SIP-based IMS Registration Analysis for WiMax-3G Interworking Architectures Arslan Munir and Ann Gordon-Ross* Department of Electrical and Computer Engineering University of Florida, Gainesville, Florida, USA e-mail: amunir@ufl.edu & ann@hcs.ufl.edu *Also with the NSF Center For High-Performance Reconfigurable Computing at the University of Florida, USA Abstract The 3rd generation partnership project (3GPP) and 3GPP2 have standardized the IP multimedia subsystem (IMS) to provide ubiquitous and access network indepen- dent IP-based services for next generation networks via merging cellular networks and the Internet. The applica- tion layer Session Initiation Protocol (SIP), standardized by 3GPP and 3GPP2 for IMS, is responsible for IMS ses- sion establishment, management, and transformation. The IEEE 802.16 worldwide interoperability for microwave ac- cess (WiMax) promises to provide high data rate broadband wireless access services. In this paper, we analyze the SIP- based IMS registration signaling delay for 3rd generation (3G) cellular and WiMax networks with specific reference to their interworking architectures. We also explore the ef- fects of different WiMax-3G interworking architectures on the IMS registration signaling delay. 1. Introduction The 3rd generation partnership project (3GPP) and 3GPP2 have standardized the IP multimedia subsystem (IMS) to provide IP-based rich multimedia services as well as content-based monetary charges for next generation net- works. The application layer Session Initiation Proto- col (SIP), standardized by 3GPP and 3GPP2 for IMS, is responsible for IMS session establishment, management, and transformation. The Internet Engineering Task Force (IETF) developed signaling compression (SigComp) for text-based protocol compression [12]. Even though the IEEE 802.16 worldwide interoperability for microwave ac- cess (WiMax) offers high data rates, WiMax coverage is limited when compared with 3G networks, which provide ubiquitous connectivity. On the other hand, 3G network data rates are much lower than WiMax data rates. The complementary coverage area and data rate characteristics of WiMax and 3G networks make their interworking, with the intent of ubiquitous high speed wireless data access, an interesting topic. Two popular system models for interworking wireless access networks (ANs) (such as 3G with WiMax) are tight and loose coupling. In tightly coupled systems, the con- necting AN integrates with the core 3G network similarly to any other 3G radio AN, using the same authentication, mobility, and billing infrastructures. To communicate with the 3G network, the connecting AN implements 3G radio access network protocols to route traffic through core 3G elements. In loosely coupled systems, a connecting AN in- tegrates into the core 3G network by routing communica- tion traffic through the Internet, with no direct connection between the two networks. The two networks use differ- ent authentication, billing, and mobility protocols, however, they may share the same subscriber databases for customer record management. Tightly and loosely coupled architec- tures are proposed and their cost analyzed in [7] and [8], respectively. Also the interworking architectures that in- tegrate 3G, WiMax, WLAN, and satellite ANs with IMS support are proposed and cost analyzed in [11]. An important aspect of interworking architectures is their IMS signaling efficiency, which is determined by the interworking architecture’s ability to carry out the IMS sig- naling procedures (i.e., session establishment, registration, termination, and transformation) with minimum delay and overhead. Previous work provides limited IMS signaling delay analysis. [9] studied the IMS session establishment procedure when both the source node (SN) and the corre- spondent node (CN) are in CDMA2000. In this paper we consider a more general case where SN and CN are in dif- ferent ANs. [5] studied SIP-based voice over IP (VoIP) IMS session establishment delay for 3G wireless networks using an adaptive retransmission timer for lost packet re- transmission. The authors studied different protocols such as transmission control protocol (TCP), user datagram pro- tocol (UDP), and radio link protocol (RLP). [10] studied