|
View:
New views
8 Messages
—
Rating Filter:
Alert me
|
|
|
PPTP move to a kernel module?Hi,
I've been messing about a bit, and having spent some time working on a box with a high amount of connections. Kinda made me want to look at a better implementation of PPTP to have a high capacity high reliable solution. I was thinking if it would be possible to create a kernel module to handle the data between the Net and PPPd directly. This was not having to send the PPTP from kernel to user space to basically just encapsulate / decapsulate and then send back to PPPd Kind of the same way PPPoE handles its connections. If this was possible we would perhaps no longer need 2 processes per session, perhaps 1 main controlling daemon. Using a PPP module to interface to the kernel (same way rp-pppoe.so) ? I'm not sure if this is a option, perhaps the developers could comment? Thanks, Mike ------------------------------------------------------------------------------ Stay on top of everything new and different, both inside and around Java (TM) technology - register by April 22, and save $200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco. 300 plus technical and hands-on sessions. Register today. Use priority code J9JMT32. http://p.sf.net/sfu/p _______________________________________________ Poptop-server mailing list Poptop-server@... https://lists.sourceforge.net/lists/listinfo/poptop-server |
|
|
Re: PPTP move to a kernel module?On Wed, Apr 22, 2009 at 08:41:43PM +0200, Michael da Silva Pereira wrote:
> I was thinking if it would be possible to create a kernel module to > handle the data between the Net and PPPd directly. Been done. http://accel-pptp.sourceforge.net/ Latest release 2009-01-27. Perhaps you'd like to push it upstream into the Linux kernel? -- James Cameron http://quozl.netrek.org/ HP Open Source, Volunteer http://opensource.hp.com/ PPTP Client Project, Release Engineer http://pptpclient.sourceforge.net/ ------------------------------------------------------------------------------ Stay on top of everything new and different, both inside and around Java (TM) technology - register by April 22, and save $200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco. 300 plus technical and hands-on sessions. Register today. Use priority code J9JMT32. http://p.sf.net/sfu/p _______________________________________________ Poptop-server mailing list Poptop-server@... https://lists.sourceforge.net/lists/listinfo/poptop-server |
|
|
Re: PPTP move to a kernel module?On Thu, 23 Apr 2009, James Cameron wrote: > On Wed, Apr 22, 2009 at 08:41:43PM +0200, Michael da Silva Pereira wrote: > > I was thinking if it would be possible to create a kernel module to > > handle the data between the Net and PPPd directly. > > Been done. http://accel-pptp.sourceforge.net/ Latest release > 2009-01-27. Looks like it might not be quite production ready. > Perhaps you'd like to push it upstream into the Linux kernel? Sounds a good move. --- Charlie ------------------------------------------------------------------------------ Stay on top of everything new and different, both inside and around Java (TM) technology - register by April 22, and save $200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco. 300 plus technical and hands-on sessions. Register today. Use priority code J9JMT32. http://p.sf.net/sfu/p _______________________________________________ Poptop-server mailing list Poptop-server@... https://lists.sourceforge.net/lists/listinfo/poptop-server |
|
|
|
|
|
Re: PPTP move to a kernel module?On Thu, Apr 23, 2009 at 09:34:18AM -0300, Bruno@... wrote:
> Is anyone from the poptop project involved? I'm not involved yet. I don't know who else is involved, perhaps the commit history or the source attribution could tell you? -- James Cameron http://quozl.netrek.org/ HP Open Source, Volunteer http://opensource.hp.com/ PPTP Client Project, Release Engineer http://pptpclient.sourceforge.net/ ------------------------------------------------------------------------------ Crystal Reports - New Free Runtime and 30 Day Trial Check out the new simplified licensign option that enables unlimited royalty-free distribution of the report engine for externally facing server and web deployment. http://p.sf.net/sfu/businessobjects _______________________________________________ Poptop-server mailing list Poptop-server@... https://lists.sourceforge.net/lists/listinfo/poptop-server |
|
|
Re: PPTP move to a kernel module?В Чтв, 23/04/2009 в 08:33 -0400, Charlie Brady пишет:
> On Thu, 23 Apr 2009, James Cameron wrote: > > http://accel-pptp.sourceforge.net/ Latest release 2009-01-27. > > Looks like it might not be quite production ready. Yes, it looks like that, but I'm using it in production on two servers and for more then year with 300-500 concurrent connections each they oopsed only twice and regretfully I failed to gather oops log. That's said now I'm keeping both servers with netconsole running so I hope to gather backtrace at some point of time... Anyway it works really nice: uptime/processor usage and most important for us ping delays are really small. If you are Gentoo user you may try ebuild from my overlay: svn co http://overlays.gentoo.org/svn/dev/pva/net-dialup/accel-pptp and just emerge accel-pptp :) -- Peter. ------------------------------------------------------------------------------ The NEW KODAK i700 Series Scanners deliver under ANY circumstances! Your production scanning environment may not be a perfect world - but thanks to Kodak, there's a perfect scanner to get the job done! With the NEW KODAK i700 Series Scanner you'll get full speed at 300 dpi even with all image processing features enabled. http://p.sf.net/sfu/kodak-com _______________________________________________ Poptop-server mailing list Poptop-server@... https://lists.sourceforge.net/lists/listinfo/poptop-server |
|
|
Re: PPTP move to a kernel module?On Tue, May 05, 2009 at 07:03:57PM +0400, Peter Volkov wrote:
> Yes, it looks like that, but I'm using it in production on two servers > and for more then year with 300-500 concurrent connections each they > oopsed only twice and regretfully I failed to gather oops log. That's > said now I'm keeping both servers with netconsole running so I hope to > gather backtrace at some point of time... Sounds good. Hope you catch it. Don't forget the possibility that the Oops is in entirely different code, or induced by it. > Anyway it works really nice: uptime/processor usage and most important > for us ping delays are really small. Can you be more precise about those numbers? I'm not being defensive, I'd love to see PPTP move to the kernel, but it will happen more easily if there are hard reproducible performance improvements. -- James Cameron http://quozl.netrek.org/ HP Open Source, Volunteer http://opensource.hp.com/ PPTP Client Project, Release Engineer http://pptpclient.sourceforge.net/ ------------------------------------------------------------------------------ The NEW KODAK i700 Series Scanners deliver under ANY circumstances! Your production scanning environment may not be a perfect world - but thanks to Kodak, there's a perfect scanner to get the job done! With the NEW KODAK i700 Series Scanner you'll get full speed at 300 dpi even with all image processing features enabled. http://p.sf.net/sfu/kodak-com _______________________________________________ Poptop-server mailing list Poptop-server@... https://lists.sourceforge.net/lists/listinfo/poptop-server |
|
|
Re: PPTP move to a kernel module?Hi James. Sorry for long delay, I haven't noticed you mail earlier.
В Срд, 06/05/2009 в 08:49 +1000, James Cameron пишет: > > Anyway it works really nice: uptime/processor usage and most important > > for us ping delays are really small. > > Can you be more precise about those numbers? Right now I have 360 clients running and channel is loaded 80 Mbit/s. At the same time my Intel(R) Xeon(R) CPU 3040 @ 1.86GHz (with hyper-threading enabled) system loads system like this: vpn-1 ~ # vmstat 1 procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu---- r b swpd free buff cache si so bi bo in cs us sy id wa 1 0 0 350552 2076 401976 0 0 0 4 3 0 0 13 86 0 0 0 0 350544 2076 401976 0 0 0 0 22978 1693 0 24 76 0 The most important improvement we got was ping delay. Previously it jumped from 2 to 40-60 miliseconds. Right now it's just 2-4ms. > I'm not being defensive, I'd love to see PPTP move to the kernel, but > it will happen more easily if there are hard reproducible performance > improvements. Well in my case it's not easy task to do better tests but probably it's really interesting to compare nubmers better. If I'll managed to do tests there I'll mail results here. -- Peter. ------------------------------------------------------------------------------ Register Now for Creativity and Technology (CaT), June 3rd, NYC. CaT is a gathering of tech-side developers & brand creativity professionals. Meet the minds behind Google Creative Lab, Visual Complexity, Processing, & iPhoneDevCamp asthey present alongside digital heavyweights like Barbarian Group, R/GA, & Big Spaceship. http://www.creativitycat.com _______________________________________________ Poptop-server mailing list Poptop-server@... https://lists.sourceforge.net/lists/listinfo/poptop-server |
| Free embeddable forum powered by Nabble | Forum Help |