Announcement

Collapse
No announcement yet.

Linux Distributions vs. BSDs With netperf & iperf3 Network Performance

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

  • #21
    Originally posted by aht0 View Post
    rofl... and I sorta remember pal666, Pawlerson, SystemCrasher and such preaching some months a go in other BSD-related threads how ***BSD networking stack is way inferior to one of Linux in every conceivable way...
    Two questions:

    1. Was firewall enabled in benchmarked systems? (in Linux distributions it's usually enabled, but it's probably not enabled by default in FreeBSD).
    2. Does it matter?
    Last edited by Pawlerson; 12-07-2016, 03:51 PM.

    Comment


    • #22
      Originally posted by labyrinth153 View Post

      Dozens of benchmarks where FreeBSD "smashed out" Linux. Stop being a fanboy.
      Keep this advise for bsd fanboys or shut up.

      Comment


      • #23
        Something is clearly off here. Having results that differ almost by an order of magnitude between linux distros and between linux and freebsd on the same hardware is IMO simply not possible. There must be a huge difference in default settings somewhere. That doesn't invalidate the results, by the way, because they still show what you get out of the box by choosing the various OSes, but it would be good to know what it is that is being really measured.

        Comment


        • #24
          Originally posted by Pawlerson View Post

          Keep this advise for bsd fanboys or shut up.
          sadly, you are defining the term "fanboy". When you have zero arguments left, you always resort to plain bullying, insults or sneering.
          Originally posted by thomasj View Post
          What happened with DragonflyBSD? I thought that it was supposed to have the best performance of the BSD's?
          I think it's wishful thinking. It'd have to have monolithic kernel to beat other OSes with monolithic kernels. Those have the best performance (but worst security).
          Originally posted by Pawlerson View Post
          Two questions:
          1. Was firewall enabled in benchmarked systems? (in Linux distributions it's usually enabled, but it's probably not enabled by default in FreeBSD).
          2. Does it matter?
          It'd be pretty stupid to turn on firewall while benchmarking network performance. Do you want untainted results or not? Or maybe stream 4K movies as well while benchmarking? Just to kill the time while benchmark is running.

          And I'll remind you that BY DEFAULT FreeBSD kernel variables are quite un-optimized in turn. Not to mention slightly less-optimized binaries compared to binaries compiled using GCC.

          Ask Micheal what settings he was using.
          Originally posted by Pawlerson View Post

          That's why enterprise computing uses FreeBSD. Oh wait, they don't! Just because FreeBSD stole ZFS from Solaris doesn't mean it's better when comes to filesystems. It only has ZFS while Linux has many filesystems for different workloads.
          And enterprise computing is mostly using 1. XFS.

          Having bunch'a stuff working half-assed, accompanied by "works in case of..", "when you have.." and so forth is not very impressive. Just bragging rights.

          Comment


          • #25
            I'd bet some systems use a firewall, were others don't AFAIR FreeBSD doesn't enable one by default.

            Comment


            • #26
              Originally posted by aht0 View Post

              And enterprise computing is mostly using 1. XFS.

              Having bunch'a stuff working half-assed, accompanied by "works in case of..", "when you have.." and so forth is not very impressive. Just bragging rights.
              To be fair, enterprise computing is not really using "XFS". It uses RHEL which incidentally happens to use XFS. How many companies really know or care about the FS, let alone compare the various options and choose one based on some objective criteria? Companies pick an OS based on support, software availability, hardware certification and, in some cases, overall performance. Heck, an awful lot of enterprise computing is running on NTFS, but that definitely doesn't imply that it's a great filesystem. Which XFS is, by the way.

              Comment


              • #27
                Originally posted by Pawlerson View Post

                Two questions:

                1. Was firewall enabled in benchmarked systems? (in Linux distributions it's usually enabled, but it's probably not enabled by default in FreeBSD).
                Fedora - Yes
                Ubuntu - No (apparently doesn't even have the kernel module installed by default)
                If he was just testing the default install of FreeBSD the firewall (or rather your choice of PF, IPFW, or IPFilter) would be disabled

                I don't have CentOS or debian installed in a VM ready at hand to test them

                Originally posted by Pawlerson View Post
                2. Does it matter?
                Going by the benchmarks In some cases yes, in others no

                Comment


                • #28
                  Originally posted by jacob View Post

                  To be fair, enterprise computing is not really using "XFS". It uses RHEL which incidentally happens to use XFS. How many companies really know or care about the FS, let alone compare the various options and choose one based on some objective criteria? Companies pick an OS based on support, software availability, hardware certification and, in some cases, overall performance. Heck, an awful lot of enterprise computing is running on NTFS, but that definitely doesn't imply that it's a great filesystem. Which XFS is, by the way.
                  I seriously doubt there is something "incidental" in RHEL using XFS. They must have tested the fuck out of various file systems and picked file system according to test results and not by simply tossing a coin. Even Novell now recommends using XFS for data partitions on it's SLES.

                  Originally posted by TeamBlackFox View Post

                  Just one look at OpenBSD-derived PF vs the Linux iptables and you can see an emphasis on clarity and user friendliness over feature-itis.
                  I'm sort of starting to like NetBSD's npf as well. One dude has done amazing amount of work in a ~5 years..
                  Last edited by aht0; 12-07-2016, 06:58 PM.

                  Comment


                  • #29
                    Originally posted by aht0 View Post

                    I seriously doubt there is something "incidental" in RHEL using XFS. They must have tested the fuck out of various file systems and picked file system according to test results and not by simply tossing a coin. Even Novell now recommends using XFS for data partitions on it's SLES.


                    I'm sort of starting to like NetBSD's npf as well. One dude has done amazing amount of work in a ~5 years..
                    XFS was made by SGI and for its time SGI did some innovative work with its stuff. In my experience a properly configured XFS or even JFS system can smash an ext4 out of the water.

                    I've only used NetBSD in mostly embedded ARM boards and the occasional ancient desktop/laptop, hardly a good place to test their firewall, however, I imagine it's good. I just need to see if their network code is giant locked because if it is, then that isn't good.

                    Comment


                    • #30
                      Originally posted by aht0 View Post

                      I seriously doubt there is something "incidental" in RHEL using XFS. They must have tested the fuck out of various file systems and picked file system according to test results and not by simply tossing a coin. Even Novell now recommends using XFS for data partitions on it's SLES.
                      You misunderstood my post. I don't doubt that Red Hat did their homework. My point is that enterprises don't give a toss about the technicalities, pros and cons of the various filesystems. They use whatever the OS vendor supplies and if Red Hat selected FAT16 as their FS of choice, then business computing would run on that. Thus claiming that XFS is somehow acclaimed by enterprise computing is meaningless, like bragging that Fortune 500 companies embrace the Linux syscalls ABI.

                      Comment

                      Working...
                      X