Already well-regarded as a security-focused network OS, OpenBSD stands to further upgrade its reputation with ver. 3.6, which combines enhanced services with wider hardware support. OpenBSD 3.6, which was released last month, will be a good fit for companies that wish to put services at the network edge, such as firewalls & VPNs, with more flexibility than appliance-based options could provide.
OpenBSD Widens Its Scope
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Huh?
OpenBSD being used for firewalls and VPNs is widening its scope? I haven’t gotten around to using it yet, but isn’t that what it’s always been used for? What else, exactly, is the scope of “a security-focused network OS”?
Why don’t you just say “OpenBSD 3.6 was released last month, and oh yeah, it had support for SMP”?
SMP on a router, here we come!
Speaking of OpenBSD widening its scope:
I wonder if OpenBSD will ever widen its scope so much that it will get 3D card/nVidia support too? FreeBSD has it, and even NetBSD has some beta grade nVidia/3D support already.
Of course it won’t be important at all for firewalls and servers – but one can always hope… 😉
Other than that, and as a more serious note, better performance and scalability are always good goals that could make make OpenBSD more popular in all sorts of server use besides of just firewalls/routers and (relatively) low load server use.
It’s a fact that crackers would have much harder time if people would just use OpenBSD (or at least its innovations and technology in their products) more. So making OpenBSD more popular world wide would be good for us all (except maybe crackers).
It can’t happen until nVidia releases either the source under an X.org compatible license or releases documentation for proper drivers to be written.
Yeah, ok. But how has NetBSD managed to get its version of the 3D nVidia drivers?
http://cubidou.nerim.net/
According to the above link/page:
“the interface to the kernel is open-sourced (the licence is a bit weird though), thus allowing a port to a different kernel.”
And also:
“Work would be easier if parts belonging to userland had an open-sourced interface too, but unfortunately it is not the case.”
Anyway, the nVidia 3D driver for NetBSD, although quite buggy and far from perfect, does exist. Would something like that be possible with OpenBSD too? Of course, enough people would first have to be interested in it for it to happen.
Sure it would be possible, but not done. A ‘weird’ license is not going to be put inside OpenBSD’s kernel, ever. OpenBSD will only become more open, not more restrictive.
It will also never happen because it is a crappy driver, if it does not work right there is no reason to have it in the operating system.
Good points and I kinda agree with you.
But if not the current nVidia 3D stuff, maybe something else then, at least later, hopefully?? I also mentioned nVidia just because they just happen to have the best 3D support for Linux and FreeBSD.
Ye its sad that corporations like nvidia doesnt have the balls to release info on their hardware so that proper drivers can be written..
“Ye its sad that corporations like nvidia doesnt have the balls to release info on their hardware so that proper drivers can be written..”
The last time this was posed as a question to Nvidia, they claimed it was because of the legal entanglements for some of the code they use in the drivers. Whether those patents extend to the hardware itself is anyone’s guess currently.
Obviously, I can’t speak for the OpenBSD developers, but somehow, I kind of doubt that getting the latest & greatest Nvidia drivers into the OpenBSD tree is not really a priority. I mean, sure, they’ll gladly support as much hardware as possible, but I’ve always gotten the impression that if they have to jump through a lot of hoops just to get a driver, they’ll just say “screw it, we don’t really need it that badly.”
In contrast with FreeBSD, NetBSD and Linux, the OpenBSD developers tend to be more strict with what can be included in the core OpenBSD operating system. It has to be very “open”…that is, freely redistributable, free to modify, etc. For example, when the author of the firewalling package “IP Filter” (ipf) put terms into his license that basically said “you can’t make derivatives of ipf without prior consent from the ipf author”, OpenBSD promptly ripped ipf out of their tree, wrote their own firewalling packet filter called “pf” and licensed it under a very permissive “do whatever you’d like with it” license.
So, like someone above said, until we see source code from Nvidia that can be freely modified and redistributed, I doubt we’ll see it in OpenBSD. At the very least, Nvidia would probably have to provide an OpenBSD binary driver that is freely redistributable, and then OpenBSD would have to decide whether it’s important enough to them to include something they can’t modify and don’t have source code for.
I think the OpenBSD developers would probably tell you “if it means that much to you to have support for 3D on this particular Nvidia card that you have, you’d probably be better off with another operating system that supports it.” From what I’ve seen, the OpenBSD developers have no qualms about telling people to use whatever tool is best for the job and whatever works best for you. There’s only so much they can do as a mostly volunteer effort.
Personally, the video card in my OpenBSD firewall is a 2 MB VRAM piece of junk that isn’t even connected to a monitor at the moment, and only exists in the first place for accessing the stupid non-serial-port-capable PC BIOS. I guess it serves me right for using junk hardware.
In response CISCO’s patented VRRP, the OpenBSD team has released CARP in OpenBSD 3.5 (Six months before 3.6 was released) For firewalls and many servers, this seems like a quite significant evolution. In addtition to allowing for high availabilty, CARP allows for load balancing and ‘rolling upgrades’ (take one box of a cluster offline & patch & return it to cluster, repeat N times.) CARP has already been ported to the other BSDs and Linux. It is most often used with pf, but it can be used with other services as well. With SMP support and CARP, OpenBSD is suddenly a highly scalable server platform.
So, like someone above said, until we see source code from Nvidia that can be freely modified and redistributed, I doubt we’ll see it in OpenBSD.
I don’t totally agree with that statement. I believe the Nvidia driver issue might be similar to the wireless adapter firmware issue they’re having right now. With regards to the firmware issue, OpenBSD isn’t asking for open source firmware. They’re asking for freely distributable firmware.
This principle might apply to Nvidia drivers. I’m assuming (I might be wrong) that OpenBSD doesn’t care that they are closed source drivers so as long as they can freely distribute the drivers. I don’t know if Nvidia prevents that or not.
Anyway, I’m just sayin’
Has anyone else noticed or is it just me, that online reviews invariably lack any real depth to them? The OpenBSD review OS News is linking to here is hardly more than just a rehash of the press release from the OpenBSD team. What happened to performance analysis, usability studies, and indepth looks at what these systems can actually do well or not so well? There used to be investigative analysis of claims made by software and hardware producers. Now it’s little better than high school journalism class projects. If I wanted rhetoric I could just read press releases.