IJITN Editorial Board

This paper is devoted to modeling and simulation of traffic with integrated services at media gateway nodes in the next generation networks, based on Markov reward models (MRM). The bandwidth sharing policy with the partial overlapped transmission link is considered. Calls arriving to the link that belong to VBR and ABR traffic classes, are presented as independent Poisson processes and Markov processes with constant intensity and/or random input stream, and exponential service delay time that is defined according to MRM. Traffic compression is calculated using clustering and learning vector quantification (e.g., self-organizing neural map). Numerical examples and simulation results are provided for communication networks of various sizes. Compared with the other methods for traffic compression calculations, the suggested approach shows substantial reduction in numerical complexity. offer unrestricted access to different service providers and support generalized mobility that allows consistent and ubiquitous provision of services to users. The NGNs that we consider in this paper can be defined with the following fundamental characteristics, including: • Packet-based transfer; • Control functions that are separate for bearer capabilities, call/session, and application/service; • Service provision that is largely independent from the network; DOI: 10.4018/jitn.2011070101 2 International Journal of Interdisciplinary Telecommunications and Networking, 3(3), 1-14, July-September 2011 Copyright © 2011, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. • Support for a wide range of services, applications and mechanisms, based on service building blocks that include realtime/streaming/non-real-time services, and multi-media; • Open interfaces; • Broadband capabilities that provide endto-end QoS, and transparency; • Ability to interconnect with legacy networks; • Generalized mobility support; • Converged services between Fixed/ Mobile; • Backward compatibility and support for IP based addressing schemes, including for a variety of IP address recognition schemes designed for routing in IP networks; and • Unrestricted access to different service providers. Furthermore, the next generation networks, considered in this paper, have to support unified service characteristics as well as convergence of broadcast and telecommunications. The NGNs architecture is layered, including the transport layer and the service layer, with the boundaries that are strictly defined, and with the following interfaces that have to be available: • User-to-Network Interface (UNI); • Network-to-Network Interface (NNI); • Application-to-Network Interface (ANI). The transport layer provides a connection between the outer NGN elements (such as, for example, the user terminals), and the elements located at the NGN servers (such as, for example, the databases and media gateways), with access that fully depends on the applied technology. For example, fixed access can be provided through the DSL and wired LAN, and wireless access can be provided through the WiFi, WiMAX, and CDMA. The service layer provides session and other related services and delivery methods, as soon as the media gateway nodes (MGW) represent the above interfaces between the NGNs and other networks. This paper is devoted to modeling and simulation of traffic with integrated services at the media gateway nodes, based on Markov reward models (MRM), using clustering and learning vector quantification, e.g., self-organizing neural map (SOM). Compared with the other methods for traffic compression calculations, this approach provides substantial reduction in numerical complexity. next GenerAtIon networkS And MedIA GAtewAy nodeS There are various views on next generation networks, which have been introduced by operators, manufacturers and providers, and that have been subject of research (Cochennec, 2002; Fazekas et al., 2002; Radev & Lokshina, 2008; Lokshina & Bartolacci, 2008). The foremost NGN concept is based on integration of currently divided voice and data networks into a simpler and more flexible IP-based network, where the transport, control and service layers are independent and interact via open interfaces. At the same time, all IP networks allow different access options seamlessly integrated with an IP network layer. Particularly, the next generation networks contain both wired and wireless access networks. The most important NGN requirements include simplicity to provide new services, portability and accessibility through different networks, and support for Quality of Service. A most popular access to the NGNs is based on the media gateways with changing transfer and switching. The media gateway nodes are often implemented as independent devices; however, they can also be integrated in another system. In the traditional circuit-switched networks, the intelligence is concentrated in parts of the core of the network (e.g., in specific central switches). International Journal of Interdisciplinary Telecommunications and Networking, 3(3), 1-14, July-September 2011 3 Copyright © 2011, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. In the NGN model, the intelligence for transfer and switching is expected to be decentralized, including to the edge of the network. At this point, the NGN architecture is conceived to achieve an independence of the applications and services from the basic switching and transport technologies. The entire independence of the applications and control mechanisms from the access and transport layers represents the fundamental feature of the next generation network architecture. That can possibly be achieved with migration of the applications and call functions to open platforms, and introduction of common control protocols supporting communication between the control functions and network resources. Particularly, that can be obtained with the gateways providing a conversion between different communication media and providing protocol adaptations. The NGN with open architecture consists of the next three main layers: • Connectivity layer; • Control layer; and • Application and service layer. The connectivity layer consists of the following key elements: • Multi-Service Core: the IP-based transport backbone that carries multiple services over high-speed optical links. This part of the network acts as a long haul transport system providing connectivity among geographically distributed nodes, and is shared by such services as, for example, the phone calls, Web sessions, video-conferences, multi-player games, and movies. • Gateway Elements: they are needed to convert the information between different standards and representations. • Access Segment: consists of various different broadband access technologies (e.g., the xDSL, broadband wireless, optical technologies, etc.). The control layer (e.g., the call control) is clearly independent from the transport (physical) layer, which provides open and programmable interfaces towards the independent application layer that seamlessly mediates between the signaling protocols of different interconnected networks. The access layer includes the both wired and wireless network technologies. The core transport network might be built around Dense Wave Division Multiplexing (DWDM) transport system. The media gateways and soft-switches are important parts of the NGN architecture. The media gateways can be employed to interconnect networks based on different representation of the same signal. The multi-service softswitches are common elements of the control layer, which are able to operate in spite of different protocols they have to mediate between. The soft-switches are designed as software applications that run on the server or switch to manage the MGW switching activities. The MGW nodes are located at the ends of the next generation network, and consist of the following important components: • Interface with the networks with circuit switching (e.g., TDM network); • Interface with the packet networks (e.g., LAN connection); • Digital Signal Processor (DSP) for signal processing between circuit-switched networks and packet networks. There are three categories of the media gateway nodes in dependence of their size: • Small Office/Home Office (SOHO) for small peripheral networks, including voice, VoIP, data and video devices; • Office, for medium size peripheral networks; • Provider or carrier grade with high capacity in terms of simultaneous sessions and aggregate bandwidth. 4 International Journal of Interdisciplinary Telecommunications and Networking, 3(3), 1-14, July-September 2011 Copyright © 2011, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. bAndwIdth ShArInG ModelS wIth PArtIAl overlAPPed trAnSMISSIon lInk At MedIA GAtewAy nodeS The partial overlap of the bandwidth sharing model is defined in the following way: traffic of service i obtains part of the bandwidth equal to rimi bandwidth units, and the rest of traffic classes concur for sharing the rest of the link capacity C-n1m1-n2m2 bandwidth units, where C is the whole capacity. The input traffic is described as traffic of service i, if it has mi existing items in reserved capacity of rimi bandwidth units, or in sharing capacity of C-n1m1-n2m2 bandwidth units; otherwise connection is blocking and lost (Balsamo et al., 2001; Radev & Lokshina, 2007b). In this way, the schemes for access in full sharing and full separating of the traffic flows can be introduced as particular cases of the partial overlap scheme, where for r1=r2=0, full sharing is obtained, and for r1m1+r2m2=C, full separating is obtained (Gross & Harris, 1998). The obtained function for retranslation with the partial overlapped transmission link for (n1,n2)∈Ω can be presented according to (1).