Abolade Gbadegesin and team - Networking in Windows Vista

This is great stuff guys.  When I ran the McAfee Personal Firewall group we came up to Seattle in 2003 to talk about what we wanted to see in the stack to make implementing our products easier.  Great to see not only that this has all worked out with the improvements over TDI, but the performance improvements as well.

So much stuff we used to have to do ourselves, now baked in.  Very cool, but I remain glad I got out of that line of business

Is the deep stuff going to be exposed to .NET?  Will it be possible to write complex security apps solely in managed code?

That sounds really great guys!  Thank you.  The team seems like a bunch of great folks.  I have some questions on network stack regarding server side tech to allow easier server side code to be developed that will scale.  This may be more socket and OS related, but I am sure it is all intermixed. 

Today, to write a socket server, you pretty much have two models.  You can have a thread per socket model which is easy to write, but does not (or can not) scale much beyond 1000 threads because of windows limitations and RAM.  Fibers, may help here, but not sure, maybe some talk on that option.  The other option is an Async server that handles IO with async reads and writes.  This scales well, but is difficult to write and more difficult to prove its correctness and debug because of its disjointed nature (i.e. its hard).  Another model that begs to work is to add sockets to a Wait event, so the event will signal when some IO is on a socket.  This does not work beyond 64 objects and has some fairness issues IIRC, because of some limits in WaitForMultipleObjects.  

It would seem like a simplier model may be to have some native SocketQueue that I could add thousands of sockets (if needed) to and be "called back" in an efficent manner for reads and writes.  Maybe you handle the thread pool or we can spin up X number of threads that Wait for reads on the SocketQueue.  Then I can still do my logic in some Syncronous way, but the system handles the async IO.  Hope that makes some sense.  Any work in this area?  TIA
--
William Stacey [MVP] 

One of the best interviews i've seen, these seem like a greate bunch of gyes that realy get into there work and have greate personalities. There is nothing more boring than when you interview someone with no sens of humour, not the case here, greate buch of gyes.

 

I haven't finished watching yet, but a question about the comment about speed up on a DSL link. You would only see that improvement if you were directly connected to the internet and not connecting through a NAT/firewall ?

Are there any other OS's that support this change in the Transmission Control algorithim? Is it a MS only algorithim? Is there any possability that this will be back ported to XP?

Sounds similar to the FAST improvements to TCP, getting rid of  the overly cautious doubling of the delay when acknowledgements are late. Don't know if it's the same thing. Apparently , while TCP has max 25% throughput, FAST can achieve 95% efficiency.

http://netlab.caltech.edu/

Is the new firewall going to be able to take advantage of integrated hardware firewall processors, like the nForce firewall?

Great video. Best combination of accents in one room...

You may want to take a look at the IO Completion port architecture – that will let you work with thousands of sockets on very few threads and the programming model is quite intuitive.

Thanks.

My previous response was to staceyw.

 

Tragic_hip: if the firewall/NAT is doing IP level forwarding/natting ONLY, then the improvements will work through them as well. On the other hand, if that device is doing split TCP connection, it may not.

 

The new congestion control algorithm is Microsoft’s intellectual property. We have not explored the possibility of backporting this to XP yet.

 

MyCroft: This is not FAST – it is called CTCP (Compound TCP).

Another excellent video! . Im dying to get my hands on Vista now!

Great Video -
Scooble - Great Interview,

The only suggestio I have is that since channel9 is dev oriented, you should minimize the "what does it give to my mom" questions.

I disagree. I love the, "What about my Mom?" It brings us back to the reality of what software is supposed to do on a large scale. Technology isn't just for geeks. I actually do think about whether or not new technology will make my Mom's life better or harder. Or, to put it another way, the number of support calls I anticipate from my Mom directly impacts my decision on whether or not to to upgrade her from one rev of software to another.

"You may want to take a look at the IO Completion port architecture – that will let you work with thousands of sockets on very few threads and the programming model is quite intuitive.
"

Thanks.  How is that exposed in .Net classes?

Programcsharp – The default firewall in VISTA is not integrated with any of the hardware firewall processors. However, VISTA provides a platform/API for 3^rd party firewalls to plug into the system – and any maker of a hardware firewall could use this to do hardware based firewalling.

 

Staceyw – I don’t’ think completion-ports are exposed as a programming primitive in .NET classes, but I’ll have to double check though.

<sorry about the formatting hiccup with the previous post, correcting it>

Programcsharp – The default firewall in VISTA is not integrated with any of the hardware firewall processors. However, VISTA provides a platform/API for 3^rd party firewalls to plug into the system – and any maker of a hardware firewall could use this to do hardware based firewalling.

 

Staceyw – I don’t’ think completion-ports are exposed as a programming primitive in .NET classes, but I’ll have to double check though.

Tell me about the speed improvements of file copying across a broadband link (PPTP, or direct connections to a Windows box on the internet with it's port 445 open).

If i'm VPN'ed to work, and i want to transfer 100MB: it is faster for me to Email it, or to copy the file to a web-server, or copy the file to an FTP server.

Isn't it time to improve the file sharing protocol, or to add a new file sharing protocol altogether? i could listen to an entire MP3 over an RDP link, faster than i could copy that same MP3. i know that SourceSafe2005 comes with a new 'file sharing service', so speed up file copying.

What is Vista doing for me in this area?

Usually, Scoble laughs uncontrollably and makes one wonder about the level of his IQ so I'm glad to see Scoble isn't horking like an oversexed pig in this video.
The root of Scoble's loud and abrupt laughter is probably derived from various factors of happiness and one way to achieve a state of happiness is to become immersed in ignorance.
However, there could also be a mental factor, such as brain damage caused by an excessive diet containing pork and/or alcohol.
Scoble, please keep the laughs down in your videos.

That was a fantastic interview, although generally I always enjoy Robert's interviews - he is so animated and communicates enthusiasm so well with others.

The *Nix guys have always in the past criticised the networking technology in Windows and clearly this new work is going to give them good reason to reconsider. It will be interesting to see the third party verifications of the many changes to networking.

I am truely astounded about the number of milestones being set, across the board, in the Vista technology timeframe and see it all as a real coming of age for Microsoft.

This team seems like a great bunch of guys who are really focused on both the developer side of things and the IT pro side of things and also the general users. Great to see such a broad consideration. Fantastic.