $25 000 just to run a website. SUN still thinks clients like those exist today. I hope they have enough thirty year long contracts to milk to survive.
You clearly have no idea what you’re talking about. If you change the above statment to read IBM, though, that would make a lot of sense. Sun has been leading the industry lately in lowering cost and complexity for their software packages. I recently talked with some friends who run an IBM shop and they told me that IBM charges them upwards of $12,000/month to support a 2-way server on which they run their website. That is rediculous. Sun’s support costs, are nowhere near that, nor do they force you to sign a GS agreement. You could buy the box with the software for less than $5k and probably spend a lot less on support and software licensing.
If SUN wants to survive they have market trends to look at and follow. But Solaris will only get them so far.
So what you’re saying is that every Systems company should do exactly what the others are doing? That doesn’t make any sense. Sun has a different view about where Linux belongs in its product line. I hardly think imitating every other IT company would be a good idea. You clearly don’t understand any part of their systems strategy. You don’t seem to understand that Sun may see the market trends differently and believe that they have an alternate strategy which can also provide value and flexibility for their customers.
Also, you claim that this argument isn’t about OSes and then make outrageous statments in favor of Linux and against Solaris. You’re just as guilty as anyone you accuse of being “fanboyish”. Fanboy.
If a technology does not give the user control over the tool, than the technology is weak and is of poor quality.
Er, I had to extract this quote from the attached anti-whatever rant it was imbedded within.
Java does give the user control of the tool. You write a program in source code, compile it to byte-code, and then run it in a VM on the platform of your choice. What runs, if it runs, is exactly what you wrote. So, the user very much has control of the tool and the behavior of the resulting program. The above statement is nonsense. And it’s author’s ability to reason is weak and of poor quality.
Those of us who know Sun and how they’ve treated Linux in the press do chuckle at Sun these days. The constant Linux bashing they used to due in the past just makes their recent attempts to embrace Linux all the more desperate. Its too late and nobody wants to see Sun succeed with Linux.
Er, okay. Sure. So since Sun didn’t adopt Linux when all the cool systems comapanies were, it will never be as cool as IBM or HP? This is an illogical argument. Sun is the first company to put forth an actual Linux desktop solution. IBM isn’t, HP isn’t, RedHat just killed theirs off. You may not want to see Sun succeed with Linux but I’m afraid it’s completely out of your control. Fundamentally adoption of Linux by IBM or HP hasn’t changed anything. Customers are still locked into the platforms that IBM and HP sold them, it has just turned into a sweet deal for IBM and HP as it’s a great new vehicle to shove additional hardware, software, and services down the throats of their customers. Linux doesn’t fundamentally eliminate vendor lock-in, at least not the way the markets are behaving right now. You’d do better to outgrow this childish attitude that because Sun scorned Linux at one point, now they’re no good for it. I mean, get a grip. Sun has stated that on x86 Solaris and Linux will be equals. So, they’re going to start offering systems based solutions for both. If you don’t like it, fine, but consider that Sun has provided a lot more flexibility (and will continue to do so) for their customers than the other pushers of Linux (like IBM, for example).
“Of course it’s their friend. So many people that work for NULL for increasing the value in their moneybox. I would call all free volunteers ‘friends’ too as long they work without any salery.”
All open-source developers who are not part of one of these corporate open-source houses (i.e.: IBM, Sun, etc) should start getting use to the idea of working for free while watching these corporate open-source players take your code for free and make money off of it. Unfortunately, most people don’t have the luxury of getting a nice full time job writing open-source software. I wonder if one day, the majority of the patches going into gnome, kde konqueror, open office, etc. won’t be from the major corporation such as Sun, Apple, IBM. Hmmmm. I wonder if there is anyone without an ibm.com email address that has submitted a major patch to eclipse. I wonder, I wonder, I wonder.
>> Java was never a tool to leverage the sales of servers
Come on, McNealy himself talked up Java selling Sun boxes.
Remember the purported Java ASICs? Sun boxes would run Java apps the fastest and you would flock to Sun to get this performance. Sun would license said ASICs and shareholders would rejoice.
>> And, frankly, with all of the money Sun has invested in Solaris it would be stupid to discard it
No, but you can end-of-life it. There are dozens of platforms, hardware and software, that are technically sound but no longer commercially viable, and they go into maintainence contract mode. Take Alpha for example.
This tells your customers that you are not dead-ending them. The customers know when a platform is no longer viable even if the vendor doesn’t. Play stubborn on the platform issue too long and all of a sudden you are marketing IRIX or BeOS (two perfectly capable yet not viable products).
>> Companies like RedHat and Google are able to run mission critical Linux because they put an *enormous* investment into modifying Linux to suit their needs.
Well of course RedHat has made an enormous investment in linux.
As to Google, they run a modified kernel, but you don’t know what work they put into it so stop speculating. I am sure the net labor is less than the cost of Solaris x86 licenses for each box.
>> So what you’re saying is that every Systems company should do exactly what the others are doing? That doesn’t make any sense.
Yes it does. Sun doesn’t have the market presence to make a proprietary platform viable. If they did, they would not be where they are today. Face it, the industry is past Sparc/Solaris for anything but the upper most tier.
>> Java does give the user control of the tool. You write a program in source code, compile it to byte-code, and then run it in a VM on the platform of your choice. What runs, if it runs, is exactly what you wrote. So, the user very much has control of the tool and the behavior of the resulting program. The above statement is nonsense. And it’s author’s ability to reason is weak and of poor quality.
But the same thing can be said of any closed source program, including Windows. Its a trust issue. Do I trust Sun not to inject malware into the VM? Should I have to? The only reason I presume this has not happened is that the JVM has no become the de facto platform they wished it had. Otherwise I presume we would in fact be going through a JVM toll booth.
As for the previous author’s “Ability to reason” – yours reads like a press release.
The university I am at is replacing all our Unix labs(Mixture of IRIX HPUX Solaris) with Linux. Ask them why? “Better performance, much cheaper” is what will echo in your ears. Try telling them “but Linux doesn’t crashdump”.
Or they listened to an idiot like you instead of looking at either clustering Solaris x86 Opteron servers or an HP Itanium running HP-UX.
Oh, I’m sorry, that would be actually using ones commonsense, my god, aren’t the Linux crowd stupid today. 9 years ago, people like root would not have even come within 100kilometre radius of the Linux community, they would have seen a text based installer and kealed over at the sight.
Ever since the media jumped on the Linux hype we’ve had 5minute teenie bopping linux fanboys like root who couldn’t code his way out of a paper bag, even if he was given the algorith and all he had to do is transpose it into C or C++.
As for “Linux will take them places”, the *ONLY* problem with Solaris RIGHT NOW is the crap hardware support, apart from that, as long as you stick to the status quo, mainstream hardware, it will be supported. Solaris is more reliable, cheaper and performs better than Linux under a heavy load. It is about time that there is MORE to Solaris than just SPARC.
How childish? So the fact that my IP says, clarion.edu means that I am a student of Clarion University and I work at Clarion?
Like I said before, you are very silly. Oh, I’m not of a student of Clarion University neither do I work their. So what’s your next hypothesis?
Oh, I’m sorry, that would be actually using ones commonsense, my god, aren’t the Linux crowd stupid today. 9 years ago, people like root would not have even come within 100kilometre radius of the Linux community, they would have seen a text based installer and kealed over at the sight.
9 years ago, I started programming for fun on Sparc machines. What where you doing then?
Ever since the media jumped on the Linux hype we’ve had 5minute teenie bopping linux fanboys like root who couldn’t code his way out of a paper bag, even if he was given the algorith and all he had to do is transpose it into C or C++.
As a previous Unix zealot, I would never had touched a Linux box with a million foot pole. Well that was until I had hardcore proof showing underpowered x86 Linux machines doing exactly what $30000 Unix machines where doing without so much as a sweat. That year was 1999. I came back home that day and installed debian. It took me weeks to get it working right.
As for programming, I’m a Java and C zealot. I hate C++ and I’ll never partake in large Java project. I’ve done once, I’ll never do it again. Java is a nice language, but it’s a step backward in computer linguistics.
If your idea of a diligent discussion is personal attacks on my character then I’ll continue to make you look like an idiot.
I’m not sure I understand the many messages that imply Solaris is (currently, or soon-to-be) a dead Unix. How much of the Unix marketshare belongs to Solaris? How many copies of Solaris are being shipped every year compared to the other Unices? What is more prevalent in the server market, Solaris or Linux? What suits Sun’s hardware better, Solaris or Linux?
It just seems odd that people have already relegated Solaris to the back alley with IRIX and the rest. Reading some of the messages here you’d think nobody’s buying Solaris servers, and everyone is rapidly abandoning their Solaris machines for Linux. Is that reality? I’d like to see some hard numbers if anyone has them.
As to Google, they run a modified kernel, but you don’t know what work they put into it so stop speculating. I am sure the net labor is less than the cost of Solaris x86 licenses for each box.
From where does this speculation come? If the x86 boxes were purchased from Sun, the total license cost would be $0.
John
The line somehow remindes on of Gollum: “Master is our friend.” And yet it was half-hearted.
The price/performance has nothing to do with SUN and everything to do with x86. As I said, again, what has stopped you from using Solaris on x86? you scream Linux yet you’re ignorant of Solaris x86. What does that tell me? it tells me you have NO experience in IT what so ever, or otherwise you would have addressed the issue in your post.
Face the facts, you’re a student at Clarion, not a master guru, not an all seeing oracle of osnews.com. Face the facts, read a book and grow up.
Remember the purported Java ASICs? Sun boxes would run Java apps the fastest and you would flock to Sun to get this performance. Sun would license said ASICs and shareholders would rejoice.
Sure sure. Java, when it was a new technology, got overblown because people had unrealistic expectations about what its capabilities, strengths, and weaknesses were. Of course, back in reality it was realized that maybe you shouldn’t use Java for everything and that Java wasn’t the one true whatever. Frankly, the hype-backlash cycle sounds suspiciously like what’s happening with Linux now. Perhaps some people have become overly excited about Linux and aren’t being realistic about the challenges it faces, and what commercialization means for the Linux market? It seems quite plausibile to me that at some time not too far in the future people will make similar comments about Linux. Linux, just like Java, or any other technology for that matter, is no more of a cure-all than anything else.
As to Google, they run a modified kernel, but you don’t know what work they put into it so stop speculating. I am sure the net labor is less than the cost of Solaris x86 licenses for each box.
Don’t I? I’d be very careful about inferring what it is a do an do not know. If you do any work with Google, or know any of their systems people, it should be clear to you that they have a significant investment in getting Linux to do the things they need to for them. I’m not talking about one or two people, either. They have a huge IT and software investment, in Linux and do a real amount of work to patch the kernel to serve their needs.
As for Solaris x86, you get the license for free when you buy the box. So, no, I doubt the net cost of hiring a staff of Kernel Programmers is cheaper than buying a $2k x86 box, and that assumes you don’t get a volume discount.
Sun doesn’t have the market presence to make a proprietary platform viable. If they did, they would not be where they are today. Face it, the industry is past Sparc/Solaris for anything but the upper most tier.
Their platform is just as viable as Linux. It’s cheaper to get a license for JDS, which is Sun’s Linux, and JES, their entire Solaris OE and Middleware stack, than it is to buy a seat of RedHat Enterprise Linux. If you don’t believe me, go check it out. Sun has done more to revolutionize the way software is priced and systems are built and sold recently, than any other company. And I wasn’t even talking about SPARC. Sun is about to release Solaris for x86-64. AMD’s Opteron is faster than anything Intel has right now. So, even if you don’t want systems built on “proprietary” hardware, you can still get systems with Solaris or Linux on affordable commodity hardware that can be configured for any solution you need. I fail to see why you think Sun should give up the game when they already have, and continue to create viable affordable platforms for their customers.
And, since you decided to drag SPARC into the mix, SPARC is less proprietary than any chip released by Intel or AMD. There’s an open consortium that decides what the SPARC standrd will be. There is no such thing for x86 or x86-64.
As for the previous author’s “Ability to reason” – yours reads like a press release.
I’m glad you decided to actually move away from the facts and cut straight to the name-calling. That’s very classy.
The price/performance has nothing to do with SUN and everything to do with x86. As I said, again, what has stopped you from using Solaris on x86? you scream Linux yet you’re ignorant of Solaris x86. What does that tell me? it tells me you have NO experience in IT what so ever, or otherwise you would have addressed the issue in your post.
Face the facts, you’re a student at Clarion, not a master guru, not an all seeing oracle of osnews.com. Face the facts, read a book and grow up.
Your ignorance amuses me. Have you tried running Solaris on x86? No, thank you, I’ll pass. I’d rather run Solaris on a Sparc. Heck, I’d use Win3.X on x86 over Solaris on the x86. What a horrible experience.
By profession, I’m a consultant for a large software company that has clients such as Penn State University, and Rutgers University. I happen to work on projects for both Universities via my employer.
As for the clarion.edu IP, ever heard of public proxy servers? You are a really naive individual just from your comments. To base judgements on my character, experience, profession and age based on an IP address that I can change, manipulate and falsify, via IP agents, web cache proxies, anonymizers, to mention but a few, is the height of ignorance.
So much for your depth of knowledge about the internet, talk less of Solaris, SUN or Linux. How do you expect us to have an intelligent when you don’t know that I could use IP subnets for the University of China while I reside in Australia to post a comment on osnews?
No really, you couldn’t be more silly. Now tell me, who needs to grow up and read some books.
Your ignorance amuses me. Have you tried running Solaris on x86? No, thank you, I’ll pass. I’d rather run Solaris on a Sparc. Heck, I’d use Win3.X on x86 over Solaris on the x86. What a horrible experience.
I run Solaris SPARC, Solaris x86, and Linux x86 (RedHat 7.3), and use the machines on a daily basis. I don’t find Solaris x86 and Solaris SPARC to be substantively different, aside from the types of machines on which they run. I think your bias is unfounded, if you’ll run Solaris SPARC but not Solaris x86. I’ve run Solaris x86 just as easily as Linux. The install and setup were certainly no more difficult. If you try to get fancy and use hardware not on the HCL you might have issues, but the same goes for Linux. You can get pre-FCS Solaris 10 through Sun’s Solaris Express program. It’s a pretty neat way to get your hands on new features as Sun puts them into their builds, and it’s not hard to get a free license if you’re using it for personal purposes.
The link is at:
http://wwws.sun.com/software/solaris/solaris-express/sol_index.html
Well I sure hope Rutgers University-NB isn’t the 45,000-student university migrating to Linux that you mentioned earlier. When I was there (5 years ago) Solaris was running the show, and their systems were absolutely rock solid. That’s where I first found Sun/Solaris and have been impressed ever since.
Since graduating and moving into “the real world”, I’ve been rather unimpressed with the reliability of many of the server solutions out there. (Of course some of that downtime is due to the quality of the IT staff.)
John
As for the clarion.edu IP, ever heard of public proxy servers? You are a really naive individual just from your comments. To base judgements on my character, experience, profession and age based on an IP address that I can change, manipulate and falsify, via IP agents, web cache proxies, anonymizers, to mention but a few, is the height of ignorance.
Why would you go to so much trouble to hide your IP on a public forum unless you were afraid that your trolling would get you banned? Why would you go through so much trouble to hide yourself if you are the confident professional who has insight into what the future holds for Sun and linux as you claim?
I am aware of all the technologies you mentioned. But we have been giving you the benefit of doubt. One can usually find out the experience of the person by they way they interact with others, by what they say and the way they say it.
But if you need to hide behind IP anonymizers I don’t know what to think of you, your comments and nor do I care. One more thing when you first get involved in linux, the root account holds a strange fasination and you gloat in the power you have being your own box’s superuser. Your fanboyish attitude, you style of responding to comments and the choice of your alias/name or handle, speaks loads on your psychological make up and exudes University student with first intro to linux.
That is my theory, as in all of science, theories are formed and broken. you have yet to give me a counter argument to break my theory. You are just going to call me names and dismiss this as being silly, so be it. If you have nothing useful to contribute don’t. This is my last post adressing you on any discussion here on osnews unless you say something intelligent. Good bye
Well said. That my fellow posters is a post that makes much more sense and rationale than most of the posts that go on this forum.
Again, I still see people jumping on the open-source bandwagon without really understanding the issues of open-source and closed-source.
THERE IS A PLACE FOR BOTH. Having only one does not allow us as a community to achieve higher goals.
Let’s face it, these language comparisons are virtually useless and Eric needs to wake up and smell the reality coffee. Open-source is not the cure-all of everything. Someone earlier asked me why comparing perl and python to Java from the perspective of one language taking over another is like apples and oranges . . . think about it for a sec before you post from you bandwagon . . . many languages are suited for different tasks . . . why would you want to limit yourself by trying to replace one language’s different abilities for another langugage’s different abilities. On a given day, I use the world of scripting for easy text manipulation (awk, shell scripting, etc. and sometimes even sed) while I can create quick prototypes and finished bioinformatics research software via Java (and when JDBC comes into play its great), and there are most certainly days where parallel computing comes into play with lots of C code and tons of C APIs. Eric’s statement regarding perl and python is asinine.
I hope to God you are not a system/network administrator. Because if you think the only reason people use the technologies I mentioned are solely for trolling, then you have a lot to learn in the world of network security and internet privacy.
So lets begin with your unfounded “theories”, as you put it.
1 Because I have .clarion.edu subnet, I am a student of Clarion University and I work there.
2 Based on that, I have no experience with Solaris and I am talking off my buttocks
3 My alias suggest I have a superiority complex as such my comments are henceforth null and void.
4 I make use of certain web technologies, hence I am a troll.
5 Conclusively, Solaris is better than Linux.
I would leave you to reflect on your comments once again.
>> Or they listened to an idiot like you instead of looking at either clustering Solaris x86 Opteron servers or an HP Itanium running HP-UX.
Itanium??? Haha. Why not just invest in PA-RISC or MIPS? Lets get all the dead architectures in one room!
My experience with Solaris with on x86 left me with sore impressions. Three years ago, my ordeal began when Solaris 9(I think) failed to detect most of my hardware (no fancy hardware–A Dell machine). I remember having to swap my ati card for an older matrox card, and also swap my network card for a better supported card amongst several gruelsome adjustments.
When I finally got Solaris 9 to boot, its performance was horrid. I mean really terrible. In fact, I find it hard to believe that your experience with Solaris on the sparc is very similar to that on the x86. My experience has been wildly the opposite. Solaris is awfully slow on the x86 arch and I can swear all the people I know who have used Solaris on but archictecture corroborate that.
While I never had an original copy of Solaris 10 for personal use, my experience with Solaris 10 on the x86 or sparc is no different. Worse on x86, better on sparc. I”ve also used HP-UX machines which are equally slow.
Now slow does not necessarily mean bad or horrible. But for workstations, performance and responsiveness is essential. Many of the Medical and Engineering labs at the university I have mentioned have conducted extensive testing since mid 2000 comparing Linux against the Unix they use in their labs. Mind you, this is long before I even joined the project. Based on their findings, Linux performed better and was as stable as the Unix they. I wasn”t a part of the testing process or project so I don”t know what the stress tests entailed.
Migration started slowly even before I joined the project. And it does not seem to me that they are regreting the decision. In contrast, now they employing consultants to work out strategies on a large scale migration. Okay, these are mainly for workstations in computing intensive laboratories.
Lets go to reasearch projects. All the supercomputing projects and cluster projects I”ve seen are based on Linux, period. I haven”t seen one, repeat, one, Unix box, be it Solaris, HP-UX, AIX, you name it in any of the projects in both Universities. As to why this is so, I don”t know. These where areas where I felt the big rigs would have dominated. Frankly, this finding shocked me.
Lets go to administration. Prior to my arrival to the project, multi-way sparc machines and mainframes(with an operating system I”m not familiar with, though I think it”s from IBM) where used predominantly as database, productivity, data backup servers. According to the admins they would have never considered Linux as a replacement had it not been for the success of aforementioned research projects carried out by the admins in the academic departments (the clustering and supercomputing projects I mentioned above that incorporate Linux).
Lets go to internet and networking. This area are primarily dominated by the BSDs, Linux and Windows. Only on few occasions have I seen some sort of Unix and that was SCO which was used as a network server in some accounting department somewhere.
Now this is not headlines news, hype or FUD or whatever you wish to call it. This my experience in the environment I work in. And anyone who says Universities don”t manipulate a large amounts of data, is just plain naive and inexposed. I”d wager that many Universities manage more data than Business Corporations, and perform more computing intensive processes that most Corporations.
I do agree Solaris is more mature. But that”s besides the point. Take a look at the experience I narrated above. The areas in computing that I was made to believe were once dominated by Unix have been largely replaced. I”m sorry to say, but Solaris is suffering the same fate. I also know of a sister project working with OSU (Ohio State Universtity) to replace their Unix labs with Linux powered ones.
Now base on my perspective, is it just me or Solaris” target audience and market shrinking, despite the fact that it might be a flawless OS? Now take a look at the areas Linux is divesting into. Database services, Backup services, networking services, Computing intensive workstation, embedded devices, multiple patforms, desktop. As opposed to Solaris” high-end completely niche computing intensive environments. I think if SUN invested wisely in Linux it would generate wider sources of revenue in long run as opposed to Solaris. Not only is it cheaper to develop for, (because the cost of developing Linux is being shared by thousands of people around the world while SUN solely bears the burden of Solaris) it is easier to customize solutions for clients than Solaris is. Yes, not all customers want to buy hardware and software. Some want both, some want just services, some want software only, some want software but on their own hardware etc.
That is why I argued that SUN should invest in Linux because it will provide them greater flexibility and most likely new sources of revenue. Holding on to a shrinking and niche market is not good business practice, at least for SUN which used to be dominant in the highend market. No product lasts forever. Eventually products will have to be replaced at some point or another. With SUN extensive experience with Solaris they could customize a powerful Linux system that scales well up 512 processors like SGI did.
Now based on my experience, do you suggest SUN hold Solaris ever dearly to their chest? Didn”t I state several times in the beginning this is not a “Solaris vs Linux” slug fest. But SUN”s market scope with Solaris keeps shrinking while other Unix vendors are doing their best to expand their client, their scope and their market. Heck, SCO is even divesting into Unix legal affairs to generate revenue. Yes, that”s a terrible example but businesses need to survive.
If further comments are a personal attack on me, then I”m afraid I”d refrain from responding henceforth.
Some links about high end, enteprises and Linux.
“SGI Altix (Linux) Achieves World Record Memory Bandwidth of 1 Terabyte per Second on Stream Triad Benchmark”
http://www.sgi.com/newsroom/press_releases/2003/november/benchmark….
“Backed by Linux, Amazon Thinks Big”
http://www.eweek.com/print_article/0,3048,a=117985,00.asp
“How Linux saved Amazon millions”
http://news.com.com/2100-1001_3-275155.html
“Top Ten TPC-C by Performance Version 5 Results As of 15-Feb-2004”
http://tpc.org/tpcc/results/tpcc_perf_results.asp
“Top Ten TPC-H by Performance Version 2 Results As of 15-Feb-2004”
http://tpc.org/tpch/results/tpch_perf_results.asp
I’m sorry to hear that you’ve had unpleasant experiences with Solaris x86. Mine have been nothing short of positive. I’ve got a Solaris Express build running quite happily on a Dell Precision 650 and I have a dual Opteron running Solaris x86 as well. I had to find drivers for its network-card on the internet but that was the extent of the effort required for me to get my configurations to install properly. I’ve also got Sun Blade 2000 running an Express SPARC build. I’ve had very few issues, even with pre-FCS software.
My objection to your statements is not necessarily with the conclusions you reached based upon your personal experience, but that you’re letting your understanding of Sun’s strategy 1-2 years ago color your judgement about their current strategy.
I disagree that Solaris x86 is slower than Solaris SPARC. I’ve done a large amount of benchmarking of the two platforms and have to tell you that, at least with Solaris 10, x86 on fast hardware is faster than SPARC. It’s hard to make a 1.2gHz UltraSparc-IIIi+ keep up with a 3.06 gHz Pentium 4. Especially when you consider all of the other hardware factors. I’d release my numbers, however, they’re priviate information of the company I work for, so I’m really not allowed to say much more than I already have. I wish I could, though.
I’m glad the University you were consulting for was able to find solutions to their problems with Linux, but I do beleive a lot has changed since 2000 when you said you started this investegation. What Linux vendor did you choose for your platform, and what hardware did this University decide to go with? I ask, since that does affect a lot of the cost that one of your clients might pay. My understanding of Sun’s software pricing model is that it is very competitive with Enterprise Linux solutions.
Your statement that high-end hardware, produced by many large systems vendors, hasn’t kept pace with the performance of commodity hardware/chips produced by AMD & Intel isn’t that far from the mark. Some large high-end systems do offer advantages to the commodity systems, but it’s easy to appreciate how they can become obsolete quickly if major systems producers don’t beef up their own chip offerings. (IBM, Motorola, TI, etc…)
However, I think you’re far from the mark to suggest that Sun’s strategy currently commits it to be stuck with high-end hardware that nobody will buy. Certainly that was their strategy 2 years ago, maybe even last year. But within this last year they’ve made a significant push in the low end. They’ve introduced x86 servers that are cheaper than Dell (so they claim — I haven’t investegated). They’ve put a significant investment in x86-64. They’ve entered into an agreement to acqurie Kealia systems:
http://www.sun.com/smi/Press/sunflash/2004-02/sunflash.20040210.8.h…
Now, what you might not know about Kealia systems, is that they were involved in building computer systems on the x86-64 platform that were on the size of the large boxes Sun was selling. It’s probably way to early to tell but I would suspect that Sun is probably looking to create a new line of high-end servers based on commodity parts. They never would’ve been able to get Bechtolsheim back on board otherwise, as he left because Sun was ignoring their potential cost savings from using commodity parts/hardware.
There have also been rumors of Sun & Fujitsu collaborating on the high-end. See:
http://www.theregister.co.uk/content/archive/33592.html
Given that Fujitsu’s SPARC chips are presently much faster than Sun’s, it has the potential to breathe new life into their high-end platforms until they can get around to building systems on commodity parts.
I think the argument that Sun doesn’t take its markets seriously and is content to let itself shrink away to irrelevance simply isn’t true. You also haven’t acknowledged that Sun has made a commitment to Linux. They’ve got the first commercial Linux desktop solution out there. You can get their servers with Linux or Solaris, and they’ve begun an effort to get all of their software to work on both Linux and Solaris. If that’s not an investment, I don’t know what is.
I don’t think it’s correct to suggest that inherently you get more flexibilty from Linux than you do from Solaris, nor that Sun has been stuck in a neiche. Flexibility isn’t inhernetly about the Operating System you run, it’s about the platform you build and the stack of software solutions you implement as a part of it. Obviously, the operating system is a part of that, but there are many other parts. Along with Sun’s push at the low-end, they’ve gone ahead and created a bunch of customer-ready solutions, and reference-architectures which are setup to be quickly configured for specific tasks that customers in certain business spaces would need. They’ve got information about this on their website:
http://www.sun.com/products/architectures-platforms/refarch/index.h…
This sortof lays out their strategy for quickly generating solutions for customers. I’ve never purchased anything from their Reference Archietctures, but I have a friend who was involved in helping them get these assembled. There was a lot of input from experts in various fields of business and it goes a long way for cutting up-front costs and complexity, as you have less configuration to do out of the box. I’m not sure this is necessarily true for Linux. Especially if they’ve hired you as a consultant to help them customize their operations.
I disagree that Sun’s market is shrinking, and think that they’re doing everything they can to open up other market opportunties for themselves. They’ve made a Linux investment, and are also maintaining their Solaris investment. I think giving up Solaris would be a mistake, especially with all of the new features they’ve been adding to it in 10. Many of the other U*IX vendors became obsolete because they put their unicies on life support and made the incorrect decisions about what hardware and software to support. You mentioned SGI which is a great example. They completely blew it, but not just by holding on to IRIX & MIPS. But they also turned themselves into just another systems company with expensive stuff. I doubt Sun would make the same mistake. Sun has invested in Linux to generate wider sources of income, but you don’t seem to be aware of this. Investing in Linux doesn’t preclude maintaining a good investment in their existing platforms which are still relevant. And that, sir, has been my point all along.
I would be surprised if Rutgers University-New Brunswick has abandonded — or has plans to abandon — Solaris for its main operations. Rutgers seems heavily tied to Sun… they probably get educational discounts.
http://uptime.netcraft.com/up/graph/?host=www.rutgers.edu
http://uptime.netcraft.com/up/graph/?host=www.eden.rutgers.edu
http://uptime.netcraft.com/up/graph/?host=www.rci.rutgers.edu
http://uptime.netcraft.com/up/graph/?host=www.rucs.rutgers.edu
http://uptime.netcraft.com/up/graph?site=remus.rutgers.edu
For more information… http://rucs.rutgers.edu/
I wouldn’t be surprised if they have dabbled with Linux systems in their non-critical areas.
Has their HPCF abandoned this: http://www.caip.rutgers.edu/~e10k/ ????
John
I little correction is in place. Firstly, I joined the project I’m working for mid last year. The migration process had started before I joined the project. I need to emphasize, I am not a part of the migration project. I’m a IT auditor and consultant that needs to work with several teams including the migration project team.
Extensive testing is still being carried out as we speak to validate the migration process. As to what enterprise solution they’d be using. I’d say it depends. For critical data, most of the migrations have been Enterprise solutions from Red Hat and SUSE on Dell and HP hardware. And recently there was talk of purchasing some high end PPC machines with Linux preinstalled from IBM.
For non-critical stuff, where the admins believe they will not need support from a vendor both Universities are taking bold steps to provide support for their own customized version of Debian and Gentoo. And many of the formely Red Hat Linux machines will be upgraded to Fedora.
So the migration is a mixture of commercial enterprise Linux with support and services for critical stuff and customized non-commercial Linux for tasks adminstrator are confident they can provide support for. I have noticed that much of the clustered Linux machines, many of which are academic oriented, have mostly been customized non-commercial Linux solutions. I’m not sure if that’s because Red Hat or SUSE don’t have good support for Linux clusters, or if the academic admins felt they could provide support and maintainance themselves.
Many of the old Unix hardwares that are no longer supported will be migrated to Linux and will be maintained and supported by the admins or sold. In order words, in places where Linux can easily be replaced with Unix, at little to no cost, it’s a no brainer it will be replaced over time. And believe me, you’d be shocked to see the places where Linux is replacing Unix.(network services, workstation, desktop)
In the highend computing area, IBM and HP provide to some powerful Linux solutions(hardware and software) with special licensing from vendors like Oracle. These solutions are usually server arrays and racks, and this is the area where I think Linux is killing Unix, frankly. All the “Linux doesn’t scale well” comments begins and ends on osnews and slashdot.
In the real world, Linux can be arranged creatively to tripple sometimes quadriple the peformance of mainframes from SUN and IBM, and for cheaper. You won’t believe me until you see it yourself. Dell also has some very powerful server array and cluster solutions they demonstrated to one of the Universities. This is the area I think SUN should step up in and provide powerful cluster/server array solutions especially for Linux if customers need them.
Forget JDS/OpenOffice.org/GNOME. Leave that for the small boys. The thing is SUNs attitude leave many, myself included, that SUN is threatened and bitter by all this. Because, I honestly doubt business will be upgrading to 128 way mainframes as opposed to server arrays/racks and clustered machines.
You see more often than not, geeks fail to realize that developing and touting innovation for intellectual satisfaction and or pride do not translate into business strategies or solutions. Yes, solaris is powerful. It has amazing technologies and capabilities. But that’s not what really matters.
Businesses want to call SUN, submit to then what their IT staff/consultant have designed for them, and all SUN should do is execute it. SUN on the other wants to tell businesses, Solaris does this better so use that instead, Solaris has that feature use that instead,etc. That pisses customers off. Use Solaris for the server, Linux for your PDA, etc statements makes clients look like morons. What if I don’t want Solaris. What if my consultants said I should use Win2k3 instead, what if they said Linux on Sparc will suit my needs.
If you take a look closely, that’s exactly what HP-UX and IBM are doing. Doing exactly what their customers want even when it doesn’t make sense. SUN until recently shoves decisions down your throat, instead of listening to what people want. If tommorrow SUN would wake and say, you know what, we will take care of your migration from Solaris to Linux and vice versa, the money I and several people are making today will all be directed to SUN’s revenue.
As to SUNs comittment to Linux, my original post was modeded down where I acknowledged suns OpenOffice.org contribution to Linux. SUN also contributes to the GNOME project, but from the comments coming out of SUN’s mouth for the past few months or so, SUN seems to be unclear as to its linux strategy. SUN is like a ship with no destination with regards to Linux.
I’d go as far as saying Linux is just a contigency plan in case all else fails. In reality this might not be so, but that is the sentiment myself and many others get. In fact, they pitch Linux and against Solaris as if they are competitive products, instead of treating them like complimentary solutions. “Solaris for the Server, Linux for the desktop” mantra, only tells clients SUN is divided in its comittment. Changing that mantra the following blue moon doesn’t help them either.
Shouldn’t they be telling customers we provide both Solaris and Linux solutions for both the desktop and server? I mean that sounds more coherent than “Our solution of Solaris on the server is stronger that our solution of Linux and vice versa for the desktop”. “Well, perhaps IBM’s solution on both is better” is what many clients will think. Hence you loose customers and public confidence.
Back to the Solaris performance on x86 and sparc. On the dell machine I purchase some years back, Linux ran a lot smoother on it than did Solaris. On the sparc box however, Solaris ran slighter better with deamon tasks, while Linux seemed very responsive for basic desktop related tasks. This was tested on the earliest version of Linux-2.4.* or even the later versions of Linux-2.2.* and Solaris 9.
In college I played a lot with the sparch machines and grew to love them. There are lovely; they are nice, but compared to the Linux boxes I had built at home, they were slow. Like you said, I’m sure many of this have much to do with disparate hardware and architecture. But ever since, the impression of Solaris being slow has always stuck, and after my ordeal with it on the x86, I was never inspired to give later versions a try on that architecture.
I’ll however try to procure and evaluation copy of Solaris Express for the x86, and give it a spin one more time. I’m sure things most have changed. I love SUN, like I said earlier my earliest Unix adventures where on Solaris 8 or 9, can’t remember. But their actions of late have made my interest in the PPC arena a more tempting exploration. SUN needs to be more flexible and more aggressive. And for God sakes, let them begin to leverage OSS to their advantage like the wiser Unix vendors are doing.
Yes, Rutgers will be migrating many of it’s Unix facilities over to Linux. In fact, they are the University I mentioned have been conducting tests with Linux for the past few years. I seriously doubt many of their next upgrades will be Unix’.
Rutgers has spent millions of dollars on high end computing intensive network servers many of which could be replaced with cheaper x86/Linux solutions. In brief, any critical services that is too costly to be migrate will not be touched.
Migration from Unix doesn’t mean an entity trashes their Unix machines and installs Linux over them. Migration is usually done almost one system or unit at a time. Migration occurs usually when a contract service has expired or time for upgrade (software/hardware) approaches. So many of the Rutgers services might still be Unix based, but I’m positive their Science/Engineering labs are carrying forward with some migration toward linux workstations.
Yes, I am aware the Rutgers Computing Services, they hired us. With regards to the Rutgers High Performance Computing Facility project, I think that spurred some of the decisions to look into Linux as I heard the project cost a fortune. However some obstacles cropped up, one of which where very specialized software applications that were not available on Linux and for which compilers on Linux did were not reliable(certified) among other contrainst I can’t remember.
So the University is thinking of similar projects but based on Linux clusters instead as opposed to Unix. I’ll try to see if I can get a friend on the migration team to post more elaborately on the topic here and what may or may not be migrated.
Will all the Unix services be migrated? Very unlikely. Are they considering Linux services over Unix services? Yes. Will they be updating/upgrading there Unix services to Unix if Linux is cheaper? Not likely. Again, everything is not going to happen over night. And a lot of the migration explorations are still in the testing phase.
SUN needs to be more flexible and more aggressive. And for God sakes, let them begin to leverage OSS to their advantage like the wiser Unix vendors are doing.
This is already underway at Sun. That’s been my point. They may not be doing this the way you think they should, but they are leveraging OSS.
In any case, our discussion has gotten pretty off-topic. If you want to take this offline, I’d be happy to continue the discussion. The e-mail address I’ve got attached is real, so drop me a line if you wanna talk about this further.