 Tukaj. Zvukam izgledano rečenčaka S whipped server, zelo več jaznjenja, da se je tudi prav dešli bo bo У nekaj evoje z vse, ne bočnje neče začetnjavne interfece in dejščke. Zrani sem glasba bil in tudi in mišljimo. Srečo, o srečenju, smo zelo vsečnji institut in Berlinu, prijev njih razdaj. Zanimar, projekto je sebexpejzroutor, ki je to še vsečnja, implementacija za srečenje, in je tudi nekaj referens in opensučnjej. We created a branch by 2005, during some, let's say, more political discussion within the old project, which was named OpenServe, and actually we discovered that you get a quite good and big project when someone suze you because they say you infringe some trademarks, so we had to change the name in Kamailio by 2008, because U.S. company claim we infringe the third trademark they have registered. However, although we started a branch in 2005, now we are working again with the original team under so-called SIPR Outer project starting by 2008, because we discovered that we can share a lot of knowledge and efforts in the SIP area, and we keep kind of direction, self-direction, mainly related to user provisioning and architecturing of the back ends as extensions. So, just to give you a feeling of how the web pages are looking for all these projects and sites, Kamailio org and SIPR Outer org, with short-to-screen shorts down. In terms of development, we are using Linux, we use C, because it's a real-time communication engine, we have to be fast, we focus on scalability, just to give you an idea, right now we are able to set up a couple of thousand calls per second, we talk here about set up, so imagine 10,000 people calling in the same time, and one usual server can deal with it, and we use all three tools available there to build very fast libraries in terms of parsing SIP messages, routing, doing list, cost routing, doing load balancing, doing location of the users, all those core features for a SIP server. As administration and organization of the project, we are very community-driven, there is no big company behind the project, we have kind of management board based on individuals from different companies, from different countries, so we are pretty worldwide established, and we don't have actually much activity there, because we try to focus on development, not on political discussions. Last year was a pretty good one, and we got some nice awards out there, and it was not by luck, because the project right now serves a couple of most biggest deployments, and we have some of those that are in charge or guilty by deploying this in some corners of this room. I want just to mention I290 in Germany, one-to-one that right now they have over 3 million subscribers, because they just integrated with FreeNet, so probably they are over 3, more close to 4. SIPgate, another very popular voice provider in Germany, it's using a solution based on this project, but also other projects that are present over here after my talk, like Asterix and other examples, just to give you an idea. Telio in Norway and Switzerland, Trufone, which is actually doing Voice Over IP for mobile devices, which is a very interesting right now, and mobile communication is going to grow more and more. Shorty about the features, because right now we have over 150 extensions, so I cannot list all of them. It's basically a SIP server, we don't deal with media, it's about signaling level and most of the work we do in the application is when the call is established to do authentication of the user, to do location discovery for the destination, to apply all these fancy rules, right now you want for your calls, if you are in vacation already directed to a voicemail or to your secretary or whatsoever, and applying a lot of filters, maybe you don't want to allow calls from the people that are not in your contact list and so on and so forth. And we are going more and more as an application server because communication is no longer voice or video, a lot of people like to add more instant messaging, presence like you saw in the previous presentation, these are very important aspects of today's communication and of course integration with social networking. We are not a SIP phone because there are other guys doing this better like Emil and the SIP communicator team and also we are not a media server or a PBX or a back-to-back user agent because there are others doing this good job and they are going to present after me. So we are somehow in the middle, we connect users with services and we tell platforms to the large number of subscribers. One of the most important as everywhere with the project is the interface because this allows to add a lot of features without affecting the core. We are in telecommunication business where the most important aspect is to be fast and connect because it's not like with email you stay and listen to massing for about 10 seconds you said, oh, what happens? You have to give a feedback quickly because this is real-time communication. Also telecommunication is for emergency services if someone is going to die, is not going to wait one minute to get okay, we are trying to find the closest service to you. So we try to keep a reliable small core and if you have fancy ideas you can implement it as module. We implement all these basic requirements from the CIP protocol, like CIP proxy redirect registrar, then we have extension for the telephony service like integration with enum numbering or connection to LDAP directories, advanced routing for list code routing, lot balancing, couple of variants, we do support a lot of transport layers. If you want to communicate secure, you can use TLS. If you have a smaller one, you can use UDP and we have a lot invest in the last month in the SCTP because skills in terms of multi-streaming, multi-homing, this kind of features. We are not tied to a specific backend because we have to host a lot of records. Think about username, password, blacklist, whitelist, where to redirect your mobile phones so you can use any kind of database which is more appropriate for you. You can interact with the service via control interface like XMRPC. We have some extension to gateway with XMPP and other open specified communication protocol. Now I'm going to present and this is the main topic for this presentation. The latest version, 30, which is three weeks old by now, which was actually the result of the collaboration within the CBRAUTER project just to give you a comparison with CBRAUTER project because now we have Kamaelio and CIP Express Router coming out of CIP Router project is more like with Linux distribution. We have the kernel, which is the parser, the router and then we have a set of modules that we package and release Kamaelio and there is another set of extensions that other team release as CIP Express Router. The first thing was to clean up a bit the architecture because there are a lot of deployments using the project on embedded systems to optimize the traffic, for example, in Wi-Fi networks to find the best path to a service and there you have to be very small. So we moved out a couple of core components as libraries to make it more flexible and smaller in footprint. So actually the core, if you think like Swedish buffet doesn't do much, it's like an empty plate and then if you want to be a very scalable load balancer to dispatch thousands of calls per second you just load dispatcher module. You need a list cause router to route calls to the cheapest alternative of your termination providers then you use appropriate module. So basically you can be very, very small in providing some services in footprint. As I said, a lot of effort was on the transport layer because in the past it was about voice. That was very easy, you just established a connection and hard job was done by the RTP. But now we have instant messaging, we have a lot of presence, we have people with hundreds of contacts so now we have to use more and more TCP to exchange this information and of course we have to protect their privacy so we actually have to use TLS. And now we have the platform bigger and bigger in terms of number of servers so inside the platform you can use a CTP with multi-homing, multi-streaming that gives a very good scalability service. Also in terms of hiding your internal network there is an extension to mask some privacy headers into the C messages topology hiding and also number portability as more and more prefer to take their number from provider to provider. DNS caching, you know, all basic actually the root of the routing over the internet is DNS, if DNS is low you feel like connection is slow. So we have here very good implementation to speed up all this one better control interface via XML RPC reload of the config, you don't need to restart if you want to change parameter of the configs. A better structure of the routing block via include files, dev and dev, some parts of the config, you know the C style of defines fallback from very fast or lower resource usage transport protocols like UDP to TCP or TLS if it's the case. So quite a lot of improvements it's hard to list all of them and beyond VoIP we do a lot of extensions in the terms of presence and instant messaging even enable all telephony devices via user location to show active online, offline states. Links if you want to check more about what's new and although it's like three weeks old by now it's still a lot of new for the upcoming release 3.1 and I invite you to visit the websites not all of them are listed here because in the last week we have like two other addition Python, API and debugger for the configuration file. So this is let's say quick run to the project because it's a lot to talk about it and I'm just in time I will take questions on side of the room or outside if you have. Thank you.