Slashdot Mirror
BSDi's Software Divisions Acquired by Wind River
big knife writes: "Check here for the official press release." Wind River is the company that puts out vxWorks, the embedded OS that runs our great Arrowpoint Web Content Switch we have here at OSDN. Evidently, BSDi will become iXsystems, dealing with BSD product solutions, and hardware (essentially a beefed up Telenet Systems). The status of FreeBSD and their commitment to Open Source, and FreeBSD in specific is mentioned, definitely a good thing.
The article concludes with the question . . . "Wind River has made it clear that they plan to support -- and hope to gain the support of -- the BSD open source developer community. But the question remains -- how warmly will FreeBSD developers welcome the potential dominance of BSD UNIX by the world's leading proprietary embedded OS vendor?"
No. They couldn't even if they wanted to.
Warner Losh
FreeBSD Core
Wind River's products aren't all good. They're also responsible for WinPOET, the horrid PPP-over-Ethernet client most residential DSL customers live with. It's got memory leaks--bad ones--and has had them for a few versions straight. Use it, and you'd be led to think PPPoE is a terrible technology with a lot of overhead. The installer is pretty big and involved, too.
What's even stranger is that this comes from an embedded-systems software company. You'd think they'd be well-positioned to pull off a PPPoE implementation at least as small and stable as some of the (downright tiny) Linux ones. Setting aside the kernel modules now available, there are things like rt-pppoe, a userspace client that manages to chug along merrily in a few dozen kilobytes of RAM.
For anyone interested, it seems Slackware will remain with the old BSDi company (which will become iXsystems).
Oh damn! What am I going to do? I've got a subscription to both FreeBSD *and* Slackware! I hope they don't ship me blank CDs.
.. .. . . [end transmission]
If BSD is dead, then what's running this compp p&8 K nb!
A Government Is a Body of People, Usually Notably Ungoverned
I have been working with VxWorks for the past year.
You pay through the nose to get any tiny little bit of functionality beyond the barest essentials, the development envorionment sucks (constant crashes, horrible IDE performance), the OS claims to be POSIX compliant but isn't, and support is absolutely TERRIBLE. They have blown us off several times when we needed problems solved and questions answered.
The three reasons for using VxWorks today are:
1) Hard real time. And if you can stand latency of a couple hundred ns, Solaris 8 or IRIX can fulfill your needs.
2) Experienced workforce. In the embedded arena, there are many experienced VxWorks developers. If your team is composed of such people, VxWorks may become a good choice when you need your product out the door in a hurry.
3) Availability of various software components and drivers. Many vendors create drivers and modules (e.g. net protocols) for VxWorks first and often exclusively. The quality of these products is often mediocre, however one still pays through the nose for them as there is no other solution.
WRS' monopoly may have been fraying at the edges (embedded Linux offerings and others such as QNX are beginning to show a following), however, the technological advantage these competitors are displaying may be erased by this move (the BSDi acquisition). At long last, WRS will have a Real Operating System which will show all those haughty bastards with their Protected and Virtual Memory, POSIX compliance and OSS tools who's the boss.
I'd be absolutely thrilled to work with BSD instead of VxWorks (FreeBSD is my workstation OS of choice), but I fear this may well seal the fate of other companies such as QSSL and Be, which are true innovators and do not have the financial clout to push their competitors out by buying them (or equivalent solutions) out.
They use GCC extensively as their cross-compiler, but are not great contributors to GCC development.
Enjoy your job, make lots of money, work within the law. Choose any two.
I have been using and contributing to BSD since 1983 and FreeBSD since 1994. I have absolutely no wish for FreeBSD to gain market share in embedded devices or any other area. There are plenty of other systems in those areas and if that was what I wanted that's what I'd use. What we need is a greater diversity of approaches, rather than everything converging on what's currently successful.
As Dennis Ritchie might say, if you want QNX you know where to find it.
Now if they managed to crash a plane with most of those commiters, they might get somewhere.
Boss of nothin. Big deal.
Son, go get daddy's hard plastic eyes.
Expanding a vast wasteland since 1996.
I work with VxWorks on a daily basis as part of my job, and I'm not sure why people think this is a good thing? Wind River is *anything* but an Open Source company in their business model:
Strict licenses which must be renewed annually for access to tech support, bugfixes and updates (we're talking about tens of thousands of dollars here).
Mediocre tech support for smaller customers (unless, of course, you are willing to pay for them to send out one of their consultants).
A draconian closed-source policy that, among other things, forces customers to pay exhorbitant fees for read-only access to source code. Right-to-modify licenses are even more exhorbitant.
And the Tornado tools are supported on any OS you like... as long as that OS is Windows NT or Solaris 2.x.
Okay fine, WRS isnt Microsoft. But the enemy of your enemy is *not* necessarily your friend.
:Michael
How do you pronounce it? Presumably just like "uniX systems" but without the un.
-bugg
If Wind River is going to "integrate" FreeBSD the same way they "integrated" PSOS, then you can kiss it goodbye.
The BSD license allows for proprietary modules built on top of open software. When people at my workplace refer to "Microsoft embedded", they mean Wind River.
Embrace, extend, extinguish.
Go to their web site you'll see PSOS listed as a product. Then call them up and try to buy it without buying VxWorks.
Good luck!
--
What happens when you outlaw guns
Just as any one company cannot kill Linux, no one company can kill FreeBSD. BSDi does not own FreeBSD -- nobody owns FreeBSD. While it would suck big-time if they fired their FreeBSD developers, it wouldn't take more than a day for each of them to find a new job developing FreeBSD.
Wind River could kill BSD/OS. But by the time they do that, I suspect that the really cool stuff will have already been assimilated into FreeBSD. That process is well under way now.
If Wind River stopped selling FreeBSD CDs, how long do you think it would be until the FreeBSD Foundation picked up the slack? Or until Jordan Hubbard found a new employer who would do it for them?
I have the misfortune to have specified VxWorks for a project I am designing. I am sorry, but I trust Wind River about as far as I can throw them.
First, full disclosure. My project isn't the typical embedded project - it's large, it runs the X windows system, it uses TCL/TK, it has a good-sized hard disk. This is not exactly what VxWorks is targeting, and many of my problems stem from that. For a small project, it's probably good.
OK, now for the dirt.
1) I have NEVER, NOT ONCE, gotten any significant help on a problem from Wind River. This does not mean that they didn't try, but in the end every problem I ran into I ended up solving myself. For instance, when I was first building the system, I built it to load from IDE hard disk. I found out, after weeks of wasted effort and several calls back and forth to Wind River that the networking stack was disabled if you weren't booting from the network, unless you turned on an option that was not documented.
2) We purchased the RtX server that Wind River sells. It turns out the X server won't work with the latest version of VxWorks. The version of VxWorks it does work with does not support DNS nor DHCP.
3) The IDE driver doesn't use DMA, and the system does not understand partitions. I had to write the code to parse the partition table of the hard disk. And the lack of DMA sucks when you are trying to log a large amount of data to the hard disk.
4) The VxWorks file system is DOS. They have a hack to support long file names, but that hack isn't VFAT. So, you cannot simultainously have long file names and compatiblility with any other OS.
5) When we signed up with Wind River, we asked them as to when and if they were going to support development using Linux as a host. Since they already support Solaris, this SHOULDN'T be rocket science. We were told "within 6 months". That was two years ago. Guess what.
6) The DOS file system has bugs where it tries to free() a static buffer. I don't have the source, I cannot fix the bug. RMS has a good point here.
7) The X server only supports 256 color mode. The server is derived from XFree, but since XFree isn't under GPL, Wind River doesn't distribute the source. Therefor, I cannot build it to support TrueColor. I cannot change the keysyms (which would be damn useful since I have keys that a standard X server doesn't have.)
8) On the StrongArm platform, they claim to support that chip, but they don't enable the on-chip cache, nor do they give you a good place to do so. Therefor, the chip runs at half speed.
9) Their implementation of the TCP stack has bugs. They pass your driver a buffer you are to place the packet into, but they don't tell you that because of a bug in the driver, you have to place the packet at bufferstart+2.
10) Their debugger does not run reliably on a network with a lot of traffic.
Given my experiences with Wind River, I don't feel warm and fuzzy about their purchase of ANYTHING. As soon as I can convince management (and that won't be long) I will be porting my project over to Linux.
www.eFax.com are spammers
Tech: What kind of system is that?
Admin: That's an ick systems box.
Tech: Well if you don't like it, why don't you return it?
Even though PicoBSD is a stripped FreeBSD....and "freeBSD will continue as before", PicoBSD is most like the space WindRiver is normally associated with.
Will they kill Pico?
If it was said on slashdot, it MUST be true!
pSOS is an embedded kernel (and all the various add-ons), formerly written and sold by "Integrated Systems, Inc" (ISI). A little over a year ago, Wind River purchased ISI. Only a couple months after the aquisition, they announced that pSOS would be discontinued. Maybe there's another explaination, but it appears that they purchased ISI to bury the competitive pSOS kernel.
Only several months before this happened, I had started a little embedded project where I work, and we decided to purchase a kernel and TCP/IP stack. I spent about a week checking out the various vendors, and I investigated pSOS quite a bit, since some of our customers had some very successful products based on it. I also took a good look at Linux. I really wanted to use linux, but the truth is that it requires quite a lot of memory and a 32 bit chip, and I was hoping to stay with a 16 bit chip and smaller memory. Still, I started out with an absolute requirement that the vendor provide source code. ISI never provided (normal) customers with source, and Wind River is about as closed as closed source gets. I ultimately went with US Software, who provide source and have a product targeted at smaller systems. ATI Nucleus was a close second choice. I utlimately obtained both vendor's API reference manuals, and USSW's were more down-to-earth (provided a much better conceptual model of what their code was doing) and they were easier to use. USSW's TCP/IP stack comes with two interfaces, the usual sockets interface we all know and love, and their own very light weight interface. If you use the light weight one, you can compile without sockets (saves about 12k code space). I needed to add a tiny feature to the light weight one, and within 1 day I was able to read through the TCP/IP stack source and understand it enough to add the thing I needed with good confidence I was doing it well. I made a similar addition to their kernel, in the space of about one day. With these small changes to customize their interfaces to my needs, I got my code running in only a few weeks, and I was able to produce a 16-bit x86 (real mode, yuk) executable image that was about 50k that included their multitasking kernel, tcp/ip stack, and my old single-task app converted to nicely run multi-threaded to serve multiple concurrent sessions, all running on very low cost hardware, AMD's Net186 Eval Board. Having the source code for whatever kernel you're using in an embedded project is a major advantage. Don't ever let those slimey salesmen tell you otherwise!
Wind River (and the former ISI) are closed source. They put a lot of effort into sales and marketing, and they put quite a bit of effort into trying to convince me that it wasn't an advantage to have source code. Fortunately where I work the management is pretty sensible and doesn't presume to be able to evaluate kernels and network stacks. When it became obvious I'd never select them due to being closed source, they made a couple attempts to directly communicate with my managers, which raised a couple questions, but they trusted my judgement that having source code was critically important.
I know of a similar group that had started a project based on the Netsilicon chip, which at the time was only supported by pSOS from ISI. Netsilicon provided source for their device drivers, but the pSOS kernel and stack were closed source. Well, there were a lot of really unhappy campers when Wind River bought ISI and announced they would bury pSOS, and Netsilicon and their customers were certainly not amused. Getting stuck with an obsolete object-only library that has a bug really sucks.
So before anyone gets really excited about the great things Wind River might do for BSDi, take a little look back at the not-so-distant history where they purchased ISI and then almost immediately announced the death of ISI's core product, pSOS... with what appears (from my limited point of view) as an utter disregard for the installed pSOS customer base.
An remember, these guys are closed source proprietary software. They spout all sorts of marketiod language about "total cost of ownership", "industry leading [insert word]", blah, blah, blah. They make some pretty impressive product offerings, yet there is no shortage of horror stories of someone who hit bugs or needed to add a feature and was absolutely helpless without the source code. I saw a good example of one of those stories above, and I hope it gets moderated up to 5 (though PHB's wouldn't ever read slashdot), having the source code can make all the difference between spending a day or two customizing/bug fixing and spending weeks of frustration on the phone and ultimately working around it somehow.
PJRC: Electronic Projects, 8051 Microcontroller Tools
To see a company specialising in embedded systems embrace BSD like this is incredibly good news. Why? Because no matter how many servers there are, there will be 1000 times more embedded devices. This is the current battleground, and I would love to see FreeBSD steal some major marketshare in this arena.......
:)
That, and a constant, reliable source of funding for the FreeBSD project will ensure its dominance (flames anyone?) in the Open Source arena........
God damn, its feel good to be a FreeBSD zealot right now.
Feed The Need[goatse.cx]
This is real simple.
An analogy. Imagine that the FSF and Linus Torvalds jointly decide that any futher development they make towards GNU software and the Linux kernel will be proprietary. Forget that they wouldn't ever do that, just realize that because they own the copyright they could. But sitting on your desk is a CD of your Linux distribution of choice. It contains all the source code, and that version of the code is under a free license. You could take all that source code and continue developing it yourself. It would not die, as long as someone was willing to keep it going.
The same is true for FreeBSD. WindRiver could say that all further development they make to FreeBSD will be proprietary. But sitting on the desks of all the FreeBSD developers is a CD containing all the source code to FreeBSD, and that particular version of the source code is covered by a free license. They would simply take that source code and continue their project unhindered.
And you can believe that if WindRiver turns out to be a bad host, the FreeBSD developers will do exactly that; WindRiver can't prevent them from taking their toys and going to play somewhere else.
I suppose this means slashdot is dead too... seeing as how they were aquired by Andover and then OSDN... being bought isn't being killed... I use BSD and linux... open source OS'es are damn hard to kill, because they have loyal fans willing to carry on even if the funding and support stops... FreeBSD rocks my UNIX world... it's faster, more stable, and runs on machines with less than 128 megs of RAM... heck even /. had a post on that recently
#set prompt = $user.$group @ `hostname -s`#
root.wheel @ reality#
Remove *your pants* to send me email.
Slashdot actually does okay under the load considering that they actually use those things. I was actually surprised to hear them boast about the arrowpoints at all after their less than spectacular performance with them before, both in the load handling and the resulting customer service they received (the "Great DDoS" episode right about the time they switched toexodus).
Fist Prost
"We're talking about a planet of helpdesks."
Fist Prost
"We're talking about a planet of helpdesks."
-Jaron Lanier
This is probably more of a Good Thing (tm) for BSD than a Bad Thing. The Wind Rivers has always been a great company to deal with (I have worked quite a bit with vxWorks) and they will be able to scrounge up interest in BSD (which is much needed). As far as I understand vxWorks is basically the industry leader in RTOS's, so it's a company that will be around for quite awhile and gives BSD some security.
I thought someone said there was going to be free beer!
I cannot help but be worried about the Wind River purchase. At least, when BSDI purchased Walnut Creek, there was the feeling that they at least had some connection to the Open Source community through the shared BSD heiritage. Now, Wind River has absolutely no committment to Open Source, since the entire premise of the company is selling proprietary software. Also, they have a very bad track record with previous acquisitionns (rememer pSOS and ISI?) Even though they have Hubbard and Karels and McKusick as talking heads saying good things, exactly how long will the good feelings last? According to the article on Yahoo!, the reason WRS did this was to try to force more people to use their proprietary tools, not to promote openness.
Just ask pSOS customers what the future of pSOS became when Wind River bought ISI... --Blair
I was just on the Wind River conference call for investors/mediapeople/etc and a 'personal investor' asked "What are the plans to make FreeBSD support javascript?" I just thought some of you could use a laugh. =)
I'm highly concerned about this purchase. In my opinion, as a regular BSDi customer, I don't want them to change in the slightest. They offer the best/most knowledgeable/fastest customer service I've seen. Their sales staff are extremely knowledgeable, their tech support staff is quick and clued. And just when I thought I've found the perfect server vendor...