Slashdot Mirror
Microsoft's Implementation Of IPv6
jinx_ writes: "For those of you who were interested in the OpenBSD IPv6, Microsoft has a site of their own on the subject. 'Microsoft Research (MSR) is writing an Internet Protocol Version 6 (IPv6) implementation to further networking research on the Windows NT/2000 platform. USC/ISI East is our partner in this development. Due to external interest, we have decided to make a beta version of this implementation publicly available in both source and binary forms.' Sounds like it would be fun to play with at least." Anyone know anything more on this? Post below, please.
And the microsoft IPv6 stack has been out for over two years in an unsupported research capacity.
I'm really considering working there.
- I don't care if they globalize against free speech. All my best free thoughts are done in my head.
Comment removed based on user account deletion
Many Microsoft Research studies are purely academics. They do some great work. Only after it gets to their employer (and Bill Gates's teams of programmers get to hammer at it) does it become flawed.
You have to separate the company from the research team.
- I don't care if they globalize against free speech. All my best free thoughts are done in my head.
$20 says that, when all is said and done, Microsoft's version of IPV6 will be ever-so-slightly incompatible with the standard implimentations.
"Gee paw...i can't get to this website running FreeBSD anymore...it must be a crappy OS"
FluX
After 16 years, MTV has finally completed its deevolution into the shiny things network
"It is seldom that liberty of any kind is lost all at once." -David Hume
This is like..... moldy. It's _SO_ old. I have a copy here that I'm sure has been "touched" so the file is even older than the date on it. ...
-rwxr-xr-x 1 xxx xxx 286720 Jun 1 1998 msripv6-netmon-src-1.0.exe
-rwxr-xr-x 1 xxx xxx 507392 Jun 1 1998 msripv6-src-1.0.exe
Talk about missing the boat by over 2 years....
Bruce
Bruce Perens.
So before you start proclaiming that Microsoft has seen the light, remember that MSFT is obviously acting its own self-interest here. For whatever reasons, some PHBs decided that it would be to Microsoft's advantage to open-source this baby (and it is a neat little project, I'll admit). It's certainly likely that we'll benefit from it too, but at the end of the day, Microsoft is a publically-held corporation and is legally obligated to maximize its profits. Don't think for a minute that they're doing this out of the goodness of their hearts -- they're just trying to make money, whether it be through closed-source software or now open-source software.
variable declarations like: char *name = NULL; (where's the char *szName?), and you get the feeling you've entered a parallel universe where Unix programmers have gone to work for Microsoft.
Paul
From da soft: "... implementing IPv6 to further networking research on the Windows NT/2000 platform."
Translation: "...IPv4 was really really hard. I didn't really get it. One time I was working on the NT4 drivers and I forgot what line I needed to GOTO and it turned out that all network traffic was getting opened as an Excel spreadsheet. I was moved to the talking paperclip project, but now I'm back for IPv6 cause it's easier!"
This reminds me of a Simpson episode where the following conversation takes place:
Lisa: Please, Dad.
Homer: No.
Lisa: Please, Dad.
Homer: No.
Lisa: Please, Dad.
Homer: No.
Lisa: Please, Dad.
Homer: No.
Lisa: Please, Dad.
Homer: No.
Lisa: Please, Dad.
Homer: Oh, okay, okay.
Hitachi has had out a 9x and NT 4.0 IPv6 protocol stack out for a while now, and it is free and works well. It's nice to see that 3rd parties can contribute greatly towards the MS platform just to have them poke out and claim that something they did is the first time it has ever been done on their OS. Realisticly though I don't think they will have a stable product till at least 2nd Qrt. 2001. But right now I have an IPv6 only sandbox running on a few FreeBSD 4.1 --STABLE and OpenBSD 2.7 systems thanks to KAME.
As more and more devices (cell phones, PDAs) become IP-enabled, 32-bit IP addresses will become increasingly scarce, and eventually they will run out. Some people are predicting this will happen in just a few short years. Moving to a larger address space, such as the one afforded by IPv6 is the only answer.
Unfortunately, the fate of IPv6 rests in Microsoft's hands. If IPv6 is to ever attain widespread use, Windows will have to support it. The sheer number of Windows machines out there guarantees it. No matter how soon Linux and BSD servers support it, it will be pretty useless without widespread client-side support, and that means Windows support.
MS has had IPv6 working in the research labs for a long time, yet they are really dragging their heels when it comes to putting it into a shipping product. Beats me why. I suppose they might have some financial interesting in seeing IP numbers getting scarce ("If you want your own IP, you have to sign up for MSN!"), but somehow I don't think even Microsoft can hold back the rising demand for more IP addresses.
So, sooner or later, they are going to have to include IPv6 support in Windows by default. And not just the server-branded versions of Windows either, but the consumer versions as well. The Windows that Joe Bloe runs on his home PC will have to come with IPv6 built-in. Otherwise, Internet growth will be stifiled. Isn't it scary to think that the future of the net rests in Bill's hands?
Free Hans!
Satan, get out the snowblower, Microsoft is posting source to what could become part of their operating system.
Check out the date on the press release -- March 15, 2000. This is not news.
Ken
From the docs:
Just kidding...
What about running IPv6 on 9x/NT/2k? Anyone have any reports on it? How hard was it to get running? Did your apps play nice with it? How was performance? Who was there to talk to? Any practical immediate advantages? Are there any ISP's yet offering IPv6 support? What will AT&T @Home (is that their name this 24 hours?) do if I start running IPv6? Did the MS implementations interoperate with other vendors? Which one seems best under WinWhatever? Which one seems best overall (Linux/BSD/etc.)
I don't read ACs: If a post isn't worth so much as a nom de plume to its author then I wont bother either.
Microsoft currently has two IPv6 stacks available for download. One is the TechNet IPv6 Developer Preview, which is a 'snapshot' release for programmers and the other is the Microsoft Research IPv6 stack which they are constantly adding new features to.
The stacks are currently command-line based. No cute GUI tools, it is cryptic to setup but seems to be quite stable in my IPv6 lab. I use FreeBSD/KAME as my tunnel broker server and GNU/Zebra as my BGP4+ router with a session to Sprintlink.
The only current weakness is the real lack of applications available for the Microsoft Windows platform on IPv6. No SecureCRT, mIRC, Bulletproof FTP or what not. Now on the BSD/Linux side that is the exact opposite, almost every concievable application has been IPv6'ified.
For more information, evaluate www.6bone.net, www.kame.net or your BSD/Linux distribution's web site.
-Pat
How about creating something like Trumpet Winsock. Like in the days before Microsoft implemented Ipv4 in Windows ? In other words, take the Ipv6 spec (from OpenBSD?) and write a driver for windows, and put it in the BSD licence (eventually with a M$ clasule :). Then push the spread of it.
Ofcourse writing a windows driver may not be fun. But the cause is good :)
Hate to be picky, but Agilent is not HP Research. Agilent is a separate company that develops, markets etc all the former HP telecoms, medical and "anything that isn't directly a computer" products. They decided to split the company up because the bean-counters thought it would look better on their spreadsheets! (Not a joke, unfortunately). Anyway, HP continues to do all the computer related R&D itself as always, although it does outsource some of that to small, specialist outfits .
Stealing a rhinoceros should not be attempted lightly.
Here is an interesting story about Microsoft, and it (slightly) involves IPV6.
I went to "Microsoft's Big Day" back in March I believe. This event was (at our town at any rate) just a big propaganda machine for Windows 2000 and Office 2000.
The hotel where it took place was initally crowded with people from the buisnesses from town, but with each intermission (the "seminar" lasted a whole day).
Basically the lectures went over the features of Win2k and why you should buy it for your buisness, same thing for Office. The main presenter (other than the boring laywer who read from the EULA... No, I am not joking)was a woman who seemed quite knowlegeable about NT. She was quite sharp I thought.
I decided to test how sharp.
I walked up to her during an intermission, where people were asking very very basic questions.
My turn came up and I asked:
"When will the Windows 2000 kernel support IPv6?
Currently it only supports IPv4, and thats a serious issue with the looming IP shortage."
Just for a second her eyes went a little wide - the first question all day that she had not been able to answer. She glanced quickly at a person nearby sitting in the front row, then looked back at me and said "I don't know".
This was fine, I did not expect her to be able to answer the question, I wanted to see her true level of knowledge, whether she was plain PR or a techie at heart.
Now what got interesting is that the fellow to my left who was sitting in the front row of the presentation (dressed in "plain" clothes)and had been the man that the presenter had glanced at, got up and began to praise Windows 2000. He mentioned how "No operating system supports IPv6".
I replied, "Funny, Linux and BSD support it." He did not believe me at first, and addressed the *nix idea with a wave of his hand, as if the *nix OSes were naught but a bother. We then argued about IPv6 and it's importance, and how it loads routers etc, etc, etc. But, as we did so I noticed that he was leading me further away from the people asking questions to the presenters (I was winning the argument because I had just read Understanding IP Addressing: Everything You Ever Wanted to Know.) I was also declaring things like: "Well, Linux can do that! What do you mean that Windows can't?" Which seemed to irk him.
(Ok, before someone tells me to read the Advocacy-how-to, I was very polite about it, and not derogitory to MS, I was doing it in more of a "Gee, I thought Windows could do that too... You mean it can't?" Besides, YOU try sitting through an 8 hour MS propaganda session and see if you don't snap!:)
We finished arguing, I "won" not that it was really important. I did not really care. Still, what I thought was *really* interesting was that I did not recognize him. I live in a small town, and I know ALL the computer people here. They all know me as the local Linux geek. I never saw this guy before, and he *WAS* knowlegable, he *DID* know what IPv6 was, and was able to discuss it. I would have known if there was a guy like this in town.
I waited until the very end of the seminar, when everyone was leaving. I watched this "plainclothes" guy, (all the MS people had Microsoft shirts on). The "plainclothes" guy left in the same van that the MS people left in. I have not seen him in town since.
Interesting don't you think?
Try to hack my 31337 firewall!
Cute, but did you have to demonstrate your own lack of knowledge with you first question? The NT/2000 kernels don't "... support IPv6?" or IPv4, or IPX/SPX or AppleTalk or any transports in the kernel. The protocol stack is essentally a file system shim between the redirector (rdr.sys) and the hardware drivers (also abstracted vis the HAL). That how the IPv6 released 2 years ago was able to install on NT5b2 along side the existing IPv4 stack.
Didn't have shit to talk to and was not very stable (given the b2 system, anyway) but it did work and did not change the kernel at all. Even installed without a reboot.
The license agreement has a bit in it about not reverse engineering the product. This seems to be a rather strange clause, given that they're also distributing source; when you can look at the source, you don't need to reverse engineer. Of course, this probably just boilerplate text for a license agreement. But my more paranoid side thinks that maybe this is there so that, should Microsoft add any "embrace and extend" incomabatabilites, no one will be able to duplicate those incompatabilites by looking at the source code, since that would break the license agreement. Thus they'd be able to have their cake, and eat it too: get people to peer reivew and fix their code, but no other IP6 projects can get any use out of it.
Suppose you were an idiot. And suppose that you were a member of Congress. But I repeat myself.
Give a man a fire, and he'll be warm for a day, but set him on fire, and he'll be warm for the rest of his life.
You can get the newest version here.
Here's the stack in action:
That's over my WaveLAN wireless PC Card in my Win2K laptop to my flatmate's Libretto C100 running a recent NetBSD-current which is our WaveLAN - LAN gateway. All of our boxes are IPv6 native. No IPv4 encapsulation for us. And yes, WaveLAN kicks ass! You NEED WaveLAN.
So, in answer to one of the major questions, Microsoft's stack works with other IPv6 implementations. It doesn't keep settings between reboots at the moment, and it doesn't do ESP only AH.
Andrew van der Stock
It's put-up time at Microsoft Research. Seven years after its founding, the lab has yet to make any real breakthroughs.
Got Rhinos?
That still comes out to 6.671e5 IPs per square nanometre, or 816 IPs on each side of the square nanometre. It's unlikely that nanomachines will be packed that densely within the foreseeable future. If so, then you can just go to a, say, 256-bit address space, which would give us 1.1579e77 IPs, more than the current estimate of the total number of atoms in the universe. Hopefully that would last us a while.
32-bit IP address are not running out. Do you actually think there is a need for 4.3 billion publically addresses machines? Most computers today are in corporate situations that don't touch the internet except through firewalls, but currently use non-reserved ip blocks for the chance of one day things changing.
I will say, that without proper management, they could run out, but clearly look at this stupidity: "whois 3.0.0.0@whois.arin.net". Does GE need 16.7 million addresses?
What we need is variable subnet masking working on All products, and supernetting working on all routers, all the issues go away. That, and have companies justify having thousands of addresses when they only have a couple pingable IP's (and usually those are on a separate network anyway)
I chose the question based on that.
I knew the stack was separate. The whole point of the question was to see how much the presenter knew, not getting an answer to the question.
I wanted to see if I got back a response like, "No, the kernel and the stack are separate" or "I believe IPv6 is forthcoming" or "what kind of question is that?"
As you just caught me on the question, I wondered if she would catch me on the question. You can choose to believe that or not, but it was the point.
Try to hack my 31337 firewall!
It does say 'implementation' which means that they are going to make it available for Windows OS.
I am not that familiar with windows tcp, but my guess is that it does not currently support IPv6. This could mean that they are now going to. The problem I see is that they could potentially do the same thing to it that they did to kerbos.
I don't want a lot, I just want it all!
Flame away, I have a hose!
Only 'flamers' flame!
While i believe you, that they probably have some fine engineers over at MS the problem is, that it's not them that make the truely important design decisions , but marketing. This can be seen at the example of bundling IE with Windows: while at first this didn't have anything to do with the design of either product (hey they just gave it away for free, you pay the price with MS Windows anyway) later IE was intertwined with the OS so much that it's now really not easy to rip it out without destrying something (for the big userbase). It's obviously bad design to give up modularity and every engineer will tell you that, It's also obviously good design to have clear, well documented interfaces (maybe there are, but the documentation is well hidden) so you can exchange one thing for another.
... so most businesses who have a lot of correspondence with customers via Word documents use two versions: an old one for creating documents goin out to customers (who might not have updated to the latest version) and the latest version to be able to read their customers documents). Now all these are design decisions no sane engineer would burden a sensible piece of software with, it's just microsoft marketing screwing over the customers.
It's also very obvious that applications should be compatible with their preceding versions if possible (try to edit a Word document with an older version than Word than what last touched it
"By the way if anyone here is in advertising or marketing... kill yourself." -- Bill Hicks
1 - What do people need from IPv6? What can we do to make this a bigger need or a need exclusive to us?
2 - Can I embrace and extend?
3 - What does this cost me? Support? Development resources?
4 - Can I release just before a crisis (remember Y2K? - oh maybe just after the crisis) to force an upgrade? Most consumers haven't even heard of IPv6 yet. Thus no economic value.
Actually I think that this release is more genuine than most people think. MS _has_ to remain compatible with the rest of the market at least to a basic degree. Remember all the articles about the companies, including MS, that use something other than NT for their web servers? They aren't stupid. They would not make their OSs incompatible with half of The Net.
Hey, maybe they're waiting until all their web servers are NT based. This could take a while...
The article got turned down.
So it's not that Microsoft don't innovate, it's that Slashdot deliberately avoids covering it when they do.
Anyone with half a brain and a reasonably open mind can see that in fact Microsoft have innovated hundreds of ideas over the last decade, which the Linux crowd is still busy copying. That's why Windows is usable to the average Joe while Linux is restricted to nerds (who think it's cool that Linux is incomprehensible to the "losers" who can't configure an inetd).
Funnily enough, I used to hate Microsoft as much as anyone when I started reading Slashdot, but Slashdot's utterly biased coverage, and the idiocy of the Linux zealots, has given me new respect for the organization.
--
It's a
-- Danny Vermin
I believe he's referring to the fact that he's either some type of tech support or in the security field. Think about it:
:-)
M$ bugs + lusers = $$$ for us
I agree, let them be. Us geeks need to eat too.