Re: Questions for OpenBGPd Developers
Hi, On Tue, 13.10.2009 at 16:41:35 +0200, Igor Sobrado igor.sobr...@gmail.com wrote: requirements come first, then you can choose the best tools to get that work done, not the reverse. why is it so difficult to understand? well... short story: Your definition of better may or may not meet my definition of better, for a large number of reasons. Simple example: I've long wanted to see ISDN support in OpenBSD, but it just has not happened in a number of years (only stating the facts here). So, if I need ISDN + something in one box, OpenBSD is immediately out of the question, and no, external ISDN modems, if still available, don't cut it. See the point? Kind regards, --Toni++ PS: I'm aware of MirOS, but hadn't too much luck last I tried.
Re: Questions for OpenBGPd Developers
Hi, On Tue, 13.10.2009 at 11:33:40 -0400, and...@msu.edu and...@msu.edu wrote: The problem with this is that the interface between the other OS and the OpenBSD based code needs to be correct and secure, else there will be bugs and people will complain that OpenBSD code isn't good, etc and in general, snipe. I expect people who are looking to build a home grown BGP router to be smarter than that. I just don't see the need to move the bgp code to another system, myself. Better hardware support could be an issue. Not everyone can, or wants to, talk over Ethernet, esp. with a BGP router... Kind regards, --Toni++
Re: Questions for OpenBGPd Developers
On Tue, 13 Oct 2009 02:12:04 +0200 Henning Brauer lists-open...@bsws.de wrote: and there's a reason why it is that way - I always found the idea of making a bgp router out of a common unix box by adding a userland bgp speaker only flawed. many things can only properly or at all be done at kernel level or with kernel support. I guess that's apply to OpenOSPF and RIP too, right? Cheers -- Massimo
Re: Questions for OpenBGPd Developers
On Tue, Oct 13, 2009 at 09:59:08AM +0200, Massimo Lusetti wrote: On Tue, 13 Oct 2009 02:12:04 +0200 Henning Brauer lists-open...@bsws.de wrote: and there's a reason why it is that way - I always found the idea of making a bgp router out of a common unix box by adding a userland bgp speaker only flawed. many things can only properly or at all be done at kernel level or with kernel support. I guess that's apply to OpenOSPF and RIP too, right? Yes. Our kroute.c code uses many OpenBSD only features that are hard to work around on other systems. -- :wq Claudio
Re: Questions for OpenBGPd Developers
On Mon, Oct 12, 2009 at 05:50:58PM -0700, Barry Friedman wrote: Hi, thanks everyone for the information, this helps give me an idea of the scope and effort involved in getting OpenBGPd onto Linux. I'll look at the OpenSSH project to see how the portability is added without cluttering up the OpenBSD code. Also I am sorry, I did not mean to imply that OpenBGPd is not in a source control system or released frequently. I was referring to the quick and dirty Linux port I mentioned which is just in tarball form. Kudos to those who did that porting work because it allows Linux users to at least play around with OpenBGPd a bit but I was just trying to see if there was a more organized and source-controlled effort yet to work on OpenBGPd porting to non-BSD systems. I don't see the point in porting this to linux. Why settle for second-best ? ... don't bother answering, it's an obvious troll, but I guess it summarizes that none of us would seriously consider running this kind of critical network stuff on anything but OpenBSD.
Re: Questions for OpenBGPd Developers
On Tue, Oct 13, 2009 at 3:13 PM, Marc Espie es...@nerim.net wrote: On Mon, Oct 12, 2009 at 05:50:58PM -0700, Barry Friedman wrote: Hi, thanks everyone for the information, this helps give me an idea of the scope and effort involved in getting OpenBGPd onto Linux. I'll look at the OpenSSH project to see how the portability is added without cluttering up the OpenBSD code. Also I am sorry, I did not mean to imply that OpenBGPd is not in a source control system or released frequently. I was referring to the quick and dirty Linux port I mentioned which is just in tarball form. Kudos to those who did that porting work because it allows Linux users to at least play around with OpenBGPd a bit but I was just trying to see if there was a more organized and source-controlled effort yet to work on OpenBGPd porting to non-BSD systems. I don't see the point in porting this to linux. Why settle for second-best ? Uhm perhaps to provide a better OSPF and BGP implementation to the for an OS that is the OS of choice of millions of users and thousands of corporations? -- Opportunity is most often missed by people because it is dressed in overalls and looks like work. Thomas Alva Edison Inventor of 1093 patents, including: The light bulb, phonogram and motion pictures.
Re: Questions for OpenBGPd Developers
On Tue, Oct 13, 2009 at 04:12:40PM +0200, Ross Cameron wrote: | I don't see the point in porting this to linux. Why settle for second-best | ? | | | Uhm perhaps to provide a better OSPF and BGP implementation to the for an OS | that is the OS of choice of millions of users and thousands of corporations? Why don't they pick the better OS to run that better OSPF and BGP implementation ? They don't pick the better OS, why would they pick the better, but then stripped down routing daemons ? (no need to answer that, it's a rhetorical question) Paul 'WEiRD' de Weerd -- [++-]+++.+++[---].+++[+ +++-].++[-]+.--.[-] http://www.weirdnet.nl/
Re: Questions for OpenBGPd Developers
On Tue, Oct 13, 2009 at 04:12:40PM +0200, Ross Cameron wrote: On Tue, Oct 13, 2009 at 3:13 PM, Marc Espie es...@nerim.net wrote: On Mon, Oct 12, 2009 at 05:50:58PM -0700, Barry Friedman wrote: Hi, thanks everyone for the information, this helps give me an idea of the scope and effort involved in getting OpenBGPd onto Linux. I'll look at the OpenSSH project to see how the portability is added without cluttering up the OpenBSD code. Also I am sorry, I did not mean to imply that OpenBGPd is not in a source control system or released frequently. I was referring to the quick and dirty Linux port I mentioned which is just in tarball form. Kudos to those who did that porting work because it allows Linux users to at least play around with OpenBGPd a bit but I was just trying to see if there was a more organized and source-controlled effort yet to work on OpenBGPd porting to non-BSD systems. I don't see the point in porting this to linux. Why settle for second-best ? Uhm perhaps to provide a better OSPF and BGP implementation to the for an OS that is the OS of choice of millions of users and thousands of corporations? Why would the OpenBSD developers be interested in that? Anyone who wants to use it on Linux is free to port it, but the typical OpenBSD developer isn't *that* interested in running stuff on Linux. Joachim
Re: Questions for OpenBGPd Developers
On Tue, Oct 13, 2009 at 04:12:40PM +0200, Ross Cameron wrote: On Tue, Oct 13, 2009 at 3:13 PM, Marc Espie es...@nerim.net wrote: On Mon, Oct 12, 2009 at 05:50:58PM -0700, Barry Friedman wrote: Hi, thanks everyone for the information, this helps give me an idea of the scope and effort involved in getting OpenBGPd onto Linux. I'll look at the OpenSSH project to see how the portability is added without cluttering up the OpenBSD code. Also I am sorry, I did not mean to imply that OpenBGPd is not in a source control system or released frequently. I was referring to the quick and dirty Linux port I mentioned which is just in tarball form. Kudos to those who did that porting work because it allows Linux users to at least play around with OpenBGPd a bit but I was just trying to see if there was a more organized and source-controlled effort yet to work on OpenBGPd porting to non-BSD systems. I don't see the point in porting this to linux. Why settle for second-best ? Uhm perhaps to provide a better OSPF and BGP implementation to the for an OS that is the OS of choice of millions of users and thousands of corporations? You don't get it, do you ? There's no better OSPF and BGP on false premises, like a flawed platform. Go on, tell people to run OSPF on Windows 7, and see linux weenies cringe. See my point ? (again, troll. don't bother to read except for entertainment values)
Re: Questions for OpenBGPd Developers
On Tue, Oct 13, 2009 at 4:12 PM, Ross Cameron abal...@gmail.com wrote: Uhm perhaps to provide a better OSPF and BGP implementation to the for an OS that is the OS of choice of millions of users and thousands of corporations? users and corporations should learn how to choose the operating systems that best fit their needs instead of choosing the coolest operating system of the day and adapt it to match their real needs. requirements come first, then you can choose the best tools to get that work done, not the reverse. why is it so difficult to understand?
Re: Questions for OpenBGPd Developers
On Tue, Oct 13, 2009 at 04:12:40PM +0200, Ross Cameron wrote: On Tue, Oct 13, 2009 at 3:13 PM, Marc Espie es...@nerim.net wrote: On Mon, Oct 12, 2009 at 05:50:58PM -0700, Barry Friedman wrote: Hi, thanks everyone for the information, this helps give me an idea of the scope and effort involved in getting OpenBGPd onto Linux. I'll look at the OpenSSH project to see how the portability is added without cluttering up the OpenBSD code. Also I am sorry, I did not mean to imply that OpenBGPd is not in a source control system or released frequently. I was referring to the quick and dirty Linux port I mentioned which is just in tarball form. Kudos to those who did that porting work because it allows Linux users to at least play around with OpenBGPd a bit but I was just trying to see if there was a more organized and source-controlled effort yet to work on OpenBGPd porting to non-BSD systems. I don't see the point in porting this to linux. Why settle for second-best ? Uhm perhaps to provide a better OSPF and BGP implementation to the for an OS that is the OS of choice of millions of users and thousands of corporations? So you want to port OpenBGPD to Windows? Since that OS has the biggest market share and all managers are all for it. Windows as a core router - shudder. I think most people on this list don't care about the million other users out their with their fancy colorful OS. I for sure don't. -- :wq Claudio
Re: Questions for OpenBGPd Developers
On Tue, Oct 13, 2009 at 4:41 PM, Igor Sobrado igor.sobr...@gmail.comwrote: On Tue, Oct 13, 2009 at 4:12 PM, Ross Cameron abal...@gmail.com wrote: Uhm perhaps to provide a better OSPF and BGP implementation to the for an OS that is the OS of choice of millions of users and thousands of corporations? users and corporations should learn how to choose the operating systems that best fit their needs instead of choosing the coolest operating system of the day and adapt it to match their real needs. requirements come first, then you can choose the best tools to get that work done, not the reverse. why is it so difficult to understand? More often than not there are more reasons than the purely technical motivations for different tools/technologies being used. Either way this is becoming an off topic OS flame war, personally I see a portability layer for OpenBGP (as with the portability layer for OpenSSH) as being a good thing. It doesn't taint the OpenBSD sources and those that have a need to use it on X platform can. -- Opportunity is most often missed by people because it is dressed in overalls and looks like work. Thomas Alva Edison Inventor of 1093 patents, including: The light bulb, phonogram and motion pictures.
Re: Questions for OpenBGPd Developers
On Tue, Oct 13, 2009 at 04:41:35PM +0200, Igor Sobrado wrote: On Tue, Oct 13, 2009 at 4:12 PM, Ross Cameron abal...@gmail.com wrote: Uhm perhaps to provide a better OSPF and BGP implementation to the for an OS that is the OS of choice of millions of users and thousands of corporations? users and corporations should learn how to choose the operating systems that best fit their needs instead of choosing the coolest operating system of the day and adapt it to match their real needs. requirements come first, then you can choose the best tools to get that work done, not the reverse. why is it so difficult to understand? many corporations think that it is better to run products based on something that is very common on the market and you easily find cheap staff working on it; especially if you out-source the development or operation to countries like India or China. so from a management perspective, you do not depend on single developers because you can replace everyone with someone else from the global linux pool, it takes only two weeks to implement a new major feature, and you can drastically reduce your development costs and time-to-market. this will improve your profit, because you will not lower the price of the products, of course. and afterwards you spend two to four years with fixing the bugs of the supposedly working products if you do not get bankrupt or sold, discontinued, or replaced the products before. but you don't have to care if you itend to sell your VC-funded company within the next two years anyways. sorry for the rant, we all like each other, don't we? reyk
Re: Questions for OpenBGPd Developers
Quoting Ross Cameron ross.came...@linuxpro.co.za: On Tue, Oct 13, 2009 at 4:41 PM, Igor Sobrado igor.sobr...@gmail.comwrote: On Tue, Oct 13, 2009 at 4:12 PM, Ross Cameron abal...@gmail.com wrote: Uhm perhaps to provide a better OSPF and BGP implementation to the for an OS that is the OS of choice of millions of users and thousands of corporations? users and corporations should learn how to choose the operating systems that best fit their needs instead of choosing the coolest operating system of the day and adapt it to match their real needs. requirements come first, then you can choose the best tools to get that work done, not the reverse. why is it so difficult to understand? More often than not there are more reasons than the purely technical motivations for different tools/technologies being used. Either way this is becoming an off topic OS flame war, personally I see a portability layer for OpenBGP (as with the portability layer for OpenSSH) as being a good thing. It doesn't taint the OpenBSD sources and those that have a need to use it on X platform can. The problem with this is that the interface between the other OS and the OpenBSD based code needs to be correct and secure, else there will be bugs and people will complain that OpenBSD code isn't good, etc and in general, snipe. I just don't see the need to move the bgp code to another system, myself. --STeve Andre'
Re: Questions for OpenBGPd Developers
On Tue, Oct 13, 2009 at 7:47 AM, Claudio Jeker cje...@diehard.n-r-g.com wrote: On Tue, Oct 13, 2009 at 04:12:40PM +0200, Ross Cameron wrote: On Tue, Oct 13, 2009 at 3:13 PM, Marc Espie es...@nerim.net wrote: On Mon, Oct 12, 2009 at 05:50:58PM -0700, Barry Friedman wrote: Hi, thanks everyone for the information, this helps give me an idea of the scope and effort involved in getting OpenBGPd onto Linux. I'll look at the OpenSSH project to see how the portability is added without cluttering up the OpenBSD code. I don't see the point in porting this to linux. Why settle for second-best ? Uhm perhaps to provide a better OSPF and BGP implementation to the for an OS that is the OS of choice of millions of users and thousands of corporations? So you want to port OpenBGPD to Windows? Since that OS has the biggest market share and all managers are all for it. Windows as a core router - shudder. I think most people on this list don't care about the million other users out their with their fancy colorful OS. I for sure don't. In my case it has nothing to do with whether or not millions of people use a particular OS but simply that I am constrained to Linux for this project and it is non-negotiable. So I could use that as an excuse to ignore OpenBGPd but I think it's a nice BGP implementation and I think it may be a nice implementation on Linux too, or maybe not, that is the reason for my questions here, which have been answered very well by Claudio and Henning and others. Thanks, Barry
Re: Questions for OpenBGPd Developers
In my case it has nothing to do with whether or not millions of people use a particular OS but simply that I am constrained to Linux for this project and it is non-negotiable. So I could use that as an excuse to ignore OpenBGPd but I think it's a nice BGP implementation and I think it may be a nice implementation on Linux too, or maybe not, that is the reason for my questions here, which have been answered very well by Claudio and Henning and others. The message has been misunderstood. Let me be clear, here are the two messages: Firstly, Linux has a routing socket and in-kernel routing table which lacks the right capabilities, so you won't get all you want. Secondly, Henning and Claudio don't do slave labour.
Re: Questions for OpenBGPd Developers
Cool. Thanks for the excellent info. Regards, Barry On Tue, Oct 13, 2009 at 12:14 PM, Theo de Raadt dera...@cvs.openbsd.org wrote: In my case it has nothing to do with whether or not millions of people use a particular OS but simply that I am constrained to Linux for this project and it is non-negotiable. So I could use that as an excuse to ignore OpenBGPd but I think it's a nice BGP implementation and I think it may be a nice implementation on Linux too, or maybe not, that is the reason for my questions here, which have been answered very well by Claudio and Henning and others. The message has been misunderstood. Let me be clear, here are the two messages: Firstly, Linux has a routing socket and in-kernel routing table which lacks the right capabilities, so you won't get all you want. Secondly, Henning and Claudio don't do slave labour.
Questions for OpenBGPd Developers
Hi, I'm looking at open source BGP implementations for possible use on a Linux-based system. I'm interested in OpenBGPd and I understand that it's part of OpenBSD but in the application that I'm looking at the only choice I am given is Linux. My questions are: Are there any projects to create an OpenBGPd implementation that is portable and separate from OpenBSD? I'm aware of quick-n-dirty ports like http://wiki.version6.net/openbgpd but was wondering more formal projects to create a full-featured and portable implementation that is more like a true open source project (ie. on a public source control server with open forums, mailing lists, bug reporting, frequent releases, etc). Is the current OpenBGPd development team an open team in which anyone may contribute patches, or is it a closed team? If there was a porting effort, could the changes be incorporated into the existing project, or would a portable OpenBGPd need to be a separate project? At this stage I'm just gathering information and answers to the above questions would be very helpful. Thanks, Barry
Re: Questions for OpenBGPd Developers
Hi, it only works on OpenBSD and any efforts to port it to FreeBSD or Linux weren't really successful. The reason is that OpenBSD's routing daemons heavilly utilize the kernel's routing stack that has many interfaces and features that are not available in and is not compatible to other OSes. You end up with a half-working daemon missing many features and lots of #if 0's in the code. reyk On Mon, Oct 12, 2009 at 10:40:14AM -0700, Barry Friedman wrote: Hi, I'm looking at open source BGP implementations for possible use on a Linux-based system. I'm interested in OpenBGPd and I understand that it's part of OpenBSD but in the application that I'm looking at the only choice I am given is Linux. My questions are: Are there any projects to create an OpenBGPd implementation that is portable and separate from OpenBSD? I'm aware of quick-n-dirty ports like http://wiki.version6.net/openbgpd but was wondering more formal projects to create a full-featured and portable implementation that is more like a true open source project (ie. on a public source control server with open forums, mailing lists, bug reporting, frequent releases, etc). Is the current OpenBGPd development team an open team in which anyone may contribute patches, or is it a closed team? If there was a porting effort, could the changes be incorporated into the existing project, or would a portable OpenBGPd need to be a separate project? At this stage I'm just gathering information and answers to the above questions would be very helpful. Thanks, Barry
Re: Questions for OpenBGPd Developers
On Mon, Oct 12, 2009 at 08:12:41PM +0200, Reyk Floeter wrote: Hi, it only works on OpenBSD and any efforts to port it to FreeBSD or Linux weren't really successful. The reason is that OpenBSD's routing daemons heavilly utilize the kernel's routing stack that has many interfaces and features that are not available in and is not compatible to other OSes. You end up with a half-working daemon missing many features and lots of #if 0's in the code. While true this is mostly the reason because nobody from outside came towards us trying to come up with a portable OpenBGPD similar to the one done for OpenSSH. Neither I nor Henning have the time and nerves to fiddle around with other systems routing stack (our own is already painful enough) I doubt that we will commit needed platform glue for other OSs to our tree but it should be possible to pack a special OpenBGPD tarball with the needed glue. It comes down to replace 4 or so files (kroute.c, carp.c, pfkey.c, pftable.c) with replacement implementations. The rest should be OS independent. Some features will be impossible to implement on some systems though. On Mon, Oct 12, 2009 at 10:40:14AM -0700, Barry Friedman wrote: Hi, I'm looking at open source BGP implementations for possible use on a Linux-based system. I'm interested in OpenBGPd and I understand that it's part of OpenBSD but in the application that I'm looking at the only choice I am given is Linux. My questions are: Are there any projects to create an OpenBGPd implementation that is portable and separate from OpenBSD? I'm aware of quick-n-dirty ports like http://wiki.version6.net/openbgpd but was wondering more formal projects to create a full-featured and portable implementation that is more like a true open source project (ie. on a public source control server with open forums, mailing lists, bug reporting, frequent releases, etc). Is the current OpenBGPd development team an open team in which anyone may contribute patches, or is it a closed team? If there was a porting effort, could the changes be incorporated into the existing project, or would a portable OpenBGPd need to be a separate project? At this stage I'm just gathering information and answers to the above questions would be very helpful. Thanks, Barry -- :wq Claudio
Re: Questions for OpenBGPd Developers
* Claudio Jeker cje...@diehard.n-r-g.com [2009-10-13 00:09]: While true this is mostly the reason because nobody from outside came towards us trying to come up with a portable OpenBGPD similar to the one done for OpenSSH. Neither I nor Henning have the time and nerves to fiddle around with other systems routing stack (our own is already painful enough) I doubt that we will commit needed platform glue for other OSs to our tree but it should be possible to pack a special OpenBGPD tarball with the needed glue. It comes down to replace 4 or so files (kroute.c, carp.c, pfkey.c, pftable.c) with replacement implementations. The rest should be OS independent. Some features will be impossible to implement on some systems though. and there's a reason why it is that way - I always found the idea of making a bgp router out of a common unix box by adding a userland bgp speaker only flawed. many things can only properly or at all be done at kernel level or with kernel support. -- Henning Brauer, h...@bsws.de, henn...@openbsd.org BS Web Services, http://bsws.de Full-Service ISP - Secure Hosting, Mail and DNS Services Dedicated Servers, Rootservers, Application Hosting - Hamburg Amsterdam
Re: Questions for OpenBGPd Developers
* Barry Friedman friedman.ba...@gmail.com [2009-10-12 19:52]: Hi, I'm looking at open source BGP implementations for possible use on a Linux-based system. I'm interested in OpenBGPd and I understand that it's part of OpenBSD but in the application that I'm looking at the only choice I am given is Linux. My questions are: Are there any projects to create an OpenBGPd implementation that is portable and separate from OpenBSD? I'm aware of quick-n-dirty ports like http://wiki.version6.net/openbgpd but was wondering more formal projects to create a full-featured and portable implementation that is more like a true open source project (ie. on a public source control server with open forums, mailing lists, bug reporting, frequent releases, etc). I am not aware of any portable efforts. I portable would have to be done OpenSSH-style. We don't want the clutter in our tree. the code already is in a publich source control server (any openbsd anoncvs has it), the openbsd mailing lists suffice, as does the openbsd PR database. and there's a release every 6 months, pretty frequent i'd say. Is the current OpenBGPd development team an open team in which anyone may contribute patches, or is it a closed team? patches are welcome from pretty much anybody. in fact, if you read the changelog you'll find a lot of diffs from people withour openbsd cvs accounts committed by claudio or me. If there was a porting effort, could the changes be incorporated into the existing project, or would a portable OpenBGPd need to be a separate project? openssh style -- Henning Brauer, h...@bsws.de, henn...@openbsd.org BS Web Services, http://bsws.de Full-Service ISP - Secure Hosting, Mail and DNS Services Dedicated Servers, Rootservers, Application Hosting - Hamburg Amsterdam
Re: Questions for OpenBGPd Developers
Hi, thanks everyone for the information, this helps give me an idea of the scope and effort involved in getting OpenBGPd onto Linux. I'll look at the OpenSSH project to see how the portability is added without cluttering up the OpenBSD code. Also I am sorry, I did not mean to imply that OpenBGPd is not in a source control system or released frequently. I was referring to the quick and dirty Linux port I mentioned which is just in tarball form. Kudos to those who did that porting work because it allows Linux users to at least play around with OpenBGPd a bit but I was just trying to see if there was a more organized and source-controlled effort yet to work on OpenBGPd porting to non-BSD systems. Regards, Barry On Mon, Oct 12, 2009 at 5:09 PM, Henning Brauer lists-open...@bsws.de wrote: * Barry Friedman friedman.ba...@gmail.com [2009-10-12 19:52]: Hi, I'm looking at open source BGP implementations for possible use on a Linux-based system. I'm interested in OpenBGPd and I understand that it's part of OpenBSD but in the application that I'm looking at the only choice I am given is Linux. My questions are: Are there any projects to create an OpenBGPd implementation that is portable and separate from OpenBSD? I'm aware of quick-n-dirty ports like http://wiki.version6.net/openbgpd but was wondering more formal projects to create a full-featured and portable implementation that is more like a true open source project (ie. on a public source control server with open forums, mailing lists, bug reporting, frequent releases, etc). I am not aware of any portable efforts. I portable would have to be done OpenSSH-style. We don't want the clutter in our tree. the code already is in a publich source control server (any openbsd anoncvs has it), the openbsd mailing lists suffice, as does the openbsd PR database. and there's a release every 6 months, pretty frequent i'd say. Is the current OpenBGPd development team an open team in which anyone may contribute patches, or is it a closed team? patches are welcome from pretty much anybody. in fact, if you read the changelog you'll find a lot of diffs from people withour openbsd cvs accounts committed by claudio or me. If there was a porting effort, could the changes be incorporated into the existing project, or would a portable OpenBGPd need to be a separate project? openssh style -- Henning Brauer, h...@bsws.de, henn...@openbsd.org BS Web Services, http://bsws.de Full-Service ISP - Secure Hosting, Mail and DNS Services Dedicated Servers, Rootservers, Application Hosting - Hamburg Amsterdam