Re: multiple bearer box - Netikos?
Nisan Bloch wrote: At 11:37 AM 2/25/03 +0100, Stipe Tolj wrote: Kalle Marjola wrote: That's why I published them, altought I knew that some things are a bit too radical. As we have no resources to do any development on it right now, I hope that you can scavenge useful things out from it and this way improve the Kannel project. yep, +1 :) that's what we all do. ditto +1 i dont have much spare time, but I could over the next week or so make a list of those pieces that we can move over basically transparently (eg http libs) and those that would be relatively easy (eg adding the sms-service rules from NMGW) and those that cannot be done without some major changes to Kannel. ok, go for it. Listen anyway to the list, because we may start already with sync'ing some parts from Netikos version to the official tree. Stipe [EMAIL PROTECTED] --- Wapme Systems AG Vogelsanger Weg 80 40470 Düsseldorf Tel: +49-211-74845-0 Fax: +49-211-74845-299 E-Mail: [EMAIL PROTECTED] Internet: http://www.wapme-systems.de --- wapme.net - wherever you are
Re: multiple bearer box - Netikos?
On Sun, 23 Feb 2003, Stipe Tolj wrote: Nisan Bloch wrote: I think rolling NMGW into Kannel would be worthwhile. At the same time come up with a combined bbox+smsbox version, with the same HTTP interface. In addition an API to build XXXboxs with. Smsbox would be the first such app and maybe the smppbox and emibox that have been mentioned on the list. definetly. We should merge things from Netikos version into official Kannel. Main question would be how many efforts Netikos is *still* doing in continued development of their branch? Kalle? Practically nothing has been done since I pusblished that code. You guys could switch off to the official tree and get your great things included there. This would cause a benefit for all. That's why I published them, altought I knew that some things are a bit too radical. As we have no resources to do any development on it right now, I hope that you can scavenge useful things out from it and this way improve the Kannel project. -- kalle marjola
Re: multiple bearer box - Netikos?
Kalle Marjola wrote: That's why I published them, altought I knew that some things are a bit too radical. As we have no resources to do any development on it right now, I hope that you can scavenge useful things out from it and this way improve the Kannel project. yep, +1 :) that's what we all do. Stipe [EMAIL PROTECTED] --- Wapme Systems AG Vogelsanger Weg 80 40470 Düsseldorf Tel: +49-211-74845-0 Fax: +49-211-74845-299 E-Mail: [EMAIL PROTECTED] Internet: http://www.wapme-systems.de --- wapme.net - wherever you are
Re: multiple bearer box - Netikos?
At 11:37 AM 2/25/03 +0100, Stipe Tolj wrote: Kalle Marjola wrote: That's why I published them, altought I knew that some things are a bit too radical. As we have no resources to do any development on it right now, I hope that you can scavenge useful things out from it and this way improve the Kannel project. yep, +1 :) that's what we all do. ditto +1 i dont have much spare time, but I could over the next week or so make a list of those pieces that we can move over basically transparently (eg http libs) and those that would be relatively easy (eg adding the sms-service rules from NMGW) and those that cannot be done without some major changes to Kannel. Nisan Stipe [EMAIL PROTECTED] --- Wapme Systems AG Vogelsanger Weg 80 40470 Düsseldorf Tel: +49-211-74845-0 Fax: +49-211-74845-299 E-Mail: [EMAIL PROTECTED] Internet: http://www.wapme-systems.de --- wapme.net - wherever you are
Re: multiple bearer box - Netikos?
Nisan Bloch wrote: I think rolling NMGW into Kannel would be worthwhile. At the same time come up with a combined bbox+smsbox version, with the same HTTP interface. In addition an API to build XXXboxs with. Smsbox would be the first such app and maybe the smppbox and emibox that have been mentioned on the list. definetly. We should merge things from Netikos version into official Kannel. Main question would be how many efforts Netikos is *still* doing in continued development of their branch? Kalle? You guys could switch off to the official tree and get your great things included there. This would cause a benefit for all. Stipe [EMAIL PROTECTED] --- Wapme Systems AG Vogelsanger Weg 80 40470 Düsseldorf Tel: +49-211-74845-0 Fax: +49-211-74845-299 E-Mail: [EMAIL PROTECTED] Internet: http://www.wapme-systems.de --- wapme.net - wherever you are
Re: multiple bearer box ?
Nisan Bloch wrote: At 10:29 AM 2/20/03 +0100, Stipe Tolj wrote: ok, as all of you kick-ass with benchmark figures, I'll do too ;)) The max. throughput we reached with a Dell PowerEdge 2x CPU (1.2 GHz) Linux 2.4 Kernel in a fakesmsc - bearerbox - smsbox chain was approx. 2400 msg/sec. We had also MO values while loading the bearerbox quueue (fakesmsc - bearerbox) with over 1 msg/sec. So I guess this is pretty fast :) that is - but tell me were you using dlr acks? And if so internal or mysql store? nop, none of them ;) Stipe [EMAIL PROTECTED] --- Wapme Systems AG Vogelsanger Weg 80 40470 Düsseldorf Tel: +49-211-74845-0 Fax: +49-211-74845-299 E-Mail: [EMAIL PROTECTED] Internet: http://www.wapme-systems.de --- wapme.net - wherever you are
RE: multiple bearer box ?
Hi We have in total 14 connections to different operators SMSC in thee different countryes and in short term we have reached about 150-200 msg/sec. Illimar I think my peak was in a campaign that was delayed and then I needed every resource available to do it and I had something like 50+15+5+3+2+several 1 msg/sec connections, which gives like 80 msg/sec -- Davi / Bruno.RodriguesatLitux.Org Litux.org: 20:36:09 up 88 days, 21:51, 4 users, load average: 0.85, 0.54, 0.36 'AUTHOR FvwmAuto just appeared one day, nobody knows how. -- FvwmAuto(1x)'
Re: multiple bearer box ?
I once stress tested stand-alone bearerbox. (Standalone meaning I was using smsboxes that did not do any fetch, but returned a standard text message instead.) I got value 1200 msg/s and must use ~20 smsboxes before I reached 100 % cpu utilisation. Cpu was AMD Athlon, 800 MHz. Aarno On Thursday, February 20, 2003, at 08:58 AM, Illimar Reinbusch wrote: Hi We have in total 14 connections to different operators SMSC in thee different countryes and in short term we have reached about 150-200 msg/sec. Illimar I think my peak was in a campaign that was delayed and then I needed every resource available to do it and I had something like 50+15+5+3+2+several 1 msg/sec connections, which gives like 80 msg/sec -- Davi / Bruno.RodriguesatLitux.Org Litux.org: 20:36:09 up 88 days, 21:51, 4 users, load average: 0.85, 0.54, 0.36 'AUTHOR FvwmAuto just appeared one day, nobody knows how. -- FvwmAuto(1x)'
Re: multiple bearer box ?
On Thu, 20 Feb 2003, Aarno Syvänen wrote: I once stress tested stand-alone bearerbox. (Standalone meaning I was using smsboxes that did not do any fetch, but returned a standard text message instead.) I got value 1200 msg/s and must use ~20 smsboxes before I reached 100 % cpu utilisation. Cpu was AMD Athlon, 800 MHz. With this 'a bit optimized' merged bearerbox and smsbox of ours, I recall values of 1000-2000 msg/s in modern workstation (fakesmsc/simple answer). It is good to notice that main things that create bottlenecks are: - smsc drivers/smsc itself - smsbox utilities (http library) The current Kannel is, however, made in that way that it gets trashed if it receives messages faster than it can handle, so beware :] -- kalle marjola
Re: multiple bearer box ?
Alexander Malysh wrote: Not 100% true ;) We run one logical link per bearerbox (e.g. 3 sessions for one offerer). For now I have not seen any performance bottleneck issues with bearerbox. Some of our bearerboxes running with ~60sms/sec. Now you will ask, why multiple bearerboxes ;) It's easy... It's more simple to maintain/route messages beetwen multiple bearerboxes. For example bearerbox do not support dynamic config reload and it's just more simple to read/write ~30 Lines of config file as 300 lines ;) that's true. That's why we implemented the 'include = file' feature for the config file handling, so you can structurize the global config file for your own needs. It's like you setup an apache config for every VHost you'd like to run. The best argument for the one bearerbox per logical link ;) is indeed the lacking of hot-restart functionability. AFAIK, the Netikos version had such things. We should definetly start to port things from their version to the official tree. Stipe [EMAIL PROTECTED] --- Wapme Systems AG Vogelsanger Weg 80 40470 Düsseldorf Tel: +49-211-74845-0 Fax: +49-211-74845-299 E-Mail: [EMAIL PROTECTED] Internet: http://www.wapme-systems.de --- wapme.net - wherever you are
Re: multiple bearer box ?
Aarno Syvänen wrote: I once stress tested stand-alone bearerbox. (Standalone meaning I was using smsboxes that did not do any fetch, but returned a standard text message instead.) I got value 1200 msg/s and must use ~20 smsboxes before I reached 100 % cpu utilisation. Cpu was AMD Athlon, 800 MHz. ok, as all of you kick-ass with benchmark figures, I'll do too ;)) The max. throughput we reached with a Dell PowerEdge 2x CPU (1.2 GHz) Linux 2.4 Kernel in a fakesmsc - bearerbox - smsbox chain was approx. 2400 msg/sec. We had also MO values while loading the bearerbox quueue (fakesmsc - bearerbox) with over 1 msg/sec. So I guess this is pretty fast :) Stipe [EMAIL PROTECTED] --- Wapme Systems AG Vogelsanger Weg 80 40470 Düsseldorf Tel: +49-211-74845-0 Fax: +49-211-74845-299 E-Mail: [EMAIL PROTECTED] Internet: http://www.wapme-systems.de --- wapme.net - wherever you are
Re: multiple bearer box ?
Kalle Marjola wrote: The current Kannel is, however, made in that way that it gets trashed if it receives messages faster than it can handle, so beware :] yes, in some sense. But that would mean you have an permanent(!) input stream of 100-200 msg/sec. and this is very unlikely. Of course you could have peaks with such or even higher values, but I don't think anyone of us here arround has a permanent load of 200 msg/sec., or am I wrong? Stipe [EMAIL PROTECTED] --- Wapme Systems AG Vogelsanger Weg 80 40470 Düsseldorf Tel: +49-211-74845-0 Fax: +49-211-74845-299 E-Mail: [EMAIL PROTECTED] Internet: http://www.wapme-systems.de --- wapme.net - wherever you are
Re: multiple bearer box ?
Am Donnerstag, 20. Februar 2003 09:35 schrieb Kalle Marjola: The current Kannel is, however, made in that way that it gets trashed if it receives messages faster than it can handle, so beware :] I have sent patch to the list for avoiding this issue , but unfortunately it was not accepted ;( I will rework this patch and resend to the list. We use this patch in production and do not have any problems that kannel trashed. -- Best regards / Mit besten Grüßen aus Köln Dipl.-Ing. Alexander Malysh ___ Centrium GmbH Ehrenstraße 2 50672 Köln Fon: +49 (0221) 277 49 240 Fax: +49 (0221) 277 49 109 email: [EMAIL PROTECTED] web: www.centrium.de msn: [EMAIL PROTECTED] icq: 98063111
Re: multiple bearer box ?
Alexander Malysh wrote: Am Donnerstag, 20. Februar 2003 09:35 schrieb Kalle Marjola: The current Kannel is, however, made in that way that it gets trashed if it receives messages faster than it can handle, so beware :] I have sent patch to the list for avoiding this issue , but unfortunately it was not accepted ;( I will rework this patch and resend to the list. We use this patch in production and do not have any problems that kannel trashed. please re-send and also quote which mail thread contained your orginal patch, so we can review why we didn't commit it yet. Stipe [EMAIL PROTECTED] --- Wapme Systems AG Vogelsanger Weg 80 40470 Düsseldorf Tel: +49-211-74845-0 Fax: +49-211-74845-299 E-Mail: [EMAIL PROTECTED] Internet: http://www.wapme-systems.de --- wapme.net - wherever you are
Re: multiple bearer box ?
On Donnerstag, Februar 20, 2003, at 10:32 Uhr, Stipe Tolj wrote: Kalle Marjola wrote: The current Kannel is, however, made in that way that it gets trashed if it receives messages faster than it can handle, so beware :] yes, in some sense. But that would mean you have an permanent(!) input stream of 100-200 msg/sec. and this is very unlikely. Of course you could have peaks with such or even higher values, but I don't think anyone of us here arround has a permanent load of 200 msg/sec., or am I wrong? Stipe we once tested Kannel in a loop by some simple SMPP to HTTP convertor. so the loop was: http->smsbox->bearerbox->SMPP driver -> SMPP to HTTP back to the origin. we then injected a few SMS via HTTP and the whole chain started to oscillate at around 400msg/sec on a 500Mhz Celeron Linux machine. So this was constant load. Andreas Fink Global Networks Switzerland AG -- Tel: +41-61-333 Fax: +41-61-334 Mobile: +41-79-2457333 Global Networks, Inc. Clarastrasse 3, 4058 Basel, Switzerland Web: http://www.global-networks.ch/ [EMAIL PROTECTED] -- Member of the GSM Association
Re: multiple bearer box ?
Am Donnerstag, 20. Februar 2003 11:44 schrieb Stipe Tolj: Alexander Malysh wrote: Am Donnerstag, 20. Februar 2003 09:35 schrieb Kalle Marjola: The current Kannel is, however, made in that way that it gets trashed if it receives messages faster than it can handle, so beware :] I have sent patch to the list for avoiding this issue , but unfortunately it was not accepted ;( I will rework this patch and resend to the list. We use this patch in production and do not have any problems that kannel trashed. please re-send and also quote which mail thread contained your orginal patch, so we can review why we didn't commit it yet. I need some time to make a new patch ;) Old version is at: http://www.mail-archive.com/devel@kannel.3glab.org/msg05721.html But some things have changed since first version ;) Stipe [EMAIL PROTECTED] --- Wapme Systems AG Vogelsanger Weg 80 40470 Düsseldorf Tel: +49-211-74845-0 Fax: +49-211-74845-299 E-Mail: [EMAIL PROTECTED] Internet: http://www.wapme-systems.de --- wapme.net - wherever you are -- Best regards / Mit besten Grüßen aus Köln Dipl.-Ing. Alexander Malysh ___ Centrium GmbH Ehrenstraße 2 50672 Köln Fon: +49 (0221) 277 49 240 Fax: +49 (0221) 277 49 109 email: [EMAIL PROTECTED] web: www.centrium.de msn: [EMAIL PROTECTED] icq: 98063111
Re: multiple bearer box - Netikos?
At 10:20 AM 2/20/03 +0100, Stipe Tolj wrote: Alexander Malysh wrote: that's true. That's why we implemented the 'include = file' feature for the config file handling, so you can structurize the global config file for your own needs. It's like you setup an apache config for every VHost you'd like to run. AFAIK, the Netikos version had such things. We should definetly start to port things from their version to the official tree. yup it does. I keep wanting to do this but then worry about getting off the current tree too much. For whats its worth here are some notes that I started on NMGW a while back. * I like the merged smsbox/bearerbox. * NMGW have much better handling of concat messages. I had a situation where a user posted through a udh of 264 bytes, this will crash the current Kannel. I havent tested this with your NMGW, but from looking at your code it seems you will handle it. * NMGW reconcats MO messages. * The current Kannel however has a much more complete EMI,SMPP and AT2 implementations, * The DLR handing is a tricky one. NMGW has a better implementation and abstraction layer for the dlr stuff, and it would be relatively easy to roll the exisiting Kannel DLR functionality into NMGW but not the other way. * Utoi's reworked HTTP lib is indeed substantially better. * Without doubt the extended charset / unicode support is great. * I see that you dont have mclass support in the sms push CGI - Kannel is ahead on smsbox/config file options. * Kannel has better control over stopping and starting individual SMSC connections from the HTTP admin interface, but NMGW reloads configs better. I dont think either can add new SMSC connections on the fly. I am also not sure how NMGW handles queued messages when restarting from HTTP interface. * A bit concerned about the lack of mysql support? Is the NMGW store file support super stable? It becomes critical for deliv acks to keep state between restarts of the gateway? * The NMGW sms-service stuff is very very nice. I like the power of the conditions and matching rules. And might be helpfull for handling deliv_acks. I think rolling NMGW into Kannel would be worthwhile. At the same time come up with a combined bbox+smsbox version, with the same HTTP interface. In addition an API to build XXXboxs with. Smsbox would be the first such app and maybe the smppbox and emibox that have been mentioned on the list. Nisan
Re: multiple bearer box ?
At 10:29 AM 2/20/03 +0100, Stipe Tolj wrote: ok, as all of you kick-ass with benchmark figures, I'll do too ;)) The max. throughput we reached with a Dell PowerEdge 2x CPU (1.2 GHz) Linux 2.4 Kernel in a fakesmsc - bearerbox - smsbox chain was approx. 2400 msg/sec. We had also MO values while loading the bearerbox quueue (fakesmsc - bearerbox) with over 1 msg/sec. So I guess this is pretty fast :) that is - but tell me were you using dlr acks? And if so internal or mysql store? Nisan Stipe [EMAIL PROTECTED] --- Wapme Systems AG Vogelsanger Weg 80 40470 Düsseldorf Tel: +49-211-74845-0 Fax: +49-211-74845-299 E-Mail: [EMAIL PROTECTED] Internet: http://www.wapme-systems.de --- wapme.net - wherever you are
Re: multiple bearer box ?
Citando Stipe Tolj [EMAIL PROTECTED]: Asif Ali wrote: hello all, i am new to this mailing list, i want to know if anybody's working on multi-bearer box architecture or there are any plans as it was identified as one of the key performance bottleneck issues, also the persistance of message streams as they pass through bearer box. hmm, I had that discussion with Alex from Centrium. They run an one bearerbox per smsc link architecture. It is definetly more performative if you get rid of all the required routing inside bearerbox if you use a bunch of smsc links inside one bearerbox, but it's more of theoretical interest. We run over 80 links in one bearerbox and fully stable. BTW, the performance bottle-neck will be of significanse when you get beyond 400 Msg/sec. (!) in MT/MO traffic, which I honestly don't expect you to have ;) We have only one bearerbox and one smsbox for 16 connections and alot of messages per day, and sometimes I'm scared when I enable a 50msg/sec on one of those connections for doing mass spam but then, while I'm sending them, I'm watching and testing the other connections and kannel works like a charm. Please note that usually I use post-xml to enable me to send like 10k destinations at one time, and thus have smsbox processing a xml post with 350KBytes, have the xml processing overhead, and have one smsbox thread creating that many messages in queue. What I mean it's might be a bigger peak processing than having 10k GET's and it works perfectly Stipe [EMAIL PROTECTED] --- Wapme Systems AG Vogelsanger Weg 80 40470 Düsseldorf Tel: +49-211-74845-0 Fax: +49-211-74845-299 E-Mail: [EMAIL PROTECTED] Internet: http://www.wapme-systems.de --- wapme.net - wherever you are -- Davi / Bruno.RodriguesatLitux.Org Litux.org: 13:27:05 up 88 days, 14:42, 2 users, load average: 0.14, 0.15, 0.09 'Linux is obsolete -- Andrew Tanenbaum'
Re: multiple bearer box ?
Bruno Rodrigues wrote: We have only one bearerbox and one smsbox for 16 connections and alot of messages per day, and sometimes I'm scared when I enable a 50msg/sec on one of those connections for doing mass spam but then, while I'm sending them, I'm watching and testing the other connections and kannel works like a charm. Please note that usually I use post-xml to enable me to send like 10k destinations at one time, and thus have smsbox processing a xml post with 350KBytes, have the xml processing overhead, and have one smsbox thread creating that many messages in queue. What I mean it's might be a bigger peak processing than having 10k GET's and it works perfectly cool. What was the highest peak (msg/sec.) you had in the bearerbox (over all connections)? Stipe [EMAIL PROTECTED] --- Wapme Systems AG Vogelsanger Weg 80 40470 Düsseldorf Tel: +49-211-74845-0 Fax: +49-211-74845-299 E-Mail: [EMAIL PROTECTED] Internet: http://www.wapme-systems.de --- wapme.net - wherever you are
Re: multiple bearer box ?
Citando Stipe Tolj [EMAIL PROTECTED]: What was the highest peak (msg/sec.) you had in the bearerbox (over all connections)? I think my peak was in a campaign that was delayed and then I needed every resource available to do it and I had something like 50+15+5+3+2+several 1 msg/sec connections, which gives like 80 msg/sec -- Davi / Bruno.RodriguesatLitux.Org Litux.org: 20:36:09 up 88 days, 21:51, 4 users, load average: 0.85, 0.54, 0.36 'AUTHOR FvwmAuto just appeared one day, nobody knows how. -- FvwmAuto(1x)'
Re: multiple bearer box ?
Hi, On Wednesday 19 February 2003 11:03, Stipe Tolj wrote: Asif Ali wrote: hello all, i am new to this mailing list, i want to know if anybody's working on multi-bearer box architecture or there are any plans as it was identified as one of the key performance bottleneck issues, also the persistance of message streams as they pass through bearer box. hmm, I had that discussion with Alex from Centrium. They run an one bearerbox per smsc link architecture. It is definetly more Not 100% true ;) We run one logical link per bearerbox (e.g. 3 sessions for one offerer). For now I have not seen any performance bottleneck issues with bearerbox. Some of our bearerboxes running with ~60sms/sec. Now you will ask, why multiple bearerboxes ;) It's easy... It's more simple to maintain/route messages beetwen multiple bearerboxes. For example bearerbox do not support dynamic config reload and it's just more simple to read/write ~30 Lines of config file as 300 lines ;) performative if you get rid of all the required routing inside bearerbox if you use a bunch of smsc links inside one bearerbox, but it's more of theoretical interest. We run over 80 links in one bearerbox and fully stable. BTW, the performance bottle-neck will be of significanse when you get beyond 400 Msg/sec. (!) in MT/MO traffic, which I honestly don't expect you to have ;) Stipe [EMAIL PROTECTED] --- Wapme Systems AG Vogelsanger Weg 80 40470 Düsseldorf Tel: +49-211-74845-0 Fax: +49-211-74845-299 E-Mail: [EMAIL PROTECTED] Internet: http://www.wapme-systems.de --- wapme.net - wherever you are -- Best regards / Mit besten Grüßen aus Köln Dipl.-Ing. Alexander Malysh ___ Centrium GmbH Ehrenstraße 2 50672 Köln Fon: +49 (0221) 277 49 240 Fax: +49 (0221) 277 49 109 email: [EMAIL PROTECTED] web: www.centrium.de msn: [EMAIL PROTECTED]
multiple bearer box ?
hello all, i am new to this mailing list, i want to know if anybody's working on multi-bearer box architecture or there are any plans as it was identified as one of the key performance bottleneck issues, also the persistance of message streams as they pass through bearer box. Can someone also guide me where can i find detailed status on the projects overal status Regards Asif __ Do you Yahoo!? Yahoo! Shopping - Send Flowers for Valentine's Day http://shopping.yahoo.com
