Slashdot Mirror

Solaris 10 has something very similar to launchd called SMF: Service Management Framework. It doesn't replace cron but it does replace init.d and inetd. It also provides backwards compatibility so that existing init.d scripts and run and existing inetd entries are migrated to SMF on upgrade.

SMF also does dependancy checks and auto matic restart on failure (or some other conditions). It also uses XML for its configuration but imports that into an SQlite database so that it doesn't need to reparse the XML on every service restart or system boot.

For more information on SMF in Solaris see the links in the main architect/developers blog .

Sun has done the same thing with a feature called SMF in Solaris 10.

And, as to why anyone would switch, well of course the answer is that fortress may be, for some things, more expressive or faster. It's a few years out, but I suspect folks in my research community (fluid-flow simulation of ocean and atmosphere) might take a long and hard look at this. The big projects (e.g. weather modelling at national labs) might benefit from the maintainability aspects.

But the big win will be small projects, e.g. individual work of graduate students. As a prof who sometimes has to work with grad students on their code, I look forward to a more conventional mathematical notation and to a less error-prone language than we have today. Other nuances also appeal, related to scoping, inheritance and so on. I also like the idea of a unit library (which I assume will incur no runtime cost) to catch silly mistakes of adding apples and oranges.

I suspect that the technologies being considered will lead to faster code, partly because it can discover good things at run time as well as compile time. Providing good (read "natural") language support for parallelism will be a welcome feature; in my community we use MPI and that adds a whole ugly layer that is of no interest to the scientist at the head of the team, so a sort of curtain develops between the coder, who worries about such things, and the scientist who prefers to think in more abstract terms. (Here again, the ability to code in a mathematical way will be very welcome, indeed.)

We all have a long time to think about this proposed new language. There are commercial as well as technical concerns. As a sun investor, I may as well just say it: my guess is that Sun may vanish before this project is completed in 5 to 10 years. l sure hope the developers can keep up this work. (Licensing, a favourite /. topic, is not irrelevant here.)

I encourage folks to read the docs and listen to the talk. The Java:C issue is a red herring.

Replying to myself (how sad?) - the system restarter stuff sounds like the new process contracts stuff in Solaris. man ctrun on a Solaris 10 box.

Nice to see Apple take a page out of Sun's book...

Sun came up with (at first glance) a similar thing called Service Management Facility in Solaris 10.

Wasn't this paper co-authored with Joe Darcy .. now the Java floating point czar working on Tiger (Java 1.5)?

Much shorter version of the paper is here, and a good java floating point paper is also over here

oh .. and if you think that nobody at sun will admit java's weaknesses .. you gotta stop talking to the sales and marketing drones, and spend some time in targeted discussions with the engineers ..

This is basically Legato Prestoserve (introduced around 1989) packaged with a disk drive. Prestoserve used battery-backed NVRAM instead of flash. It was typically installed on servers, and its main purpose was not power conservation but rather accelerating an NFS server by committing writes to stable storage without waiting for the hard drive. But the idea is similar: a persistent disk cache.

Gimmick? Hmm...

This is just electronic writes. Those who have worked with a high-performance SAN like the Hitachi 9900, Sun 6920, know that electronic writes is where all the performance comes from. When our SAN's 4gig of cache goes offline, my DBAs come running and everyone complains about terrible write/read speeds.

Electronic writes (in a good amount) means that the data flies into memory and later on the disk system pumps the data out to the disk platters. Netapp is really great at doing this kind of operation. Talk to someone with a Netapp and they'll tell you that you can watch all the activity lights on an array light up in a fury of activity with the Netapp head pushes from cache->disk. It's really fantastic. Of course, it's really expensive too.

I'm not saying that Samsung's disks are going to be enterprise-class. But I'm hoping that this cache thing catches on to make SATA2 (or whatever) based arrays feasible for desktops. Imagine a chain of large-cache drives on desktops, large-cache single drives on laptops ... For example, if you had 128MB of cache per drive, then an array of 8 disks would have 1gig of cache. That would be so great for the power-user. The Sun T4 bricks cost $50k and they have 2gb of cache per controller. 2gb of electronic writes/read = amazing performance.

I can't wait to learn more about the quality and speed of the disk cache. It's what I've always wanted to see in consumer grade disks.

This is a clone of the SunRay solution (see here). With SunRay's you use a smartcard to access, which has the cool feature, that you can pull your card from the thin client, walk over to another (to discuss something with a colleague, to give a preso in the boardroom, to go home even!!) and plug your card into another SunRay, where your session continues as if nothing had happened.

I have been Skyping for a while now, my wife uses it instead of those cheap international dialing numbers; it works (which is pretty incredible if you have any insight into how). I make free and cheap calls from the US to Europe and Australia; I call my home for free from a remote hotel room. The only issue I have is if Skype get to monopoloy status and then start to milk the market but I think the market is still way too immature for that. IMO - rather Skype lead the market than AT&T.

http://blogs.sun.com/roller/page/sharps

Both version numbers "1.5.0" and "5.0" are used to identify this release of the Java 2 Platform Standard Edition. Version "5.0" is the product version, while "1.5.0" is the developer version. The number "5.0" is used to better reflect the level of maturity, stability, scalability and security of the J2SE.

The number "5.0" was arrived at by dropping the leading "1." from "1.5.0". Where you might have expected to see 1.5.0, it is now 5.0 (and where it was 1.5, it is now 5).

For the rest of the discussion, see http://java.sun.com/j2se/1.5.0/docs/relnotes/versi on-5.0.html

Of course, last year at JavaOne, the Sun presenters tended to just roll their eyes when about the naming convention.

It looks like they've been making a lot of progress. I found their advanced prototype documentation here.

Sun has already taken this track on Solaris 10 with their Service Management Facility. See SUN's quick summary of SMF

I am still getting used to SMF, and luckily they also keep backwards compatability with the rc.d init scripts.

Provided SMF isn't under 100 patents, maybe LINUX can pickup a few of Sun's good ideas once OpenSolaris comes out later this year.

Would you buy a car with a locked hood that the dealer refused to give you the combination for?

I like open source software and Linux just for the simple reason that if it breaks I know I can find and fix the problem given enough time. I've gone thru the code on lots of projects, from Pine to ArgoUML to JBoss. Do I always have the time? No, but someone else might have. Most open source projects are fairly good about taking bug reports, applying fixes you've written, or even giving you cvs access. This is my only complaint with Sun's open source Java implementation. Yeah, it's open source but it doesn't really matter if I fix their code for them, I still have to go and vote on which bugs I want fixed( top 25 Java bugs by vote rank).

Do I care if my Mom uses FireFox? Only from a support(spyware/popup tech call) perspective. To many people treat open source as a religion/political aggenda. I believe the best approach is to treat it like a club/group of people sharing ideas and working together to fix problems.

Or maybe he just uses indent(1).

I am not an MS lover by any means, but I have seen a few shots of the "Glass" UI that look pretty interesting. If you actually peruse the site, you will see that the new UI isn't scheduled for release unitl late 2005 (Beta 2).

Some of those features look lame as all get out: the scrolling option on the Start menu for one.

I can't stand the Windows UI, so I skin it and call it a day...

I think the truth is that UIs are getting stale and that the 3d desktop is going to change things (like Java project looking glass: http://java.sun.com/developer/technicalArticles/J2 SE/Desktop/lookingglass/

I will quote in its entirety a post I put on a http://forums.hexus.net/showthread.php?t=23425 Hexus forum a few months ago:

Our world is such that we commonly work with numbers between positive and negative 2 billion. Everything from monetary amounts to the number of records in a database, from distance between places to weights and quantities - we generally work with numbers well within the 32-bit range.

When 32-bit processing came around we had a general and common need to process numbers bigger than what could be handled by 16-bit CPU registers, e.g. numbers bigger than 65,000. But this need for large number processing have stayed the same over time, and so it will not assist to drive the development of 64-bit CPUs!

There is also the ability of the CPU to do more accurate floating point mathematics. This, as well as the need to work with very big numbers, play a role only in engineering and science applications, and to a smaller degree in games.

Therefore the need for 64-bit processing is driven more by the need for addressing more memory than by the need for faster processing of very big or very small numbers.

We need 64-bit processing where the data width inherent in the problem exceeds the (32-bit) processor's registers' width. (Actually this is true for database memory requirements and for games' number crunching and engineering and scientific applications too ... these applications gain from 64-bit processing either because the problem lends itself to a very large data set, eg a large memory requirement or else because of a requirement to process numbers which are wider than the CPU's register width.)

It is not generally possible to recompile or even rewrite an existing problem to "require" bigger registers or memory space. However if a problem already requires big numbers to be processed and had been "optimised" to fit into 32-bit world, then the program can be (un/re)"optimised" to utilise the full 64-bit processing capability by removing these initial optimisations, such as where 64-bit operations have been broken into multiple 32-bit operations.

In fact, someone (Adrian Cockroft) very aptly said http://www.sun.com/sun-on-net/itworld/UIR951101per f.html 64-bit CPUs increase application performance despite the 64-bit nature of these CPUs . 64-bit instructions and, in particular, 64-bit memory address pointers imposes a big additional load on memory, caching and the CPU, so if you're not using those extra bits, compiling to 64-bit actually makes the application execute slower!

To test this, take your favourite compiler and compile your favourite utility program to both 32 and 64-bit executables and run both and compare the timings on your trusted Althlon64 or Sun ULTRA 5 workstation. :-)

Unless the program either processes lots of large numbers or utilise more than 4GB ram, the 32-bit version will run faster.

A program which does not process huge numbers and which does not process numbers bigger than 32-bits will run faster when compiled to a 32-bit executable, even on a 64-bit CPU. There is also the bigger 64-bit executable to load and instructions to move between memory and CPU.

Let me add something to this - as a pojnt in case, all the general purpose utils on Solaris 7, 8, 9 and 10 come as 32-bit executables by default (Some 64-bit utils are available, but not in your path by default). This is probably because the memory bandwidth overhead (read: Wasted memory bandwidth) due to 64-bit executables needing to transfer double the amount of bits from memory to find out where pointers point, even if it is just to point to next next memory address! (Eg pointers are bigger because they can address more memory, even if you don't need it)

A very simple comparisson will prove this, eg
timex

Mate, which rock have you been hiding under?

I run Solaris Express on my noname-brand laptop amd64 with ati radeon 9600) without any problems at all. I've run Solaris 9 and Solaris Express on my athlon laptop since 2003. Perhaps you should check out the info at http://www.sun.com/bigadmin/hcl to get up to date.

Already available in beta from
http://java.sun.com/j2se/1.5.0/download.jsp
Link at the bottom of the page.

Sun Microsystems has been shipping fully 64-bit chips and operating systems for about a decade now.

Sun press release from 1995 about 64-bit chips and operating systems.

Sun should have made Java an open specification like, err, EVERY OTHER FRIGGING LANGUAGE EVER MADE, instead of fighting idiotic lawsuits with MS (who were in the right for a change)

Java is an open specification. It's documented right here: http://java.sun.com/docs/books/jls/. The JVM and the Class Library are also openly specified. There is absolutely nothing stopping anyone from implementing their own Java system.

However, Sun owns the name Java. And if you want to use it for your system, you have to pass Sun's tests. Fair Enough, Linux is a trademarked name and I'm pretty sure there'd be at least a couple lawsuits if Microsoft started selling Windows under the name Linux XP.

Microsoft created their own JVM & Class Library (fair enough), and then added a bunch of non-standard additions while calling it a standard JVM & Class Library. That was against the rules governing the name Java.

Sun sued Microsoft because they broke the rules, and in doing so, broke the promise that everything with the name Java will be fully compatible. That promise is why Sun holds on to the name Java, and is pretty much the same reason Torvalds owns the Linux trademark.

I used a Sun E3500 with 8 processors and 8GB of ram to run a GwebCache. It didn't really perform as well as I had hoped, so I moved it back to it's normal cluster of 3 intel machines.

SunRay software web page.

It looks like it's currently Solaris-SPARC and Linux-x86 as the supported configurations (yes, Slashdot, Sun _does_ sell Linux thin clients).

IMAP's closer than you think. If you don't think SQL is relational, you certainly won't think IMAP is, but you can do more than most MUAs support. You can save arbitrary tags on email messages. You can execute surprisingly-sophisticated queries. I recently wrote some crude Jython scripts that use the JavaMail API to do queries like this:

to_me = OrTerm([HeaderTerm(header, "slamb@slamb.org")
for header in ["To","Cc","Bcc"]])
to_list = OrTerm([HeaderTerm(header, "")
for header in ["List-Post","List-Id","List-Archive"]])
msgs = sourceFolder.search(AndTerm(NotTerm(utils.to_me), utils.to_list))

(sorry about the indentation; ecode apparently doesn't like it.)

You have to read RFC 2060 to know all that IMAP can do.

new JFormattedTextField(NumberFormat.getIntegerInstanc e());

Check out Java Directory Server 5.2 (a.k.a. Netscape/iPlanet/SUN One DS).

Are you providing leadership in distributed computing?

Check out Solaris Containers in Solaris 10 (formerly called N1 Grid Containers).

I ask Sun, where are you innovating now?

I'd say that dtrace has to count. Maybe you should read up on Solaris 10 instead of spreading FUD?

Check out Java Directory Server 5.2 (a.k.a. Netscape/iPlanet/SUN One DS).

Are you providing leadership in distributed computing?

Check out Solaris Containers in Solaris 10 (formerly called N1 Grid Containers).

I ask Sun, where are you innovating now?

I'd say that dtrace has to count. Maybe you should read up on Solaris 10 instead of spreading FUD?

Check out Java Directory Server 5.2 (a.k.a. Netscape/iPlanet/SUN One DS).

Are you providing leadership in distributed computing?

Check out Solaris Containers in Solaris 10 (formerly called N1 Grid Containers).

I ask Sun, where are you innovating now?

I'd say that dtrace has to count. Maybe you should read up on Solaris 10 instead of spreading FUD?

Check out Java Directory Server 5.2 (a.k.a. Netscape/iPlanet/SUN One DS).

Are you providing leadership in distributed computing?

Check out Solaris Containers in Solaris 10 (formerly called N1 Grid Containers).

I ask Sun, where are you innovating now?

I'd say that dtrace has to count. Maybe you should read up on Solaris 10 instead of spreading FUD?

Wow, if only there were a moderation "Doesn't know what the hell he's talking about"...

1) The ability to allocate more memory space to Java apps.

From the command line, run "java -X". Note the -Xmx and -Xms options to allow you to control initial and maximum heap space. These options have been around for years.

2) 64 bit support.

Ever heard of the SPARC processor? You know, the one Sun produces, one of the first mass-market 64-bit platforms available?

Assuming you're talking about AMD64, there's already a JVM that supports it. Linux-only at the moment, but if you insist on using platforms that suck, there're already AMD64 Windows JRE release candidates up too.

3) Ability to/Easier implementation of hardware specific calls to speed calculations.

Now you're just missing the point entirely. One of the major goals of Java is that the same compiled binary can run on any machine with a compliant JVM written for it. When you start making calls to specific hardware, you lose that. Besides, if you *do* want to sacrifice the platform portability of your program, you can always just write a JNI module for it, but then I suppose you don't have a clue about that one either...

Wow, if only there were a moderation "Doesn't know what the hell he's talking about"...

1) The ability to allocate more memory space to Java apps.

From the command line, run "java -X". Note the -Xmx and -Xms options to allow you to control initial and maximum heap space. These options have been around for years.

2) 64 bit support.

Ever heard of the SPARC processor? You know, the one Sun produces, one of the first mass-market 64-bit platforms available?

Assuming you're talking about AMD64, there's already a JVM that supports it. Linux-only at the moment, but if you insist on using platforms that suck, there're already AMD64 Windows JRE release candidates up too.

3) Ability to/Easier implementation of hardware specific calls to speed calculations.

Now you're just missing the point entirely. One of the major goals of Java is that the same compiled binary can run on any machine with a compliant JVM written for it. When you start making calls to specific hardware, you lose that. Besides, if you *do* want to sacrifice the platform portability of your program, you can always just write a JNI module for it, but then I suppose you don't have a clue about that one either...

How about Star Office, It is free for educational use and may not have some of the bugs or problems with Open Office. I have never used Open Office so i cannot tel you how much better/worse it is that OOo. Supports Macros too and comes with some sort of database app too

Exactly. So why is Sun dragging its feet on porting its virtual machine for the Java language to the AMD64 platform?

What are you talking about?!? Go to http://java.sun.com and click on "J2SE 5.0". Then click on "JDK 5.0 Update 2" and accept the license agreement. Right there are four packages for AMD64. (Two for Linux, two for Solaris.)

Stop spreading FUD.

What makes StarOffice so much better than OOo? The only difference that I was aware of was a builtin database app for StarOffice.

Sun's web site is of no help in answering this question; the only specific statement they offer on the differences is, "For a detailed comparison, please see TBD."

Also, there is no longer a 'secure' Solaris version, which was typically used by the US government. Solaris 10 is (apparently) secure enough 'out of the box' to be natively deployed in the CIA, NSA, etc...

No, Trusted Solaris 10 is coming.

Sun will never able to withdraw any OpenSolaris code which is released, read the CDDL.

Once it's out there, it's out there for good. Sun will not have any specific right to terminate OpenSolaris licences.

I'd like to say that running Solaris 10 on a homegrown AMD XP 1800+ with a VIA chipset is a great advance for the OS. Of course you still need a specific NIC (3Com), but it has made advances in its compatibilty with X86 hardware. Take a look at this http://www.sun.com/bigadmin/hcl/ if you are interested in running 10 on a X86 machine. It's still pretty rigourous and you'll more than likely have more success running it on a Dell/HP workstation but 10 has opened the doors for better X86 hardware support.

Why the heck don't you just download it from the official site at Sun like everybody else ? Like at
http://www.sun.com/software/solaris/get.jsp. Sun's whole Web site is plastered with "Get Solaris 10 now" banners and links. So where's your problem ? Don't have a Web browser or what ?

You know, in Windows, you can hit Ctl+Alt+Del to bring up the task manager, go to the processes tab, right click on a process, and go to "Set Affinity".

The "auxiliary documentation" that is mentioned at the article being kept up-to-date and reflecting code changes is now implemented as javadoc and its competing technologies like Doxygen or phpDocumentor.

I have to admit, XP doesn't crash very often compared to previous MS releases. I have had it crash/lockup though due to applications and the only thing to do is power it down. I can't recall seeing a BSOD with it yet. On the other hand, my Solaris systems for the last 7 years and the DEC Ultrix systems for 3 years prior to that have had better reliability with zero OS freezes/crashes. I have had applications hang, but they haven't hung the OS in the least.

Microsoft has to overcome years poor stability reputation among IT professionals. Linux will likely face the same issue. The major difference I see though between the MS and Linux systems is that when I pay for the OS, I'd expect better quality, just because I'm paying for it. I'm probably a bit more forgiving to the Linux system crashing than a Microsoft OS.

As for the hardware comments, vendor support for the dominant desktop OS makes sense. Unless the hardware vendor sees a particular reason to release drivers simultaneously for multiple operating systems, I would expect MS to be the first driver released.

If you're looking for hardware to install into your PC, then how about showing me where I can get MS Win32 drivers for Sun GigaSwift PCI Ethernet Card. Likely this piece of hardware will only provide Solaris drivers. Note: I'm only showing this as an example to disprove your statement. Generally speaking though, you are correct about the hardware.

Your RCed wrong. The latest version (ultrasparc IV) is a 2 core processor (the two cores being the same as the ultrasparc III) and can run 2 threads (one per core) from http://www.sun.com/processors/. But they are talking about their next generation (which is not ultrasparc) called Rock and Niagara which are supposed to have more than two cores per processor. Some of what you are talking about like having "virtual registers" are already being implemented by techniques like register renaming where the registers visible to a program are less than the actual set of registers on the CPU. This is useful for things like function calls which involve saving the register states, but could also be useful for multiple threads on the same core.

Sun came out with a PCi card that allowed people to run Windows NT in an X-server application window, primarily for CAD users who wanted to read Windows E-mail.

And of course, there is Wine which allows Windows application to run on Linux.

So Microsoft are going to have to compete somehow.

1. A decent computer costs nowadays, what? A few hundred dollars? How much can you save on CPU and RAM anyway? A decent CPU is under 200$. (You don't need a 3.6 GHz P4EE on everyone's desk.) So you're gonna save... what? Maybe 100$ for the whole machine?

With costs in the range of several tens of thousands per person per year, that kind of saving is a spit in the bucket. It's just not worth the loss of productivity and the learning curve.

Even assuming that all programs ran exactly as fast over a network (and they don't), and the server had enough computing power to not get stuffed when 100 people do some CPU-intensive batch processing at the same time (e.g., before the big meeting on Friday), etc, it's still a losing proposition. You only need 1-2 server crashes, or hard drive getting full, or whatnot, to turn that "profit" into a loss.

2. Precisely because salaries are high and operating costs are high, the way to go is to increase productivity, not to handicap everyone with piss-poor cheap tools. It's not even something IT speciffic.

E.g. if you have a construction company, the way to go is to buy a bulldozer and a crane, not to give everyone shovels and buckets. Yes, shovels and buckets are cheap, redundant (you can have everyone have one), reliable (no moving parts for a start), bug free (the design was tested for millenia), etc. It's still a bloody stupid business plan.

3. Is it even a win anyway? Let's say you ran 100 terminals off a mainframe. You saved maybe, what, 10,000$ by buying thin clients instead of computers?

Now let's say you connect them all to a small-ish 8 CPU Sun server, say, the Sun Fire V890. Let's take the 8-way 32 GB RAM option: $123,995.00

Net _loss_ there: over $100,000. You also want to make it _redundant_? Shall we take two of those? Net loss: almost $250,000.

And that's already a piss-poor solution, since 100 users actually running CPU and graphics intensive software on that, will make the machine crawl. I.e., you invested $250,000 into... lowering productivity. How bloody stupid is that?

I.e., please... I can see how snake oil vendors like Sun would love to convince you to pay them $250,000 for a piss-poor big-iron solution, instead of paying $100,000 to Dell for some good PCs. But is it actually in your company's interest to pay more for less? Definitely not.

Assuming you are referring to my blog entry discussing the way that the success, not failure, of MPL-style licenses is at the root of the license proliferation problem, I'm afraid I don't agree with you. LGPL does not include the explicit patent grant that the MPL includes, nor does it establish ground rules for maintaining a patent peace, and thus does not serve as the archetype for MPL licenses and their (many) derivatives. Instead it includes an exception to the scope of the GPL which depends on the language and architecture of the software in use and makes assumptions on how dynamic linking will take place. In all other respects it is the same as my third license category, GPL licenses.

Just some facts:

Sun's revenue Q3: $2.625 billion

Net loss for the third quarter of fiscal 2005 on a GAAP basis was $9 million

The cash and marketable debt securities balance at the end of the quarter was $7.357 billion.

http://www.sun.com/aboutsun/investor/earnings_rele ases/pr/2005-q3.html

So, your remark should be modded Troll....

I just read the definition of the definition of the CDDL. Outside of the fact that it's not compatible with the GPL (which can always be solved by dual-licensing), I find it reasonable. Of course, I may have missed some details, so I'd be happy if someone could point out what the problem is (outside of "it's from Sun" or "it's another license"). Also, I see that they distinguish between "Initial Developer" and "Contributor", but I don't see any difference between the rights given to one and the other. Can anyone enlighten me on this too?

Not anymore, it doesn't. SunRay server software is now available for Linux, as well. So you can run a *cheap* SunRay lab. Get some SunRays off eBay, buy the server software (it's kinda spendy, but cheaper than the Sun hardware), and run a couple of dozen SunRays off a single server.

They are really nice machines. Fanless. And their software is getting very capable. You can even mount USB pen drives off the back of them.