Slashdot Mirror

Check out the guy's presentation:
Jonathan Schwartz presentation

Page 23:
All software will move to one distribution, and three licensing models - Traditional, Predictable and Metered

So comparing what Sun plans to what Microsoft has already done is rubbish.

Check it out

Ok, after the "bsd is dead" partyline slashdot announces the "sun is dead" partyline. The whole purpose of this fire blade platform is that sun wants to bypass the "more MHZ every week" stupidity. Do you want a linux server? take an x86 blade.you want ssl? take an ssl processor blade. And mind you, this is not 10 pc's all plugged into the same bus, there is virtualization here, meaning you have a centralized point of management from which you can assign jobs and control everything. This kind of management is unlike the "login to X and do Y" paradigm , is more like having Vmware with hardware acceleration provided by the blades. With the N1 architecture it is my opinion that everything becomes a managed resource so you can do with it anything you want. So they make it a question of scalability which (IMHO) is what sun is best able to do better than (almost) anyone.

Of course there's always Sun's online store for those who can decide what they want without hand-holding.

I think that one thing people fail to realize when they are saying "Sun should convert to Linux" is that the argument for Solaris vs. Linux isn't so cut and dried.

Let's just say Linux worked out all of the scalability and SMP performance issues and was able to run on Sun's Big Iron (TM) just as fast, or even faster than Solaris. This would be great for Linux, however, I still don't think Sun would adopt Linux on a wide scale for their larger SMP boxes.

Why, you might ask? Binary Compatibility. The number one reason why a lot of enterprise customers use Solaris for their homegrown applications is that they can be confident that if they develop an application on Solaris 8 right now, when Solaris 10 comes out next year it will still run without even needing to be recompiled.

This is a huge plus for enterprise customers that spend millions developing custom in-house applications and don't want to have to worry about the FSF "breaking" glibc on every release. They also don't want to worry if Linus Torvalds' latest kernel will break their app.

Linux has been able to make a lot of headway in a short amount of time, and I'm frankly amazed at how quickly the functionality has surpassed Solaris and commercial Unices, however, that rapid growth and feature bloat comes at a price... Backwards compatibility... I would be seriously amazed if you could take an app compiled with gcc on a Redhat system from 1997 and run it on a Redhat 8.0 system from today without recompiling it.

Anyway, here is a link to Sun's Binary Compatibility Promise.

My biggest complaint with Linux is that the desktops lack the refinement of CDE - CDE was designed, KDE and Gnome evolved



Guess what Sun is using for their interface now?

GNOME 2.0 Desktop for the Solaris[tm] Operating Environment

Ain't that a bitch?

Many different machines

A Sun, an IBM Mainframe and even Ozzy and Sharon

Pah, this mod is just rubbish. I rule.

The above was a paid for advertisement on behalf of the "I can do it better association".

Also, 64 bit chips are not ready for the mainstream market yet. They need more testing, analyzation, and intrepration with an unbiased group of persons as the sample, to make reliable predictions.

Yeah, no one makes 64 bit chips that are ready to use.

> Sun last year bowed to market realities and accepted general-purpose Intel-compatible computers into its server line. Its first model, the LX50, uses Intel processors.

What about their Cobalt line of RaQ servers. Seems to me they "bowed to market realities" when they acquired Cobalt Systems back in 2000. The LX50 is just a Cobalt RaQ with a faster processor.

I think Cobalt servers make great low-end web servers, and they even run Sun's Brand of Linux (as does the LX50), which I believe is based on RedHat 7.2 (which they also acquired from Cobalt Systems).

> Sun last year bowed to market realities and accepted general-purpose Intel-compatible computers into its server line. Its first model, the LX50, uses Intel processors.

What about their Cobalt line of RaQ servers. Seems to me they "bowed to market realities" when they acquired Cobalt Systems back in 2000. The LX50 is just a Cobalt RaQ with a faster processor.

I think Cobalt servers make great low-end web servers, and they even run Sun's Brand of Linux (as does the LX50), which I believe is based on RedHat 7.2 (which they also acquired from Cobalt Systems).

> Sun last year bowed to market realities and accepted general-purpose Intel-compatible computers into its server line. Its first model, the LX50, uses Intel processors.

What about their Cobalt line of RaQ servers. Seems to me they "bowed to market realities" when they acquired Cobalt Systems back in 2000. The LX50 is just a Cobalt RaQ with a faster processor.

I think Cobalt servers make great low-end web servers, and they even run Sun's Brand of Linux (as does the LX50), which I believe is based on RedHat 7.2 (which they also acquired from Cobalt Systems).

On the low end Sun servers cost $250,000 (unless you cut a deal)

You might want to qualify that, since Sun sells lots of low-end servers for under $5k. If you count "going to their website" as "cutting a deal," then, maybe. :-)

I've worked with a bunch of el-cheap SunFires (prices) that were pretty good servers.

You mean this Linux?

I just bought a sgi on ebay and love it.

Err lets see.... I'd use JNDI which does all the work for me. So in Java a LDAP replicator would take....

A couple of minutes (and yes I've written one). Because JNDI makes all that work totally trivial.

No CSV, no intermediate phases. Give me one ldap URL and I'll copy everything over into another.

NB. This proves nothing but the fact that you don't know what you are talking about.

Gee, way to show you don't know what you're talking about. Every compiled language turns the result into assembly language.

Java is not compiled into machine code, because there is no machine.

I've HEARD some java compilers actually can compile into ASM, but I haven't seen them, because everyone whinges and bitches about how this defeats java's imaginary "write once run everywhere".

Is it true that some companies are so overcome with script bias that they'd assign years of unnecessary work rather than give it to the coding untouchables?"

this toaster

burns the weather into your toast - done by java, maybe linux is in there somewhere though

For information on Open Firmware see the Open Firmware home page.

Also, info about Sun and Apple implementations.

OpenFirmware (IEEE 1275) has a homepage. As does the IEEE working group. There's also a DMOZ/Google category.

Where again is Sun's VM capability?

Here. Or at least, it will be.

'The technology will be integrated into the Windows code, sources said.' Will Microsoft be able to pull this one off? Will their virtual machines run operating systems other than Microsoft's?

This is most likely Microsoft's response to Solaris Containers which are expected to be shipping in Solaris 10. Of course, both of these are simply implementations of ideas pioneered by IBM with VM/CMS.

The VM approach makes a lot of sense even if you only plan to use it to run multiple copies of the native OS within them. The advantages are twofold. Firstly, it prevents one malfunctioning application from impacting other applications - even on Unix this is a serious problem, since one process can devour the CPU, memory, disk space, etc. Secondly, it allows resources to be redistributed or added on the fly, especially if your VM is seamless enough to span nodes.

alext, you're completely missing the point. Java users have been able to do this for quite some time now. Take a look a JAAS - it is an excellent solution if all you do is Java. But the purpose of describing the access control policies in something language-independant like XML is that, we'll, you can implement it in other languages without having to rewrite both the rules and the access control mechanism. This is darn good stuff actually, and lots of readers here are completely missing it. You and many others are blaming Sun for developing something that can be used for more than just Java! Incredulous...

I don't think so.

Can't you see Java is the default language for XML technology since its beginning?

When you think to create web services using all infrastrctute such as provided by Sun One , you think Java, not C, nor Perl.

It's necessary to have frameworks to build complex applications, unless you think its worth to start from scratch. In this case you can implement all the standards on your favourite language.

That url should read http://www.sun.com/servers/highend/sunfire15k/inde x.xml.

Java uses number of milliseconds, expressed as a "long", which ranges from -9223372036854775808 to 9223372036854775807 (reference). According to my maths, this means we're safe until around 292,269,049AD.

Can anyone remember when our sun is scheduled to go nova? Oh, and this of course assumes people use Java correctly - its still quite possible to write non-Y2k compliant code in Java!

I guess you were never subjected to this beast

Take note of:

Esc
Control
CAPS
tilde
backslash
backspace
Help
The piece of plastic between "Help" and F1

Absolutely aweful.

Still, a working knowledge of forth can make certain maintenance operations (booting a different kernel, examining the device tree) a whole lot easier when you're having boot problems.

Mac's have been OpenFirmware based since 1994 or so (first PCI PowerMacs).

Oh, and there's an OpenFirmware song, too: sung by Mitch Bradley

The SPARC is aging and Sun has no resources for a replacement.

Where do you get that Sun has "no resources" for maintaining SPARC?

http://www.sun.com/processors/ranch/

Sun has 1,400 processor design engineers, the second largest add CPU design team in the world, using 2,000 systems consisting of 9,000 CPUs running 24x7.

Clearly, either Sun is very serious about SPARC, or Sun is completely insane.

The other fact is Itanium is not going to become the only 64-bit CPU. IBM is serious with POWER, Sun is serious with SPARC, AMD is serious with Hammer, and I even hear SGI is serious about MIPS again.

My preferred keyboard has things like cut, copy, paste, home, end, undo, help. I find it quite useful.

You can develop for Siemens J2ME phones without using any Siemens software, as long as you stick to standard MIDP 1.0. Just use Sun's J2ME Wireless Toolkit and then test on your phone.

That said, it is possible to get free SDKs from Siemens if you can manage your way through their horrible Web site. Start at www.siemens-mobile.com, then choose "Partner Program", then "Developers Village", then "Mobile Phones". The SDKs are then on the left.

This paper outlines a potential business strategy for Sun Microsystems Inc. that differs from the current strategy by limiting Java's release to Solaris only, open-sourcing Solaris and Sun ONE, and then allowing the cloning of Sun's SPARC architecture. Ideally, the market becomes suffused with inexpensive Sun compatible hardware working with Sun's free Java software (or licensed competitor offerings). To work, this strategy requires developers choose the J2EE platform instead of Microsoft .NET so then industry must use Sun or Sun compatible hardware without the choice of selecting Intel based hardware including those from IBM, HP or Dell, as Java is not available on those platforms. Sun's sustainable advantage comes from owning, implementing, and licensing the Sun SPARC and Java standards. Both of these standards become dependent on each other so that an implementation of each is required in any one system for it to operate correctly. The strategy is possible because Sun is the last vertically integrated computer company, making its own chips, circuit boards and software (Forbes, 2002) while also controlling the popular Java programming language. The strategy partly comes as a response to the speech McNealy on Strategy with the company affliction that it mentions. However, the window of opportunity for this strategy's execution is limited due to the growing enterprise functionality and corporate support of the open-source Linux operating system and the growth of the Microsoft .NET software platform. Not discussed in this paper is the additional possibility of Sun merging with Sony so that the author's GrooveTip entertainment idea may be executed. If you are happy with this strategy to save Sun then please sign the petition.

Another poster mentioned that oracle is pushing linux in a big way, so I assume that handles the software.

Here's the hardware: SGI's altix. Each linux OS image can scale to 64 processors and 512 Gigs of physical RAM. These images can then be clustered, to "span terabytes of global shared memory". Prices start at $70k for 4 CPUs and 36G RAM, and go to $1.2M for 64 CPUs.

Sun's hardware is about four times as expensive.

Okay, just eating the dogfood of the hypothetical argument, let's presume Sun goes off the radar for some reason, leaving Java to stand solely on its own merit.

What you end up holding are two things:
-The high level language. Not a bad effort, and rather well-known, even if facing competition from Ruby, Python, Perl, C#, and even C++. Everyone has their reasons to love/hate languages, but I'll argue that, on the broad scale, Java is faiiirly 'clean,'* and will always have a niche as a set of semantics for telling computers to do stuff.

-The JVM. As a crossplatform solution, it's suboptimal across "all" hardware, though various non-Sun implementations have found ways to eke out better performance. Can anyone name an architectural reason to keep faith in the JVM, and opt for processors 'built' for Java? This is an honest question; does implementing whatever those opcodes are advance computing in any way?

If there's nothing Insanely Great lurking under the covers, perhaps it's time for the crossplatform field to "advance" to solutions tuned on a lower level. As an Amiga nut, I've been following Tao's Elate/Intent products, the ill-fated open source Internet Virtual Machine (which is said to assume that All the World's an x86- oops!), and the .NET CLR. All abide by the same concept; build a lightweight, low-level virtualization that functions well on 'real' hardware; don't (completely) strap the VM to one language, whether in design or marketing.

---

So, what I'm trying to say is, Java's death could be *good* for the industry and design as a whole, if it forces us to replace it with something better. The CLR probably isn't particularly better, given the whole 'mark of the beast' aspect plus general bloat, but approaches like Tao's are compact, fast, and exist today. In that case, cutting the VM down to the bare minimum- theirs uses only a few k in binary form, allowing most of the environment itself to be kept in machine-independent code- actually lets you leverage the crossplatform aspect by appearing on new platforms quickly.

[Okay, as a bitter OS/2 -> FreeBSD user, I've gotten the impression that Sun's vision of 'everywhere' has narrowed to "Solaris, Windows, and just enough thrown bones to keep the Linux users happy."]

http://www.sun.com/hardware/serverappliances/

Sun left well enough alone and didn't change much about this product, including letting it run Linux and on an Intel platform.

This put them in a nice spot with previously invested companies like Rackspace and other hosting providers that survived the crunch.

Diversification is a successful investment strategy.

Click here for Solaris prices. It's free if you're a student, but the point of the Cringely comment was businesses. Example: 4-CPU Workgroup Server license $999.00

For example, your code isn't type safe. Check out
Substitutes for Missing C Constructs (About 1/3 down the page). The other tips are also quite useful.

Nokia and the others are just following the J2ME spec by Sun.

IPv6 information:
http://www.ipv6.org/

IPv6 for Windows:
http://www.microsoft.com/ipv6
http://research.microsoft.com/msripv6/

IPv6 for Linux:
http://www.bieringer.de/linux/IPv6/

IPv6 for Mac:
http://lists.apple.com/mailman/listinfo/ipv6

IPv6 for Java:
http://java.sun.com/j2se/1.4/docs/guide/net/ipv6_g uide/

I did not see the document you have pointed me to, yet. I have read another one, that I guess has been updated since then, that implied a total lifecycle of 3 years (from GA to end of maint. support). According to the document you pointed to, it's indeed an interval of 5 years, and the following text spells that out most clearly:

Red Hat Linux Advanced Server 2.1:
General Availability: May 17, 2002
Full Support (including hardware updates): May 17, 2002 -- November 30, 2004
Deployment Support: May 17, 2002 -- May 31, 2005
Maintenance Support: June 1, 2005 -- May 31, 2007


However, the Solaris lifecycle, in the same terms (general availability to end of maintenance support) is 10 years which is twice the joy.

I will admit that these terms look much more favourably on RHAS, though. Thanks for the link.

Comparing that to Solaris, I have no idea where you pulled out the 11.5 year life cycle. According to Sun's web page, it's 5 years from last ship date. Reference this page: http://wwws.sun.com/software/solaris/fcc/lifecycle .html

I will admit that 5 years from last ship is greater than 5 years from initial ship, but there's no way in hell it's an 8.5 year delta like you're trying to claim.

Ya know, "gobs of system management tools" and "a kernel many tricks up it's[sic] sleeve" don't exactly add to much of a review :-). I believe I can honestly claim that Red Hat Linux Advanced Server has "gobs of system management tools" and "a kernel with many tricks up its sleeve". Of course, this claim holds true for Windows too.

How you got moderated to 2 on your post is beyond me...

According to Java History on Sun's website, 1.0a didn't come out till 1995.

Glad you got a good deal on the hardware. Be careful though.

They have a whole section of the Solaris licensing pages dedicated to relicensing. Don't laugh. For some models, Solaris 9 relicensing fees are in the US$100,000's. Not sure if this link will work because the have some strange session-management junk on the pages with the pricing on them: store.sun.com/catalog

I have to agree. Every Slashdot post about Sun generates a fair number of "Sun is dying" messages. Yes, Linux on x86 is making some of their past offerings in the blade server market look a bit sad.

This is why they released their own Linux/x86 system. This is the second time Sun have released an x86 based server. This time it just might work.

Besides, Sun have a lot of diverse products - let's not forget that. Granted, the future of computing may not include some of these, but I think they are aware of that.

Actually they do sell Linux x86 servers.

Sun will not sit back and let Linux and Intel eat up their market. I also have a friend of a friend who works for Sun and is beta testing Sun's new intel workstation line. Appearently they are noticing companies like Pixar and boeing switching to dell lintel and wintel boxes. They plan to make both 3d as well as software engineering workstations that both will run Linux. Wait until this summer or next fall for the announcment.

Since their own distro is tuned for their own hardware it will be rock solid and stable. This is something thats traditionally an advantage to Unix over Linux. Corporations will love this as well as users.

the real question is if this will dissapear into obscurity as JINI has (a similar technology using Java).

We had to replace already the Lelon capacitors mentioned in that story in more than half a dozen SunRay1 clients, which had failed exactly in the way described (details and photo).

Nice to hear the newly emerging espionage-twist of that story ...

On top of it all, the GUI interface (I think they were running Gnome) was so much nicer than CDE on Solaris.

Solaris runs GNOME, by the way.

I looked for several of the bugs using Sun's bug database search engine, and couldn't find them. Constructing the URL myself, however, seemed to work.

For instance, bug# 4460368.

Well they had PRMan ported to Linux several years ago, and it certainly ran on SUN machines, even if that port was never released commercially. Sometimes Pixar takes a long time to release some stuff udes internally, like the notorious Deep Shadow maps, which they presented over two years ago and used in Monsters Inc. but just released it on PRMan 11 (probably after lots of customer complaining).

Anyway you can read this old press release from SUN (1997), specifically saying that SUN machines were used in the renderfarm:

PIXAR AND SUN MICROSYSTSTEMS CREATE MORE POWERFUL RENDERFARM FOR PIXAR ANIMATION STUDIOS

If I remember back then they had a suite of test that gave something like a RenderMark number or something like that and they said SUN got the highest marks. Some of their more recent SUN machines were 8 processor racks which also gave them bang in as little space as possible (the old buildings before they moved to Emmeryville were really tight).