Re: [JDEV] The Important Things
I think one thing we could do is go to the ISPs with an open book and say What would you want from an IM service?, don't say yes or no to anything (except the things we already have), and then start forming a to-do list from it. The list from Mike Lin can act as a start, but we need people going to these companies asking what they want to fill it out. Is anyone doing this/thinking of doing this? Al. On Thu, 2002-01-10 at 07:19, Ashvil wrote: The end-user technology is not compelling enough yet to make average users of the big IM services come over to form a critical mass of Jabber users, regardless of what their ISPs say. As a client developer that hurts but that's a true statement. How can we open a conversation between Jabber end users, server developers and client developers. One idea is an Open Survey at Jabber.org to find the compelling feature set that a Jabber client requires. Does this make sense. If so, I can develop a survey like this. Can we host this at Jabber.org? I suspect that we would have difficulty at this point supporting that many users; relative to the other services, Jabber tends to be more complex and has a considerable learning curve of its own. With regards to scalabilty, I feel that we need to move to the Apache and SendMail model. Millions of small servers in different domains, rather then AOL's centralized model. If some ISP still demands scalabilty, then they can choose Jabber.com's JCS. But we are finding new metaphors in our clients; our protocol and server software is evolving; and we as a community are slowly learning our strengths and weaknesses so that we can capitalize and anticipate. One of the biggest strengths that Jabber has is that it allows a sysadmin to setup their own IM server, that can communicate to the outside world. We need to leverage this to make Jabber grow via the grassroots way like Apache and Mozilla. Regards, Ashvil ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
RE: [JDEV] Creating a new Transport/jsm?
Hello, I agree with you, I also find it very hard to get any documentation on creating Transport/Agent/Module/Component/whatever-the-heck-it's-supposed-to-be.. . :-) -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] On Behalf Of Peter Gebauer Sent: quarta-feira, 9 de Janeiro de 2002 16:37 To: [EMAIL PROTECTED] Subject: [JDEV] Creating a new Transport/jsm? Hey there! I've been reading the documentation found at jabber.org, however, starting to write your own transport or module is obviously harder than it should be. There should be a manual on this that explains how to build your own transport and / or module. I don't even know the difference between mod_echo and, say, irc transport. Are they both transports or is mod_echo a module? And if so, can I write my own modules that are not included under jsm/modules? Too many names are used for the same thing? Agent, Transport, Module, Component? What is what? I have a lot of unanswered questions in my head and none of the documents seem help me. Even though there are many documents, they are either for people installing already existing stuff or focus on instant messaging only where as I would like to use Jabber for more than just sending one-liners to AIM, SMS, ICQ, etc. If someone has any good suggestions on SPECIFIC documents that will aid me in my quest to create a Transport/Agent/Module/Component/whatever-the-heck-it's-supposed-to-be, appreciate it! Also, if someone could straighten out the terminology, that would be good. /P ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
RE: [JDEV] Creating a new Transport/jsm?
Hi, I agree with Introfini too, I would also like to get some more info on this. Regards, Ritu Quoting introfini [EMAIL PROTECTED]: Hello, I agree with you, I also find it very hard to get any documentation on creating Transport/Agent/Module/Component/whatever-the-heck-it\'s-supposed-to-be.. . :-) -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] On Behalf Of Peter Gebauer Sent: quarta-feira, 9 de Janeiro de 2002 16:37 To: [EMAIL PROTECTED] Subject: [JDEV] Creating a new Transport/jsm? Hey there! I\'ve been reading the documentation found at jabber.org, however, starting to write your own transport or module is obviously harder than it should be. There should be a manual on this that explains how to build your own transport and / or module. I don\'t even know the difference between mod_echo and, say, irc transport. Are they both transports or is mod_echo a module? And if so, can I write my own modules that are not included under jsm/modules? Too many names are used for the same thing? Agent, Transport, Module, Component? What is what? I have a lot of unanswered questions in my head and none of the documents seem help me. Even though there are many documents, they are either for people installing already existing stuff or focus on instant messaging only where as I would like to use Jabber for more than just sending one-liners to AIM, SMS, ICQ, etc. If someone has any good suggestions on SPECIFIC documents that will aid me in my quest to create a Transport/Agent/Module/Component/whatever-the-heck-it\'s-supposed-to-be, appreciate it! Also, if someone could straighten out the terminology, that would be good. /P ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] Creating a new Transport/jsm?
On Thu, Jan 10, 2002 at 04:40:52PM +0530, Ritu Khetan wrote: I agree with Introfini too, I would also like to get some more info on this. What kind of component do you want? I've made some simple ones in Perl (Jabber::Connection) and Python (jabber.py). I don't understand the C code of the existing transports. ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] Creating a new Transport/jsm?
Hi, Can you send me the Jabber::Connection one. I would like to look at it.Do let me know what it is meant for. Thanks Regards, Ritu Quoting Migs Paraz [EMAIL PROTECTED]: On Thu, Jan 10, 2002 at 04:40:52PM +0530, Ritu Khetan wrote: I agree with Introfini too, I would also like to get some more info on this. What kind of component do you want? I\'ve made some simple ones in Perl (Jabber::Connection) and Python (jabber.py). I don\'t understand the C code of the existing transports. ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] The Important Things
Ashvil wrote: With regards to scalabilty, I feel that we need to move to the Apache and SendMail model. Millions of small servers in different domains, rather then AOL's centralized model. If some ISP still demands scalabilty, then they can choose Jabber.com's JCS. snip So i don't understand why somebody is saying something like this on a mailing list about opensource project. First of all, I don't work for Jabber.com or will gain if they do well. I do care that Jabber does well. For me, Jabber is a community that includes the Jabber open source server, other proprietary servers like JCS, Jabcast, etc., open source clients WinJab, etc., proprietary clients like Jabber IM, Vista, etc., clients libraries, server modules, innovative stuff like Jogger, etc. and finally the end users who use Jabber. From the technical side we are bound by the Jabber specification. So Jabber to me is MORE then just an Open Source project. I think this is the view in the Jabber community also, but if I am wrong, please correct me. With regard to my comment on JCS, my position is connecting to a Jabber server (whether open source or not) is better then connecting to an AOL server. Using a client supporting the Jabber protocol (whether open source or not) is better then using the AIM client. Both the actions GROW the jabber community, so if I want to use the open source versions of either the client or server, I can MAKE that choice. With the CLOSED AOL AIM protocol, that choice is not possible. it's ok we can stop developing the server. I never said in my email that we need to stop development of the server. I support your desire to for Jabber to be 'the best server'. All I said were two things with regards to the Jabber server, which I am expanding and repeating here 1. If the Jabber server is used in a distributed fashion like sendmail and Apache, then that reduces the scalability requirements. 2. For ISPs that demand more scalability, let them use JCS, etc. now instead of choosing some closed solution like Odigo or co-branding AOL, Yahoo, etc. That way, when the jabber server meets their requirements, they can adopt it easily. Regards, Ashvil ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] Creating a new Transport/jsm?
On Thu, Jan 10, 2002 at 05:20:39PM +0530, Ritu Khetan wrote: Can you send me the Jabber::Connection one. I would like to look at it.Do let me know what it is meant for. http://www.internet.org.ph/download/juddi-2002.tar.gz a basic User Directory (JUD), requires MySQL. enjoy! ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] The Important Things
The list from Mike Lin can act as a start, but we need people going to these companies asking what they want to fill it out. One big request is support for HTTP proxy. Jermie replied to my request earlier that this functionality is in CVS. Can some of the public jabber servers (jabberview.com) support this ? This would enable lot of users behind firewalls to join the jabber community. Regards, Ashvil ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
RE: [JDEV] The Important Things
Title: RE: [JDEV] The Important Things I'm new in the discussion but need to install a server in an ISP, about 3000 concurrent users, do I have to use JCS or open source jabber is OK for this? THANKS in advance! dedalo -Mensaje original- De: Ashvil [mailto:[EMAIL PROTECTED]] Enviado el: Jueves, 10 de Enero de 2002 09:12 a.m. Para: [EMAIL PROTECTED] Asunto: Re: [JDEV] The Important Things Ashvil wrote: With regards to scalabilty, I feel that we need to move to the Apache and SendMail model. Millions of small servers in different domains, rather then AOL's centralized model. If some ISP still demands scalabilty, then they can choose Jabber.com's JCS. snip So i don't understand why somebody is saying something like this on a mailing list about opensource project. First of all, I don't work for Jabber.com or will gain if they do well. I do care that Jabber does well. For me, Jabber is a community that includes the Jabber open source server, other proprietary servers like JCS, Jabcast, etc., open source clients WinJab, etc., proprietary clients like Jabber IM, Vista, etc., clients libraries, server modules, innovative stuff like Jogger, etc. and finally the end users who use Jabber. From the technical side we are bound by the Jabber specification. So Jabber to me is MORE then just an Open Source project. I think this is the view in the Jabber community also, but if I am wrong, please correct me. With regard to my comment on JCS, my position is connecting to a Jabber server (whether open source or not) is better then connecting to an AOL server. Using a client supporting the Jabber protocol (whether open source or not) is better then using the AIM client. Both the actions GROW the jabber community, so if I want to use the open source versions of either the client or server, I can MAKE that choice. With the CLOSED AOL AIM protocol, that choice is not possible. it's ok we can stop developing the server. I never said in my email that we need to stop development of the server. I support your desire to for Jabber to be 'the best server'. All I said were two things with regards to the Jabber server, which I am expanding and repeating here 1. If the Jabber server is used in a distributed fashion like sendmail and Apache, then that reduces the scalability requirements. 2. For ISPs that demand more scalability, let them use JCS, etc. now instead of choosing some closed solution like Odigo or co-branding AOL, Yahoo, etc. That way, when the jabber server meets their requirements, they can adopt it easily. Regards, Ashvil ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] The Important Things
On Thu, Jan 10, 2002 at 05:41:41PM +0530, Ashvil wrote: 1. If the Jabber server is used in a distributed fashion like sendmail and Apache, then that reduces the scalability requirements. If I understand correctly, you mean that all the servers that speak the Jabber protocol will be able to deliver to other servers. Right now, jabberd's aren't configured to take connections from just about anyone, so we'll have to make changes in the paradigm or the protocol for that. ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
RE: [JDEV] The Important Things
Title: RE: [JDEV] The Important Things Good question especially if you'd like to authenticate using LDAP? - RM -Original Message-From: Rodrigo Roman [mailto:[EMAIL PROTECTED]]Sent: Thursday, January 10, 2002 2:22 PMTo: '[EMAIL PROTECTED]'Subject: RE: [JDEV] The Important Things I'm new in the discussion but need to install a server in an ISP, about 3000 concurrent users, do I have to use JCS or open source jabber is OK for this? THANKS in advance! dedalo -Mensaje original- De: Ashvil [mailto:[EMAIL PROTECTED]] Enviado el: Jueves, 10 de Enero de 2002 09:12 a.m. Para: [EMAIL PROTECTED] Asunto: Re: [JDEV] The Important Things Ashvil wrote: With regards to scalabilty, I feel that we need to move to the Apache and SendMail model. Millions of small servers in different domains, rather then AOL's centralized model. If some ISP still demands scalabilty, then they can choose Jabber.com's JCS. snip So i don't understand why somebody is saying something like this on a mailing list about opensource project. First of all, I don't work for Jabber.com or will gain if they do well. I do care that Jabber does well. For me, Jabber is a community that includes the Jabber open source server, other proprietary servers like JCS, Jabcast, etc., open source clients WinJab, etc., proprietary clients like Jabber IM, Vista, etc., clients libraries, server modules, innovative stuff like Jogger, etc. and finally the end users who use Jabber. From the technical side we are bound by the Jabber specification. So Jabber to me is MORE then just an Open Source project. I think this is the view in the Jabber community also, but if I am wrong, please correct me. With regard to my comment on JCS, my position is connecting to a Jabber server (whether open source or not) is better then connecting to an AOL server. Using a client supporting the Jabber protocol (whether open source or not) is better then using the AIM client. Both the actions GROW the jabber community, so if I want to use the open source versions of either the client or server, I can MAKE that choice. With the CLOSED AOL AIM protocol, that choice is not possible. it's ok we can stop developing the server. I never said in my email that we need to stop development of the server. I support your desire to for Jabber to be 'the best server'. All I said were two things with regards to the Jabber server, which I am expanding and repeating here 1. If the Jabber server is used in a distributed fashion like sendmail and Apache, then that reduces the scalability requirements. 2. For ISPs that demand more scalability, let them use JCS, etc. now instead of choosing some closed solution like Odigo or co-branding AOL, Yahoo, etc. That way, when the jabber server meets their requirements, they can adopt it easily. Regards, Ashvil ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
RE: [JDEV] The Important Things
Title: Mensaje I'm no decided about authentication platform, I can use whatever I need, but it must be stable of course.. :) -Mensaje original-De: Riyaad Miller [mailto:[EMAIL PROTECTED]] Enviado el: Jueves, 10 de Enero de 2002 09:43 a.m.Para: [EMAIL PROTECTED]Asunto: RE: [JDEV] The Important Things Good question especially if you'd like to authenticate using LDAP? - RM -Original Message-From: Rodrigo Roman [mailto:[EMAIL PROTECTED]]Sent: Thursday, January 10, 2002 2:22 PMTo: '[EMAIL PROTECTED]'Subject: RE: [JDEV] The Important Things I'm new in the discussion but need to install a server in an ISP, about 3000 concurrent users, do I have to use JCS or open source jabber is OK for this? THANKS in advance! dedalo -Mensaje original- De: Ashvil [mailto:[EMAIL PROTECTED]] Enviado el: Jueves, 10 de Enero de 2002 09:12 a.m. Para: [EMAIL PROTECTED] Asunto: Re: [JDEV] The Important Things Ashvil wrote: With regards to scalabilty, I feel that we need to move to the Apache and SendMail model. Millions of small servers in different domains, rather then AOL's centralized model. If some ISP still demands scalabilty, then they can choose Jabber.com's JCS. snip So i don't understand why somebody is saying something like this on a mailing list about opensource project. First of all, I don't work for Jabber.com or will gain if they do well. I do care that Jabber does well. For me, Jabber is a community that includes the Jabber open source server, other proprietary servers like JCS, Jabcast, etc., open source clients WinJab, etc., proprietary clients like Jabber IM, Vista, etc., clients libraries, server modules, innovative stuff like Jogger, etc. and finally the end users who use Jabber. From the technical side we are bound by the Jabber specification. So Jabber to me is MORE then just an Open Source project. I think this is the view in the Jabber community also, but if I am wrong, please correct me. With regard to my comment on JCS, my position is connecting to a Jabber server (whether open source or not) is better then connecting to an AOL server. Using a client supporting the Jabber protocol (whether open source or not) is better then using the AIM client. Both the actions GROW the jabber community, so if I want to use the open source versions of either the client or server, I can MAKE that choice. With the CLOSED AOL AIM protocol, that choice is not possible. it's ok we can stop developing the server. I never said in my email that we need to stop development of the server. I support your desire to for Jabber to be 'the best server'. All I said were two things with regards to the Jabber server, which I am expanding and repeating here 1. If the Jabber server is used in a distributed fashion like sendmail and Apache, then that reduces the scalability requirements. 2. For ISPs that demand more scalability, let them use JCS, etc. now instead of choosing some closed solution like Odigo or co-branding AOL, Yahoo, etc. That way, when the jabber server meets their requirements, they can adopt it easily. Regards, Ashvil ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
RE: [JDEV] Thoughts on AOL
Hi all Is there like any particular reason http://www.jabber.com should be dead? Does this site still exist or is it merely a temp prob? -Original Message- From: dlb [mailto:[EMAIL PROTECTED]] Sent: Wednesday, January 09, 2002 11:10 PM To: [EMAIL PROTECTED] Subject: Re: [JDEV] Thoughts on AOL One thing to consider regarding AOL is that they'll have to adopt some measure of interoperability, either standards based or w/ specific competing services , prior to launching IM enhanced broadband applications within cable regions covered by the AOL - Time Warner consent decree. I think that this argues for a focus on the continued development of Jabber in its role as a user /resource aware XML messaging routing architecture. The fundamental extensibility of Jabber along with the work being done w/ XML-RPC and SOAP , and the design concepts being implemented w/ Jabber2 have the potential to establish Jabber as the platform of choice for precisely these sorts of applications. I don't see any way of overcoming the network effects supporting the existing AOL service - you're never going to get a significant number of AOL IM users to adopt Jabber and AOL won't allow interop until they're forced to. One strategy that I think could work would be to position Jabber as a competitor in the area of IM enhanced broadband apps and to deploy applications and services within systems covered by the consent decree. AOL isn't going to want a rival technology to gain a significant foothold in this area and would eventually forward an interop proposal simply to enable them to introduce rival applications. preliminary work in this area could include : Developing OCAP (OpenCable Applications Platform) compatible jabber components. Developing ATVEF based clients for Liberate / AOLTV - ATVEF is based on web dev. standards . Promoting Jabber as a 'platform' for next generation BB and ITV apps. if anyone else is interested in this area pls contact me -David ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] The Important Things
Rodrigo Roman wrote: I'm new in the discussion but need to install a server in an ISP, about 3000 concurrent users, do I have to use JCS or open source jabber is OK for this? i think OSS version is not yet ready. The first problem is that standard version of Jabber server has some strange memory managment problems so it's growing quite fast. And with 3000concurrent users you will have to restart this server every few days/hours. Another thing you will have to use jpolld to let more than 1k users connect to your site. We were making tests with few thousands users connected to the jabber server trough jpolld with sucess ( if i remember well it was something about 20k users on PC with 1GB of ram ). But this was only scripts connected not really chatting, changing presence etc. I will try to send some test results that we made. I don't have experience with commercial server, don't even know anybody using it so i cannot say what it really can do. ---kuba troszok [EMAIL PROTECTED] ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] The Important Things
On Thu, Jan 10, 2002 at 05:41:41PM +0530, Ashvil wrote: 1. If the Jabber server is used in a distributed fashion like sendmail and Apache, then that reduces the scalability requirements. If I understand correctly, you mean that all the servers that speak the Jabber protocol will be able to deliver to other servers. Right now, jabberd's aren't configured to take connections from just about anyone, so we'll have to make changes in the paradigm or the protocol for that. My understanding was this was already done. A user at jabber.org and another user at jabber.cz can see each other's presence and send messages to each other. Regards, Ashvil ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
RE: [JDEV] Scale of server
Title: Scale of server Jabber can handle 3000 concurrent users with no problem. The only problem you will have is that the select() function in Linux is hard coded to a certain number of sockets. In kernel 2.4 that limit is 1024. You can either recompile the kernel to allow more, or you can use dpsm or jpolld as your socket manager. These socket managers use the poll() function instead of select() and are not subject to the limit. You will however have to increase the number of file descriptors you can have open per process. This can be done in a startup script that calls ulimit. Jason Reineri Northridge Systems, Inc. -Original Message- From: Rodrigo Roman [mailto:[EMAIL PROTECTED]] Sent: Wednesday, January 09, 2002 6:14 PM To: '[EMAIL PROTECTED]' Subject: [JDEV] Scale of server I'm planing an instalation of jabber server in a isp, we have aprox 3000 concurrent users (i dont know wich porcentage will use our messanger) our client base is 150k users. Wich is the best implementation for this numbers? Is jabber prepared for this? Are all the transports able to work in this client base? Thanks for your help!
Re: [JDEV] The Important Things
Ashvil wrote: First of all, I don't work for Jabber.com or will gain if they do well. :) i didn't want to anybody think like this I do care that Jabber does well. For me, Jabber is a community that includes the Jabber open source server, other proprietary servers like JCS, Jabcast, etc., open source clients WinJab, etc., proprietary clients like Jabber IM, Vista, etc., clients libraries, server modules, innovative stuff like Jogger, etc. and finally the end users who use Jabber. From the technical side we are bound by the Jabber specification. So Jabber to me is MORE then just an Open Source project. I think this is the view in the Jabber community also, but if I am wrong, please correct me. yes, you're right. This is jabber community. But to let it grow we will have to let all the components grow or we will finish only with free clients and some products that we can buy. With regard to my comment on JCS, my position is connecting to a Jabber server (whether open source or not) is better then connecting to an AOL server. Using a client supporting the Jabber protocol (whether open source or not) is better then using the AIM client. Both the actions GROW the jabber community, so if I want to use the open source versions of either the client or server, I can MAKE that choice. With the CLOSED AOL AIM protocol, that choice is not possible. Of course but in a world where only people running small lans with few dozens of users maximum use the open source server and all the others have to buy commercial server. no, but we can come into this conclusion in few months. As commercial server has almost everything - and oss has just the basic functionalities. I've seen already some responses here saying that - for this or for that you need to go and buy commercial server. And no discussion to change it in a way that people can download free version with these features. I never said in my email that we need to stop development of the server. I support your desire to for Jabber to be 'the best server'. All I said were two things with regards to the Jabber server, which I am expanding and repeating here no, but we can come into this conclusion in few months. As commercial server has almost everything - and oss has just the basic functionalities. I've seen already some responses here saying that - for this or for that you need to go and buy commercial server. And no discussion to change it in a way that people can download free version with this features. 1. If the Jabber server is used in a distributed fashion like sendmail and Apache, then that reduces the scalability requirement yes, and i was thinking about this already. But the problem is that people want's to have addresses like [EMAIL PROTECTED] not [EMAIL PROTECTED] . Because this will be needed in the future if your system will grow. I found that we could use some kind of a gateway - people connect to one server ( for example jabbber.org ) autheticate - get a token/session id - and then continue with a server l1.l4.dddljfds.jabbber.org that are real jabber servers. The external traffic goes always trough the main gateway jabbber.org and the internal traffic is going between machines that knows that l1.l4.dddljfds.jabber.org and l2.l4.dddljfds.jabber.org are in this same domain. The advantage is that noone have to know about our internat structure, for the external users we will always look like [EMAIL PROTECTED] 2. For ISPs that demand more scalability, let them use JCS, etc. now instead of choosing some closed solution like Odigo or co-branding AOL, Yahoo, etc. That way, when the jabber server meets their requirements, they can adopt it easily. How somebody can adopt closed source software ? Reverse engineer ? Yes you are right but in my opionion if you choose closed source software you go for closed solution. As you don't have possibility to change tcp/ip stack on M$ Windows you will not have possibility to modify anything inside JCS. ---kuba troszok [EMAIL PROTECTED] ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] The Important Things
Ashvil wrote: My understanding was this was already done. A user at jabber.org and another user at jabber.cz can see each other's presence and send messages to each other. the problem starts when you have more and more users - and i would like to repeat that in my opinion buying the commercial server is not solution - many people loose much time to convince their customers to use OpenSource technologies and this will not be a solution if at the end of the road will stand closed source server that they will have to buy, pay a lot and forget about OSS model. Customers wants long term solution, even if they don't need it. ---kuba troszok [EMAIL PROTECTED] ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] The Important Things
I found that we could use some kind of a gateway - people connect to one server ( for example jabbber.org ) autheticate - get a token/session id - and then continue with a server l1.l4.dddljfds.jabbber.org that are real jabber servers. snip Any ideas that can help in scalability are welcome. If we can use a pool of cheap PCs to build a scalable jabber network, then even more valuable then having One big Server with Gigs of memory. This will require some changes in the Jabber protocol. The MSN protocol does something like this, but takes this one step ahead by letting you connect to any server in the pool, which then refers you to the right server that can authenticate you. If you make logging in a two-step process, you can solve this problem but that would mean changing all the Jabber clients and also the S2S communication in the Jabber server. Anyway, this is an area that the Jabber server developers are the best folks to comment on. Regards, Ashvil ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] DTD or Schema for a (jabber:x:*) namespace?
All DTDs (and XML Schemas) in Jabber are purely informational - because the traffic stream of jabber is basically one long XML document, any schema violations would cause that document to become invalid would require that the parser report an error and stop. The general rule is that any deviations from the DTDs provided are to be ignored. -David Waite On Thu, 2002-01-10 at 05:10, Bernhard Pietsch wrote: All extension namespaces i found use a DTD. Is that necessary or is it possible to use XML Schema instead? thx, Bernhard Pietsch ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] Creating a new Transport/jsm?
I wrote a basic doc on starting in Component development and understanding the differences in the protocol. It's only at revision 1 and could use more work, but I haven't had time yet. http://docs.jabber.org/general/html/component-intro.html Hopefully I can work on it more soon. --temas On Thu, Jan 10, 2002 at 04:40:52PM +0530, Ritu Khetan wrote: Hi, I agree with Introfini too, I would also like to get some more info on this. Regards, Ritu Quoting introfini [EMAIL PROTECTED]: Hello, I agree with you, I also find it very hard to get any documentation on creating Transport/Agent/Module/Component/whatever-the-heck-it\'s-supposed-to-be.. . :-) -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] On Behalf Of Peter Gebauer Sent: quarta-feira, 9 de Janeiro de 2002 16:37 To: [EMAIL PROTECTED] Subject: [JDEV] Creating a new Transport/jsm? Hey there! I\'ve been reading the documentation found at jabber.org, however, starting to write your own transport or module is obviously harder than it should be. There should be a manual on this that explains how to build your own transport and / or module. I don\'t even know the difference between mod_echo and, say, irc transport. Are they both transports or is mod_echo a module? And if so, can I write my own modules that are not included under jsm/modules? Too many names are used for the same thing? Agent, Transport, Module, Component? What is what? I have a lot of unanswered questions in my head and none of the documents seem help me. Even though there are many documents, they are either for people installing already existing stuff or focus on instant messaging only where as I would like to use Jabber for more than just sending one-liners to AIM, SMS, ICQ, etc. If someone has any good suggestions on SPECIFIC documents that will aid me in my quest to create a Transport/Agent/Module/Component/whatever-the-heck-it\'s-supposed-to-be, appreciate it! Also, if someone could straighten out the terminology, that would be good. /P ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev -- Thomas Muldowney email/jabber: [EMAIL PROTECTED] msg03902/pgp0.pgp Description: PGP signature
Re: [JDEV] DTD or Schema for a (jabber:x:*) namespace?
Schemas are very possible and being worked on now. Here's an example one for jabber:iq:last schema xmlns='http://www.w3.org/2001/XMLSchema' xmlns:last='jabber:iq:last' targetNamespace='jabber:iq:last' elementFormDefault='qualified' element name='last:query' complexType mixed='true' attribute name='seconds' type='unsignedLong' use='optional'/ /complexType /element /schema I'm not a schema expert (yet ;-]) so this could be wrong, but I think it's correct and it validates =) --temas On Thu, Jan 10, 2002 at 01:10:48PM +0100, Bernhard Pietsch wrote: All extension namespaces i found use a DTD. Is that necessary or is it possible to use XML Schema instead? thx, Bernhard Pietsch ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev -- Thomas Muldowney email/jabber: [EMAIL PROTECTED] msg03903/pgp0.pgp Description: PGP signature
Re: [JDEV] The Important Things
This is the piece (http/wcs + mod_jabber) that we've talked about moving over to the Apache project -- see the logs from yesterday's meeting: http://perl.jabber.org/logs/conference.jabber.org/foundation/2002-01-09.html IMHO moving this code under the Apache umbrella would speed up development immeasurably and help us quickly develop web interfaces to Jabber, thus getting around those pecky firewalls. :) Peter -- Peter Saint-Andre email+jabber: [EMAIL PROTECTED] web: http://www.saint-andre.com/ On Thu, 10 Jan 2002, Ashvil wrote: The list from Mike Lin can act as a start, but we need people going to these companies asking what they want to fill it out. One big request is support for HTTP proxy. Jermie replied to my request earlier that this functionality is in CVS. Can some of the public jabber servers (jabberview.com) support this ? This would enable lot of users behind firewalls to join the jabber community. Regards, Ashvil ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
RE: [JDEV] Thoughts on AOL
Temporary problem with some network config or somesuch this morning. Peter -- Peter Saint-Andre email+jabber: [EMAIL PROTECTED] web: http://www.saint-andre.com/ On Thu, 10 Jan 2002, Riyaad Miller wrote: Hi all Is there like any particular reason http://www.jabber.com should be dead? Does this site still exist or is it merely a temp prob? -Original Message- From: dlb [mailto:[EMAIL PROTECTED]] Sent: Wednesday, January 09, 2002 11:10 PM To: [EMAIL PROTECTED] Subject: Re: [JDEV] Thoughts on AOL One thing to consider regarding AOL is that they'll have to adopt some measure of interoperability, either standards based or w/ specific competing services , prior to launching IM enhanced broadband applications within cable regions covered by the AOL - Time Warner consent decree. I think that this argues for a focus on the continued development of Jabber in its role as a user /resource aware XML messaging routing architecture. The fundamental extensibility of Jabber along with the work being done w/ XML-RPC and SOAP , and the design concepts being implemented w/ Jabber2 have the potential to establish Jabber as the platform of choice for precisely these sorts of applications. I don't see any way of overcoming the network effects supporting the existing AOL service - you're never going to get a significant number of AOL IM users to adopt Jabber and AOL won't allow interop until they're forced to. One strategy that I think could work would be to position Jabber as a competitor in the area of IM enhanced broadband apps and to deploy applications and services within systems covered by the consent decree. AOL isn't going to want a rival technology to gain a significant foothold in this area and would eventually forward an interop proposal simply to enable them to introduce rival applications. preliminary work in this area could include : Developing OCAP (OpenCable Applications Platform) compatible jabber components. Developing ATVEF based clients for Liberate / AOLTV - ATVEF is based on web dev. standards . Promoting Jabber as a 'platform' for next generation BB and ITV apps. if anyone else is interested in this area pls contact me -David ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] The Important Things
On Thu, 10 Jan 2002, Migs Paraz wrote: On Thu, Jan 10, 2002 at 05:41:41PM +0530, Ashvil wrote: 1. If the Jabber server is used in a distributed fashion like sendmail and Apache, then that reduces the scalability requirements. If I understand correctly, you mean that all the servers that speak the Jabber protocol will be able to deliver to other servers. Right now, jabberd's aren't configured to take connections from just about anyone, so we'll have to make changes in the paradigm or the protocol for that. ? One of the many advantages of Jabber is that it enables us to build out a network of servers, each connecting to the other. Of couse, you *can* turn off server-to-server communications, but the default is to communicate with other servers. Peter ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] The Important Things
IMHO moving this code under the Apache umbrella would speed up development immeasurably and help us quickly develop web interfaces to Jabber, thus getting around those pecky firewalls. :) s/pecky/pesky/ :) ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] The Important Things
There is also the simple hack in the server that sends a http header when you connect to a port flagged as http compatible. I forget the flag, but maybe jer will see this message and pipe up about it. Otherwise I'll dig a bit for it. Take a look at Jarl or Winjab they both support the mode. --temas On Thu, Jan 10, 2002 at 10:10:31AM -0600, Peter Saint-Andre wrote: This is the piece (http/wcs + mod_jabber) that we've talked about moving over to the Apache project -- see the logs from yesterday's meeting: http://perl.jabber.org/logs/conference.jabber.org/foundation/2002-01-09.html IMHO moving this code under the Apache umbrella would speed up development immeasurably and help us quickly develop web interfaces to Jabber, thus getting around those pecky firewalls. :) Peter -- Peter Saint-Andre email+jabber: [EMAIL PROTECTED] web: http://www.saint-andre.com/ On Thu, 10 Jan 2002, Ashvil wrote: The list from Mike Lin can act as a start, but we need people going to these companies asking what they want to fill it out. One big request is support for HTTP proxy. Jermie replied to my request earlier that this functionality is in CVS. Can some of the public jabber servers (jabberview.com) support this ? This would enable lot of users behind firewalls to join the jabber community. Regards, Ashvil ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev -- Thomas Muldowney email/jabber: [EMAIL PROTECTED] msg03908/pgp0.pgp Description: PGP signature
Re: [JDEV] The Important Things
I totaly agree with Ashvil. It's the same kind of discussion I was trying discuss in the jadmin list. You can follow the the below discussion in the jadmin list. - Original Message - From: Alper Tarhan [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Thursday, January 10, 2002 12:07 PM Subject: Re: [jadmin] When may the scalability issues solved?? Do you think these performance parameters are really necessary for the design? I mean in my opinion, anyone suffering from scalability and performance should easily be able to overcome this problem by adding a new hardware box to the farm. I know it is easy to talk about it :) and it may require radical changes in the design but it doesnt seem impossible. This issue is the most important advantage of the commercial server over the open-source one. It's for sure Jabber.com has to earn some money to stay alive and continue support for Jabber technology. But in my opinion without solving the scalability issues, the open-source server will go no furher than serving limited community of users and never being considered as a mass IM technology or an alternative to the commercial one. - Original Message - From: Peter Saint-Andre [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Thursday, January 10, 2002 6:52 AM Subject: Re: [jadmin] When may the scalability issues solved?? Well I think we'd need to set some parameters. It's easy to talk about scalability but how high do people (e.g., you) really need to scale in terms of registered users, concurrent users, messages per second, and so on? Hard numbers will give developers some design goals. Peter -- Peter Saint-Andre email+jabber: [EMAIL PROTECTED] web: http://www.saint-andre.com/ On Wed, 9 Jan 2002, Alper Tarhan wrote: Hi there, I have been following the Jabber technologies through this list for about a year. And I have also setup a server myself and wrote a simple client for test. At that time, I was searching for messaging technologies and heard about Jabber. After some research, my first question was the Scalability Issues which seemed a major problem to me. I wrote it to the list and I asked what to do or any plans for solving this. I was told that by the time I was able to collect that many users, the problem would be already solved. I havent of course that many users but without the techn. availability, you cannot promote your service with confidence. As you can see the frequent posts on this list, Scalability Issues are still the major concern for many of us. I appreciate and congratulate all the designers, developers and contributers to this technology. But in order to boost it, in my opinion, it is a Priority #1 issue to be solved. I know there are methods to increase the # concurrent users on a single machine. There are also contributions to allow for farming and multiplexing socket connections. Do it yourself may be a method, but I think this issue must be solved in design and standardly integrated in the main architecture . Again in my opinion, the best way would be the farming with support for a central xdb, which would give the most flexible architecture in terms of scalability and high-availability. You can offer the Jabber.com commercial product for those big players, but you know it wouldnt be the best solution for many of us. I would be pleased to know what you think about this issue. What priority should it be for future releases? Best regards, ___ jadmin mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jadmin ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] DTD or Schema for a (jabber:x:*) namespace?
Schema would be better IMHO and we're slowly moving in that direction with the standards cleanup that's in progress. Peter -- Peter Saint-Andre email+jabber: [EMAIL PROTECTED] web: http://www.saint-andre.com/ On Thu, 10 Jan 2002, Bernhard Pietsch wrote: All extension namespaces i found use a DTD. Is that necessary or is it possible to use XML Schema instead? thx, Bernhard Pietsch ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] Scale of server
Title: Scale of server Do you know of a Jabber server that has had 3K users on at one time? - Original Message - From: Jason Reineri To: [EMAIL PROTECTED] Sent: Thursday, January 10, 2002 7:29 AM Subject: RE: [JDEV] Scale of server Jabber can handle 3000 concurrent users with no problem. The only problem you will have is that the select() function in Linux is hard coded to a certain number of sockets. In kernel 2.4 that limit is 1024. You can either recompile the kernel to allow more, or you can use dpsm or jpolld as your socket manager. These socket managers use the poll() function instead of select() and are not subject to the limit. You will however have to increase the number of file descriptors you can have open per process. This can be done in a startup script that calls ulimit. Jason Reineri Northridge Systems, Inc. -Original Message-From: Rodrigo Roman [mailto:[EMAIL PROTECTED]] Sent: Wednesday, January 09, 2002 6:14 PMTo: '[EMAIL PROTECTED]'Subject: [JDEV] Scale of server I'm planing an instalation of jabber server in a isp, we have aprox 3000 concurrent users (i dont know wich porcentage will use our messanger) our client base is 150k users. Wich is the best implementation for this numbers? Is jabber prepared for this? Are all the transports able to work in this client base? Thanks for your help!
Re: [JDEV] The Important Things
Why not add authentication and message relaying to the S2S protocol. This would give four advantages; 1. Any user could log into any machine and the server would relay the authentication request to the relevant machine to handle authentication. 2. The messages for that user would be relayed to the server they are logged in to and then forwarded on to them. 3. Clusters or farms could be constructed to server a a single jabber community and the load shared between them. 4. This would only involve a change to the S2S protol and servers supporting it (of which there are few), and would leave the C2S protocol unchanged and thus not require any client changes. Comments? Al. On Thu, 2002-01-10 at 15:50, Ashvil wrote: I found that we could use some kind of a gateway - people connect to one server ( for example jabbber.org ) autheticate - get a token/session id - and then continue with a server l1.l4.dddljfds.jabbber.org that are real jabber servers. snip Any ideas that can help in scalability are welcome. If we can use a pool of cheap PCs to build a scalable jabber network, then even more valuable then having One big Server with Gigs of memory. This will require some changes in the Jabber protocol. The MSN protocol does something like this, but takes this one step ahead by letting you connect to any server in the pool, which then refers you to the right server that can authenticate you. If you make logging in a two-step process, you can solve this problem but that would mean changing all the Jabber clients and also the S2S communication in the Jabber server. Anyway, this is an area that the Jabber server developers are the best folks to comment on. Regards, Ashvil ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
RE: [JDEV] Scale of server
Title: Scale of server Im actually testing load on a server right now. I have one box with three instances of dpsm, one instance of jabberd, as well as the yahoo and msn transports running. The user information is stored in a mySQL db running on the same machine and jabber accesses it through a version of xdb_sql that we have modified to fix a few quirks and to make it work with the msn transport. The way we are load balancing the connections to the three instances of dpsm is to just run each one on a virtual ip and set up our dns to round robin the three ips. With a straight test that does nothing but connect sockets to the server, we can get 10,000 connections per instance of dpsm. Three instances of dpsm seems to be the limit with this test, as there are diminishing returns with any more than that. Im now using a tester app I found in the CVS repository to get an actual load test. The app has some problems with it so we are tweaking it and running tests. I dont have solid numbers yet, but I know that I have had 3,000 users simultaneously logged in and all of them sending messages. The problem with this test is that if you have 3,000 people logged in at any given time, they will not be constantly talking to each other. So with some tweaking we are trying to make a load test that will more accurately simulate a load. What we are trying to accomplish here is to have 30,000 clients connected to the one box, with a percentage of them talking that would accurately reflect a real life server. Note: Our server machine is just a desktop that we stole to test on. Its a one CPU P4 1.3Ghz I believe, with 1GB RAM. If we had multiple CPUs I suspect we could get away with running even more instances of dpsm. We are also going to try adding 2 more network cards to the machine so each instance of dpsm will be on its own network card. Let me know if you need any more info. Ive found the information on scaling Jabber to be pretty lean and hard to find. Jason Reineri Northridge Systems, Inc. -Original Message- From: Michael F. March [mailto:[EMAIL PROTECTED]] Sent: Thursday, January 10, 2002 2:38 PM To: [EMAIL PROTECTED] Subject: Re: [JDEV] Scale of server Do you know of a Jabber server that has had 3K users on at one time? - Original Message - From: Jason Reineri To: [EMAIL PROTECTED] Sent: Thursday, January 10, 2002 7:29 AM Subject: RE: [JDEV] Scale of server Jabber can handle 3000 concurrent users with no problem. The only problem you will have is that the select() function in Linux is hard coded to a certain number of sockets. In kernel 2.4 that limit is 1024. You can either recompile the kernel to allow more, or you can use dpsm or jpolld as your socket manager. These socket managers use the poll() function instead of select() and are not subject to the limit. You will however have to increase the number of file descriptors you can have open per process. This can be done in a startup script that calls ulimit. Jason Reineri Northridge Systems, Inc. -Original Message- From: Rodrigo Roman [mailto:[EMAIL PROTECTED]] Sent: Wednesday, January 09, 2002 6:14 PM To: '[EMAIL PROTECTED]' Subject: [JDEV] Scale of server I'm planing an instalation of jabber server in a isp, we have aprox 3000 concurrent users (i dont know wich porcentage will use our messanger) our client base is 150k users. Wich is the best implementation for this numbers? Is jabber prepared for this? Are all the transports able to work in this client base? Thanks for your help!
HTTP/Proxy access to 5222 (was Re: [JDEV] The Important Things)
There is also the simple hack in the server that sends a http header when you connect to a port flagged as http compatible. I forget the flag, but maybe jer will see this message and pipe up about it. Otherwise I'll dig a bit for it. Take a look at Jarl or Winjab they both support the mode. The details are at http://jabber.org/?oid=971 and it actually works with the PUT or POST methods, and should work over ssl if you have your server built to support ssl as well. Although it's built into the default socket manager that comes w/ 1.4, it's not yet been included into any of the external socket manager projects that I know of, although it shouldn't be hard for them to add support if people find it useful. Jer ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] The Important Things
First, for normal s2s, unless you were using some form of digest or zero-knowledge authentication you wouldn't want to trust remote servers to do this kind of thing for your users :) The more important answer is that the server pretty much does and allows this already. The protocol that the client socket managers use to talk to the session manager works like you describe, you can have many client handlers working as a farm running from anywhere. In a future generation I could easily see some capabilities to allow dynamic switching and redirecting of connections as mentioned. Jer On 10 Jan 2002, Al Sutton wrote: Why not add authentication and message relaying to the S2S protocol. This would give four advantages; 1. Any user could log into any machine and the server would relay the authentication request to the relevant machine to handle authentication. 2. The messages for that user would be relayed to the server they are logged in to and then forwarded on to them. 3. Clusters or farms could be constructed to server a a single jabber community and the load shared between them. 4. This would only involve a change to the S2S protol and servers supporting it (of which there are few), and would leave the C2S protocol unchanged and thus not require any client changes. Comments? Al. On Thu, 2002-01-10 at 15:50, Ashvil wrote: I found that we could use some kind of a gateway - people connect to one server ( for example jabbber.org ) autheticate - get a token/session id - and then continue with a server l1.l4.dddljfds.jabbber.org that are real jabber servers. snip Any ideas that can help in scalability are welcome. If we can use a pool of cheap PCs to build a scalable jabber network, then even more valuable then having One big Server with Gigs of memory. This will require some changes in the Jabber protocol. The MSN protocol does something like this, but takes this one step ahead by letting you connect to any server in the pool, which then refers you to the right server that can authenticate you. If you make logging in a two-step process, you can solve this problem but that would mean changing all the Jabber clients and also the S2S communication in the Jabber server. Anyway, this is an area that the Jabber server developers are the best folks to comment on. Regards, Ashvil ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] The Important Things
Distributed authentication is something that we need down the line, but it's not quite true that the changes it requires are trivial. Firstly, it requires trust between servers (the authenticating server simply asserts that the credentials presented by the client's server are valid.) Furthermore it has routing issues. If I want to log in as [EMAIL PROTECTED] on jabber.com, how does everyone know to route messages to jabber.com? This seems to imply some kind of smart multi-hop routing, which we don't have. All this can probably be done within one institution (one server farm) because you have that trust between the servers and there is probably already some kind of routing-resolving mechanism set up. But to have this on the wider Jabber cloud is going to require more careful engineering and strong crypto. -Mike |-+ | | Al Sutton| | | [EMAIL PROTECTED]| | | Sent by: | | | jdev-admin@jabber| | | .org | | || | || | | 01/10/2002 02:59 | | | PM | | | Please respond to| | | jdev | | || |-+ --| | | | To: [EMAIL PROTECTED] | | cc: | | Subject: Re: [JDEV] The Important Things | | | | | --| Why not add authentication and message relaying to the S2S protocol. This would give four advantages; 1. Any user could log into any machine and the server would relay the authentication request to the relevant machine to handle authentication. 2. The messages for that user would be relayed to the server they are logged in to and then forwarded on to them. 3. Clusters or farms could be constructed to server a a single jabber community and the load shared between them. 4. This would only involve a change to the S2S protol and servers supporting it (of which there are few), and would leave the C2S protocol unchanged and thus not require any client changes. Comments? Al. On Thu, 2002-01-10 at 15:50, Ashvil wrote: I found that we could use some kind of a gateway - people connect to one server ( for example jabbber.org ) autheticate - get a token/session id - and then continue with a server l1.l4.dddljfds.jabbber.org that are real jabber servers. snip Any ideas that can help in scalability are welcome. If we can use a pool of cheap PCs to build a scalable jabber network, then even more valuable then having One big Server with Gigs of memory. This will require some changes in the Jabber protocol. The MSN protocol does something like this, but takes this one step ahead by letting you connect to any server in the pool, which then refers you to the right server that can authenticate you. If you make logging in a two-step process, you can solve this problem but that would mean changing all the Jabber clients and also the S2S communication in the Jabber server. Anyway, this is an area that the Jabber server developers are the best folks to comment on. Regards, Ashvil ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
[JDEV] JECL questions
Hi All, Thanks for your help on jdev chat this afternoon! I am writing an external component that needs to send messages to group chats. It seems that I am always getting (on the server) Packet Delivery failed, invalid packet Dropping ... Is it possible to have a component send messages to group chats? What could I be doing wrong? Thanks Glenn
corrections... Re: [JDEV] The Important Things
no, but we can come into this conclusion in few months. As commercial server has almost everything - and oss has just the basic functionalities. Completly untrue, in fact, in many way it is exactly the opposite of that, the open source 1.4 server has newer protocol support, more options/features, and wider platform support as well. I've seen already some responses here saying that - for this or for that you need to go and buy commercial server. Jabber, Inc.'s commercial server has better scalability, higher rate of throughput, oracle/ldap integration, and is really targetted at large scale enterprises and service providers. If you want features, flexibility, the latest protocol development and prototyping support, and something that can be molded fairly easily into many environments you'll find it in open source. It's also getting better on the scalibility and will likely improve as more people need to push it further and can help with the development in that area. yes, and i was thinking about this already. But the problem is that people want's to have addresses like [EMAIL PROTECTED] not [EMAIL PROTECTED] . Because this will be needed in the future if your system will grow. Hmm, I think your confused, Jabber works very much like email and your address is user@domain, not user@localmachiene. Jer ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
RE: [JDEV] The Important Things
I really, really like this idea, although I would not consider myself an expert yet I am willing to help out/move it along in this arena. Tim Ferguson -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Al Sutton Sent: Thursday, January 10, 2002 1:00 PM To: [EMAIL PROTECTED] Subject: Re: [JDEV] The Important Things Why not add authentication and message relaying to the S2S protocol. This would give four advantages; 1. Any user could log into any machine and the server would relay the authentication request to the relevant machine to handle authentication. 2. The messages for that user would be relayed to the server they are logged in to and then forwarded on to them. 3. Clusters or farms could be constructed to server a a single jabber community and the load shared between them. 4. This would only involve a change to the S2S protol and servers supporting it (of which there are few), and would leave the C2S protocol unchanged and thus not require any client changes. Comments? Al. On Thu, 2002-01-10 at 15:50, Ashvil wrote: I found that we could use some kind of a gateway - people connect to one server ( for example jabbber.org ) autheticate - get a token/session id - and then continue with a server l1.l4.dddljfds.jabbber.org that are real jabber servers. snip Any ideas that can help in scalability are welcome. If we can use a pool of cheap PCs to build a scalable jabber network, then even more valuable then having One big Server with Gigs of memory. This will require some changes in the Jabber protocol. The MSN protocol does something like this, but takes this one step ahead by letting you connect to any server in the pool, which then refers you to the right server that can authenticate you. If you make logging in a two-step process, you can solve this problem but that would mean changing all the Jabber clients and also the S2S communication in the Jabber server. Anyway, this is an area that the Jabber server developers are the best folks to comment on. Regards, Ashvil ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
[JDEV] jabber-1.4.2test: network sockets sometimes not closed?
Hi! I'm running the 1.4.2test version of jabberd. Today I discovered a strange problem. When I make a netstat -n | grep 5222 I notice that from some IPs I have multible connections. From one I even had 199 TCP/IP connections at a time. When I looked into the record.log I saw that there should be no connection, the user had just logged in and out many times, he never was online with multible clients. - I also asked the user if he could imagine what that could be. He said that he's using the server just normal. For some other IPs I noticed up to 5 connections b ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
[JDEV] ldapv3 or kerberos mod for jabber server?
jdev, I was curious if I could spark some interest in the development of a module for the Jabber Server that could authenticate to ldapv3, which I believe is ldap + kerberos authentication protocol, probably some other stuff I'm not aware of. I understand there is an xdb_ldap ? BUT, in an ldapv3 implementation, it seems the userPassword property is not populated due to the fact that a separate kerberos database is kept for the passwords. This makes the ldap mod by itself ineffective. With connectios to the ldapv3, my thoughts on the goal would be to take away from Jabber's separate user db and use the ldap db of users as its userbase and place of authentication. I believe this is what happens with the ldap mod already available when connecting to a true ldap server? The problem then comes back to the authentication, which cannot occur in ldapv3 b/c of the kerberos password db. Ideas or thoughts? Thanks for you time! Dave David McDowell Network Administrator DynPro, Inc. 919.969.7076 x236 http://www.dynpro.com ~~ Kindness in another's trouble, courage in your own. ~~ Princess Diana ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] Some jabberd problems
I have noticed that my welcome message was being truncated as well and I'm running on a Solaris box. I also noticed that with the JIM client, in a conference room and sometimes with broadcast messages, that messages were truncated. The latter seems to happen when a single quotation mark is included in the message. I shortened my welcome message and have not since seen a recurrence. I have not seen the other errors that you describe though. Kevin [EMAIL PROTECTED] wrote: The first is that the welcome message: gets truncated to: I also get: Tue Jan 8 21:00:48 2002 modules.c:164 mapi_call returning unhandled in the debug output. Should I worry? Finally, although I can create a new user, and send messages to the machine, I can never log back in as that user. It appears that I get a 401 Unauthorized. Now I admit, whatever is causing the first problem could be causing other problems, but I thought I'd ask if this looked familiar to anyone... Unfamiliar errors to me. Anyone else? You're runningon HP-UX, right? Peter ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] Creating a new Transport/jsm?
http://docs.jabber.org/general/html/component-intro.html does that help? -mikem --- Peter Gebauer [EMAIL PROTECTED] wrote: Hey there! I've been reading the documentation found at jabber.org, however, starting to write your own transport or module is obviously harder than it should be. There should be a manual on this that explains how to build your own transport and / or module. I don't even know the difference between mod_echo and, say, irc transport. Are they both transports or is mod_echo a module? And if so, can I write my own modules that are not included under jsm/modules? Too many names are used for the same thing? Agent, Transport, Module, Component? What is what? I have a lot of unanswered questions in my head and none of the documents seem help me. Even though there are many documents, they are either for people installing already existing stuff or focus on instant messaging only where as I would like to use Jabber for more than just sending one-liners to AIM, SMS, ICQ, etc. If someone has any good suggestions on SPECIFIC documents that will aid me in my quest to create a Transport/Agent/Module/Component/whatever-the-heck-it's-supposed-to-be, appreciate it! Also, if someone could straighten out the terminology, that would be good. /P ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev __ Do You Yahoo!? Send FREE video emails in Yahoo! Mail! http://promo.yahoo.com/videomail/ ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: [JDEV] The Important Things
On Thu, Jan 10, 2002 at 10:18:05AM -0600, Peter Saint-Andre wrote: If I understand correctly, you mean that all the servers that speak the Jabber protocol will be able to deliver to other servers. Right now, jabberd's aren't configured to take connections from just about anyone, so we'll have to make changes in the paradigm or the protocol for that. ? One of the many advantages of Jabber is that it enables us to build out a network of servers, each connecting to the other. Of couse, you *can* turn off server-to-server communications, but the default is to communicate with other servers. Yup, I was mistaken. I got misled by my work which is focused on a single Jabber server :) Anyway, I'd like to help by working on code (to the best of my ability) that can tie up servers to form a network. Does it sound like a 'Freenet' or a 'Gnutella' to you? :) ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
[JDEV] 1.4.2-test2
Sorry about the big gap since the first test release, it's amazing how time flies when your not looking :) Anyway, snag a copy here: http://download.jabber.org/dists/1.4/final/jabber-1.4.2-test2.tar.gz I believe I have all the submitted patches and platform fixes included, as well as numerous bugs. It should work now on cygwin as well as 64bit systems. If it doesn't compile (or even if it gives you warnings) or is broken in some other way when you test it, send me an email to [EMAIL PROTECTED] with any necessary details. I'll start writing up the hefty changelog and features in preparation for a release within a week, barring any major bugs with test2. Thanks, Jer ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
Re: HTTP/Proxy access to 5222 (was Re: [JDEV] The Important Things)
The details are at http://jabber.org/?oid=971 and it actually works with the PUT or POST methods, and should work over ssl if you have your server built to support ssl as well. I think we discussed this before. This works with the Jabber 1.4.1 servers if the HTTP proxy in question keeps the connection open (mainly transparent type proxies). But most proxies close the connection after the response, so the client has to poll and requires some kind of token session id, ... etc. I assume that the http/wcs module that Peter mentioned solves this problem. My wish is that the public Jabber servers support it, so that Jabber use behind HTTP firewalls will be possible. Regards, Ashvil ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev
RE: [JDEV] The Important Things
Jer has said the capability is pretty much there, so I think the next thing is try to put together a proof of concept. My C is rusty and my day job is hectic, so I'm unlikley to be able to do anything for a while, so if you want to get stuck in feel free. Al. On Thu, 2002-01-10 at 22:45, Tim Ferguson wrote: I really, really like this idea, although I would not consider myself an expert yet I am willing to help out/move it along in this arena. Tim Ferguson -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Al Sutton Sent: Thursday, January 10, 2002 1:00 PM To: [EMAIL PROTECTED] Subject: Re: [JDEV] The Important Things Why not add authentication and message relaying to the S2S protocol. This would give four advantages; 1. Any user could log into any machine and the server would relay the authentication request to the relevant machine to handle authentication. 2. The messages for that user would be relayed to the server they are logged in to and then forwarded on to them. 3. Clusters or farms could be constructed to server a a single jabber community and the load shared between them. 4. This would only involve a change to the S2S protol and servers supporting it (of which there are few), and would leave the C2S protocol unchanged and thus not require any client changes. Comments? Al. On Thu, 2002-01-10 at 15:50, Ashvil wrote: I found that we could use some kind of a gateway - people connect to one server ( for example jabbber.org ) autheticate - get a token/session id - and then continue with a server l1.l4.dddljfds.jabbber.org that are real jabber servers. snip Any ideas that can help in scalability are welcome. If we can use a pool of cheap PCs to build a scalable jabber network, then even more valuable then having One big Server with Gigs of memory. This will require some changes in the Jabber protocol. The MSN protocol does something like this, but takes this one step ahead by letting you connect to any server in the pool, which then refers you to the right server that can authenticate you. If you make logging in a two-step process, you can solve this problem but that would mean changing all the Jabber clients and also the S2S communication in the Jabber server. Anyway, this is an area that the Jabber server developers are the best folks to comment on. Regards, Ashvil ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev ___ jdev mailing list [EMAIL PROTECTED] http://mailman.jabber.org/listinfo/jdev