The convergence of voice (i. e. circuit switched) and data (i. e. packet switched) networks for oﬀering seamless voice and multi-media services is rapidly progressing. In order for packet based voice services to ﬁnd acceptance in the marketplace, however, it is paramount that the quality of service, as perceived by the end user of such services, is at least as good as that experienced in presentday circuit switched networks. While there are a number of other important aspects of convergence, such as addressing , this paper concentrates on the reliability and performance of signaling in a converging environment. For the PSTN/ISDN a number of strict reliability, quality, and performance requirements can be found in the E-series of Recommendations of the ITU-T, for example. One cornerstone for delivering these requirements is the reliability of the signaling network. This has been discussed extensively, for instance at the ﬁrst DRCN in  and  . Even where the convergence of circuit and data networks is not yet an issue, it is becoming more and more important to be able to combine classical SS7-based networks with IP-based networks using the latter to transport/tunnel SS7signaling messages. Deploying such a combined architecture enables operators to make use of the advantages of IP-based equipment in an SS7-based environment, avoiding some of the problems increasingly appearing in the rapidly growing SS7 networks, such as linkset capacity and load sharing . The standardization of a protocol suite for transporting SS7 signaling over IP networks is currently beingdeveloped by the Signaling Transport (SIGTRAN) working group of the Internet Engineering Task Force (IETF). In order for convergence and/or SS7 over IP becoming a success it is therefore important that the protocol suite developed by SIGTRAN can oﬀer the same or a better performance and reliability for signaling than the current SS7 network. This paper discusses some of the explicit and implicitperformance and reliability features of the SS7 transport layers (MTP levels 2 and 3) which together with proper signaling network planning and dimensioning deliver the performance and reliability required by the signaling applications. We will then show that using the protocol parameters given in  for the deployment of SCTP (the basic transport protocol of the SIGTRAN protocol suite) in thepublic Internet will not result in the necessary protocol behavior for signaling transport. Therefore it is necessary to use special parameter settings to be able to fulﬁll the SS7 performance requirements. The relation between these parameters is analyzed and values are suggested which can be used for signaling transport. Finally we will show that, when interworking between SS7 and the SIGTRANstack is on MTP level 3 (as opposed to on MTP level 2 or on the SCCP level), a number of restrictions to the allowed addressing options in the SIGTRAN stack are necessary in order to enable MTP’s network management to fully include the elements in the IP domain and thus to guarantee the reliability required from the signaling network.
∗ NeuStar, Inc., 45980 Center Oak Plaza, Sterling, VA 20166, USATel: +1 571 434 5652. Fax: +1 571 434 5601. e-mail: Klaus.Gradischnig@neustar.com. † Siemens AG, ICN WN CS SE 5, D-81359 M¨ nchen, Germany. u Tel: +49 89 722 47210. Fax: +49 89 722 48212. e-mail: Michael.Tuexen@icn.siemens.de.
The classical PSTN/ISDN separates the networks used for the transmission of the voice information of a call and the call control messages needed forcontrolling the calls. The latter network is called the signaling network and normally uses the protocols deﬁned for the signaling system no. 7 (SS7). These SS7 networks are considered in this paper. In the SS7 terminology the physical layer is called message transfer part level 1 (MTP level 1 or MTP1), the link layer is called message transfer part level 2 (MTP level 2 or MTP2) and the network...