Announcement

Collapse
No announcement yet.

OpenIndiana 2018.10 Released With MATE 1.20 Desktop, GCC 8 & Python 3.5 Support

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • OpenIndiana 2018.10 Released With MATE 1.20 Desktop, GCC 8 & Python 3.5 Support

    Phoronix: OpenIndiana 2018.10 Released With MATE 1.20 Desktop, GCC 8 & Python 3.5 Support

    A new release of the Illumos-powered OpenIndiana Hipster operating system is now available as one of the leading open-source Solaris-derived operating systems...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Serious (not snarky) question: what is the sales pitch for OpenIndiana vs Linux or one of the *BSD projects? What does it do better?

    Comment


    • #3
      Originally posted by Michael_S View Post
      Serious (not snarky) question: what is the sales pitch for OpenIndiana vs Linux or one of the *BSD projects? What does it do better?
      Multiplicity of choices? I'd say Linux monoculture or BSD monoculture would be worse than having multiple choices. It's not like OI devs would certainly develop for Linux if they didn't have OI.
      Can't see the point of such picking "whats the point, whats the sale pitch"

      Comment


      • #4
        Originally posted by aht0 View Post

        Multiplicity of choices? I'd say Linux monoculture or BSD monoculture would be worse than having multiple choices. It's not like OI devs would certainly develop for Linux if they didn't have OI.
        Can't see the point of such picking "whats the point, whats the sale pitch"
        I'm not against a multiplicity of choices. I think that's a valid value, and I too oppose monocultures. I just thought there might also be a killer feature or two, or at least some nice-if-not-quite-killer features.

        Comment


        • #5
          I guess one difference compared to Linux is that Solaris then illumos have been developed as a whole operating system: there is a tighter coupling between kernel and userland.
          Therefore technologies are very well integrated with each other: for example ZFS is leveraged in the package management (Boot Environment, snapshots, build Zones and package repositories in virtual envs), the user interface (navigation through different version of files in the ZFS history), operating system level virtualization (Zone integration with snapshot, update of virtual Zones).
          The Crossbow network virtualization framework is also very well integrated.

          Solaris/illumos offers a few nice features that have equivalents (Zones are inspired from FreeBSD Jails), successors (dynamic tracing now with eBPF), or have been replicated on other systems (DTrace, ZFS, Zones, kernel scheduler, slab allocator etc...).

          The Project management and Role Based Access Control (RBAC) model frameworks are also very well integrated.
          I have read that something adapted from RBAC has been proposed for integration in Linux.
          The system offers APIs for different aspects, that are pretty transparent from userland.
          For instance Locality Groups (NUMA affinity) have been around for more than 15 years and provided by the system instead of relying on "third party" libraries.

          I would say that observability and integration of the technologies in the operating system are strong points of illumos.
          But on the other hand reliability and observability are not sacrified for performance so illumos may not be your number crunching platform of choice.

          The community of users and developers is much smaller than FreeBSD and Linux ecosystems so the hardware support is not as good.
          However considering the cross-pollination of technologies with other systems like FreeBSD and Linux, Solaris/illumos has contributed to innovation to some extent (even with more moderate manpower). IMO Illumos remains a interesting system that strives for consistency and observability.

          Comment


          • #6
            Thanks for the walkthrough. I appreciate it.

            Comment


            • #7
              Originally posted by Michael_S View Post
              I'm not against a multiplicity of choices. I think that's a valid value, and I too oppose monocultures. I just thought there might also be a killer feature or two, or at least some nice-if-not-quite-killer features.
              Zones, DTrace (not sure if Linux equivalent has reached parity), better vertical scaling, RBAC, IPS package manager.. Linux have equivalents or near-equivalents for everything but experience might be smoother and give impression of being thought out/designed better on Solaris/OpenIndiana.

              Comment


              • #8
                Glad to hear OpenSolaris it't not dead, yet.
                Actually Oracle open-sourced DTrace under the GPL.
                Last edited by onicsis; 24 October 2018, 11:25 AM.

                Comment


                • #9
                  OpenIndiana has been main desktop platform for years: never regretted it. Very satisfied with Illumos itself in general, Joyent's work on it and OI development team... wish the best to them

                  Comment


                  • #10
                    Originally posted by Mifune View Post
                    OpenIndiana has been main desktop platform for years: never regretted it. Very satisfied with Illumos itself in general, Joyent's work on it and OI development team... wish the best to them
                    That's interesting to hear. Are you using the latest version with MATE? Does OpenIndiana include the package mate tweak in its repos?

                    Comment

                    Working...
                    X