Slashdot Mirror


Interview with Vita Nuova CEO Michael Jeffrey

Little-Fat-Sheep writes: "Lots of talk on Slashdot and elsewhere lately about the future of Operating Systems being massively distributed. Well, the technology exists for years now in the two operating systems offered by Vita Nuova: Plan9 and Inferno. OSNews features today an interview with Vita Nuova's CEO, Michael Jeffrey."

28 of 104 comments (clear)

  1. Read the article... by slugfro · · Score: 2, Informative

    It mentions Bell Labs several times including the relationship between bell labs and Vita Nuova (i.e. bell labs spun off Inferno to Vita Nuova).

    --

    -- Find the Truth...
  2. Fermentation... by sterno · · Score: 3, Insightful

    It amazes me how long it's taken for these ideas to ferment. I mean I was talking with people about the distributed OS concept back in 96 or so. I have to wonder why the concept has sat mostly unexplored for this long. Perhaps more importantly, I wonder why it's suddenly hot again. Is there some actual practical use for the technology that's bringing it back into the light? Or are people just thinking this is the next logical step of P2P and thinking that it will be hot because P2P is?

    --
    This sig has been temporarily disconnected or is no longer in service
    1. Re:Fermentation... by Anonymous Coward · · Score: 4, Insightful

      Wow, amazing. One would think that if you had, indeed, invented the concept of the Distributed OS back in 96 or so, you would know that Plan 9, a fully distributed OS had been around since the 80's. It's amazing how long it takes for a clue to ferment these days.

    2. Re:Fermentation... by _typo · · Score: 2
      I mean I was talking with people about the distributed OS concept back in 96 or so. I have to wonder why the concept has sat mostly unexplored for this long.

      Because of no real killer app has been found yet. For most distributed computing work PVM/MPI is all you need. Distributing the OS is a neat concept, but, in itself, gets you nothing. It's one of those concepts that was to be the One True Solution (tm) but was proven not essential. Sort of like microkernel OS's. When Linus started Linux Andy Tanenbaum declared it a monster. He argued that at that time starting work on a monolithic kernel was an idiotic idea. He was proven wrong. Same here, I guess. Unix was supposed to be obsolete, but it has served us well thus far, and probably will for alot longer. It's nice to know we have good alternatives to take off when Unix is truly dead.

      --

      Pedro Côrte-Real.

    3. Re:Fermentation... by Graymalkin · · Score: 2

      I don't think Linux becoming popular makes Andy Tenenbaum wrong about micro kernels. Linux got popular because it was free (always good for poor college students), ran on their 386 computers (yet again a shot in the arm from poor college students), and somebody got the bright idea of porting the GNU toolset onto it making it a functional cheap Unix-like OS that ran on cheap computers. I think had Minix been more libre licensing slashdot geeks would be chanting Andy's name rather than Linus's. Windows NT, MacOS X, and BeOS are good examples of how Andy hit the nail on the head talking kernel architecture. BeOS was technically cool but poor marketing and Microsoft and Apple heavy handing killed it. Kinda off topic I know.

      --
      I'm a loner Dottie, a Rebel.
    4. Re:Fermentation... by _typo · · Score: 2
      I don't think Linux becoming popular makes Andy Tenenbaum wrong about micro kernels.

      But it does. Tanenbaum predicted monolithic kernels were on their way out of the mainstream and that only micro-kernel architectures would survive. Linux is here to prove otherwise.

      I think had Minix been more libre licensing slashdot geeks would be chanting Andy's name rather than Linus's.

      Minix was a pile of crap. Period. It had no real multitasking, and it's cross-platform support was done by using a lowest-common-denominator aproach at the cost of performace. If all we had was Minix (no FreeBSD, no Linux), Windows would have taken over completely by now and I'd probably wouldn't be taking an Informatics Engineering course since I wouldn't have had the chance to play with an OS at such a low level when I was at the PFY stage. That's life. :)

      --

      Pedro Côrte-Real.

  3. Evolve together... by MosesJones · · Score: 2, Insightful


    Okay so Plan 9 is cool. Useful ? Probably not as it doesn't have any support or applications of note.

    Where as Linux is a poorer OS from a next gen perspective but has the applications and support.

    OS/390 is old school but has great memory management, io and SMP etc.

    The first two are already open source, the third owned by the Big Blue Linux supporter. Wouldn't it be better to have a directed 2 year plan to create a merged platform ? The reality is that Linux right now is in the Bazaar and to get to that end game we need some form of Cathedral project to guide and drive. But picture the end game, a networked OS, with loads of apps, the best SMP, io, memory and domaining support you can get.

    This would be the great killer platform for servers, and a kick-ass gaming platform.

    Unfortunately it won't happen because the only people who could really run this successfully would be a combination of Bell Labs and Thomas J Watson. Damn that would be kick-ass, but Big Blue don't seem to want to take the lead in Linux, and the linux community probably wouldn't let them anyway.

    --
    An Eye for an Eye will make the whole world blind - Gandhi
    1. Re:Evolve together... by baka_boy · · Score: 2

      I'm really getting tired of the argument that a single OS should be targetted at every imaginable application. That's exactly the kind of thinking that has turned Windows into the painful environment it is for any system administrator or programmer.

      For example, the attributes that make a great gaming platform (low latency, lots of multimedia device support, good graphics libraries, easy single-user setup and configuration) are not the same as those that define the server market (high reliability, clustering and RAID, easily automated and administered remotely). Why should Linux (or any single OS) be the "One True Way" for both of these, much less for any and every potential market out there?

      Personally, I'd rather boot into something like OpenBeOS (once it's ready, of course) for media work, switch to Linux when I'm doing network code development, and maybe leave a copy of OpenBSD around for the times I'm feeling paranoid.

    2. Re:Evolve together... by EvlG · · Score: 2

      You are quite right.

      Any OS tectbook or course will be full of lines like "but we can't say which of [x y z] is the best because it will depend on the application at hand."

      This is true for scheduling algorithms, distributed deadlock detection algorithms, etc...

      Something like an OS is very much intertwined with the problem you are solving. And that means that when you generalize, you lose some of the efficiency.

    3. Re:Evolve together... by DrSkwid · · Score: 2

      Okay so Plan 9 is cool. Useful ? Probably not as it doesn't have any support or applications of note.

      It does have users & applications.
      It's lacking in desktop software like word processors & spreadsheets & image editors but that does not mean it's not "useful".

      I have a plna9 box on my netwrok and I much prefer working on it than any of the others (FreeBSD & win2k).

      It is used by non-programmers in a few instiutions for text editing and email.

      The mailing list consists of about 50 active posters.

      It's a research OS and the programmers that use it are happy that it is a clean sheet implementation and that it doesn't carry a lot of the hair that some of the other OS's have collected on their voyage through userland.

      --
      There are places where the networks are not touching,and there are places where they are-Boeing's Lori Gunter
  4. A correction and thoughts by clump · · Score: 4, Interesting

    One little nit-pick is that the article mentions both Plan9 and Inferno are not Open Source. Also, its important not to look at the significance of these operating systems as in current market saturation, but what new and exciting features they can bring.

    Regarding the 'killer platform', im not sure that Holy Grail exists. However the world proves daily that implementation is more important than design, so just pick what works best for you.

  5. Michael Jeffrey sums it up in one sentence by evil_roy · · Score: 4, Informative

    " I believe Inferno achieved what Java set out to do. "

  6. Re:Open Source? by anothy · · Score: 5, Informative

    note the capitalization. Plan 9 is open source, but due to some traits in the license, it's not considered Open Source as per the requirements of the OSI. Inferno is open source except for a few core components, which are based on a subscription license model. i'm not sure if the license covering the non-core software is Open Source, although it is open source. the core software is clearly not.

    --

    i speak for myself and those who like what i say.
  7. Beg to differ by clump · · Score: 2
    From the article:
    15. What licensing covers Plan 9 and Inferno?

    Michael Jeffrey: Neither OS is Open-Source. However, we think that the licensing and pricing is attractive to individuals and corporate users alike.

    Because you can look at the code does not mean it passes the criteria of the OSD or the FSF's guidelines. Put some Plan9 code in your 'Hello World' app and im sure you will be hearing from someone...
  8. Re:Inferno Plugin (fractal program) - not a fracta by furiousgeorge · · Score: 5, Informative

    >>Eventually it gets all pixellated

    you're reaching the numeric precision of the hardware. Most fractal viewiers out there have this problem. They may do things in 64 bit math or 128 bit math or use their own custom routines, but eventually you zoom in so far the math falls apart.

  9. Re:A correction of the correction by dallen · · Score: 2, Interesting

    Correction: plan nine ISN'T open source according to Richard Stallman.

    - the plan nine license requires all changes be sent back to them;

    - you (possibly) can't sell your code for a profit;

    - lots of other problems with the license. (see here for Stallman's take on it.)

  10. Re:JavaOS and Inferno by anothy · · Score: 5, Informative

    Yes. Inferno applications are written in a concurrent programming language called Limbo. the language reference manual is available online, as are varous descriptions of programming in the language (and some other papers as well). the language is C-like in structure, with influences from many other places, like Pascal and Algol. of particular note are channels, a data type for inter-process communication which makes writing multi-threaded and/or distributed apps easier than in maybe any other system. it's a beautiful language.

    --

    i speak for myself and those who like what i say.
  11. Re:Plan 9 / Open source by anothy · · Score: 4, Informative

    Plan 9 is open source, but it is not Open Source. that is, it doesn't meet the criterion set forth by OSI for its license to get the "Open Source" mark. you can, however, get full source to Plan 9 for free and use it towards any end, commercial or otherwise.

    note also that the commentary you're linking to is commenting on an older version of the Plan 9 license; most (not all) of the issues have been addressed.

    --

    i speak for myself and those who like what i say.
  12. Re:What about bandwidth? by anothy · · Score: 2

    the protocol, Styx, is very efficient with its bandwidth; very little overhead (it's more latency sensative than bandwidth sensative). of course, if you're pumping mpeg movies around, there's not much to be done about that...

    any connection in the system can be optionally securely hashed and/or encrypted, using well-known algorithms like sha, rc4, des, or idea (among others). authentication is based on a public key model.

    --

    i speak for myself and those who like what i say.
  13. One platform != one choice by MosesJones · · Score: 2


    See Microsoft and modular argument. Linux is modular (not that nice an impl IMO), so are the mainframe architectures. Only have one proc ? Don't install SMP. Don't need domaining, don't use it. Don't need X,Y,Z then don't use them. Having a standard OS platform from which you can build your targetted OS is the approach I was talking about. In the same way as you don't compile the ISA support into Linux if you don't need it.

    OSes should be modular, the aim should be to get the best modules available from the best people to create the most flexible platform.

    One size does not fit all, just look at the size of the SUSE distro.

    --
    An Eye for an Eye will make the whole world blind - Gandhi
  14. Re:Plan 9 / Open source by the+gnat · · Score: 2


    Umm, Plan 9 isn't [gnu.org] open source...

    I'm amazed anyone really gives a shit. How many people whining about it are actually systems programmers? I'm happy that they're distributing the source at all- it doesn't bother me that I can't redistribute modified versions. RMS gets tiresome quickly- he doesn't seem to draw a distinction in moral terms between this kind of license and the MS licenses. Moronic.

    Also, I'm sick of hearing RMS talk about how "I'm not a supporter of the Open Source movement." Jesus, put the bong down and join the real world, where you have to cooperate with people.

  15. inferno plugin by Z4rd0Z · · Score: 2

    I decided to download the IE Inferno plugin and run the demos. I was really impressed with the performance. It really seemed much faster than any java applets I've used. And the download times were very short. It looks like a really cool system. But I really doubt it will be used widely. It's been around for a long time and I would venture to guess that only a relative few have ever heard of it. Too bad.

    --
    You had me at "dicks fuck assholes".
  16. Re:Inferno Security was examined in 2600 Winter 20 by anothy · · Score: 2

    i've responded to this article several times in several forums. the basic summary of the article is "if you mis-configure something you've installed as root, you can make bad things happen". well, gee, thanks. there's nothing specific to inferno here. inferno itself, either installed on raw hardware (like a normal OS) or hosted on top of another OS installed properly, as per the directions is quite secure, and does not have any known holes in it, nor does it expose any in the underlying system. you are instructed to install the installation as a user other than root - the fact that the author of the 2600 article gets it wrong from step 0 sort of taints his findings.

    --

    i speak for myself and those who like what i say.
  17. Re:Plan9 users? by slickwillie · · Score: 2

    Well, now that you asked, everyone at Bell labs supposedly uses it.

  18. Or maybe not eh ? by MosesJones · · Score: 2

    Java has become a ubiquitous development language on devices as diverse as mainframes and mobile phones. Inferno on the other hand has bugger all.

    Like me saying "I believe that Fluffy dinosaurs rule the world" it says more about the gullibility of the believer than the statement.

    --
    An Eye for an Eye will make the whole world blind - Gandhi
    1. Re:Or maybe not eh ? by konmaskisin · · Score: 2

      Like me saying "I believe that Fluffy dinosaurs rule the world" it says more about the gullibility of the believer than the statement.

      You can bet your ass taht if there *were* fluffy dinosaurs out there and they were smart enough they would rule the world. Maybe even if they weren't smart but there wre a lot of them and they were really hungry ...

  19. Doing things right. by karlm · · Score: 2, Informative
    Plan 9 and Inferno are examples of people doing things the right wayinstead of the easy way. They looked at UNIX and its problems and set out to fix them. I've read a few papers. Some of the features I attribute to Plan 9 may have also carried over to Inferno.

    Private namespaces -> Inferno gives each user/app a private namespace. If you're not allowed to see a file, it'not in your namespace, so there's no way you can even ask to see it. This is a good example of capabilities-based security. This is lightyears past the MS-DOS idea of each disk partition or network share being painfully appearant to the user.

    JIT optimized VM -> DIS, the Inferno VM, is based on a memory machine instead of a stack machine (a la Java and CLR/Mono). This allows for more efficient register allocation durring just-in-time compililation. Stack machines are great for writng smpleinterpreters with small memory footprints. Memoery machines are great for easily recompiling into fast native code. If I could, I'd start on an Open Source VM based on DIS. Toasters are great, but I don't want a crippled VM just so that it's easy to run on an 8-bit microprocessor in a toaster. You guys running SPARC, MIPS, POWER, PPC, IA64, etc. CPUs should notice the performance advantags of DIS more than us poor x86 users because the x86 is pretty register starved.)

    Distributed resources -> in Plan 9, there is a crippled user account without a password that pretty much can't doanything but present cryptographic credentials that prove it's doing work on behalf of a priveledged user. This would allow your dnet client to run on your CPU farm, but not actually be able to log in as you if it got compromised. As far as I can tell, the system is very similar to Kerberos with more types ofcredentials and tickets that never expire. I don't like the lack of ticket expiration , but it's better security than almost anything else out there. Most Beowulf implementations use rsh for performance, so you need to isolate the Beowulf compute nodes from anything remotely hostile, since rsh gives you a root prompt without a password based on the source TCP port number.

    --
    Copyright Violation:"theft, piracy"::Anti-Trust Violation:"thermonuclear price terrorism"<-Overly dramatic language.
  20. Re:Why do they keep ignoring us? by anothy · · Score: 2

    hey, he didn't say they won't get customers in antarctica, just that they don't have any yet. the distributed OS development community gets a little thin that far south...

    --

    i speak for myself and those who like what i say.