Min-Xiou Chen

SSIP: Split a SIP session over multiple devices

Recently, most of Voice over Internet applications are based on the Session Initiation Protocol (SIP). This paper addresses issues of allowing a user to continue to communicate with a remote party while changing terminals over multiple devices. Specifically, we propose an SIP extension header to improve the Call Transfer mechanism and hide the changing of the terminal from the remote party. We also propose a mechanism to solve the problem of the user needing to terminate all devices separately when a session is split over multiple devices. Finally, the proposed mechanisms are implemented using Sip-Communicator, an open source of SIP.

An Efficient Adaptive Grouping for Single Code Assignment in WCDMA Mobile Networks

For achieving high utilization and efficient code management of the OVSF code tree in 3G WCDMA networks, several researches have extensively studied. Based on combining both the code assignment and the reassignment mechanisms, it increases obviously high utilization and reduces completely the code blocking. Nevertheless, the required rate of traffic should be powers of two of the basic rate, i.e. 1R, 2R, 4R, …, etc., which is impractical and results in wasting the system bandwidth while the required rate is not powers of two of the basic rate. Several multi-code assignment mechanisms have proposed to reduce the waste rate. Nevertheless, these methods bring two inevitable drawbacks including, high complexity of handling multiple codes, and increasing the cost of using more rake combiners at both the base stations and mobile nodes. Therefore, we propose an adaptive grouping code assignment herein to provide a single channelization code for any possible rate of traffic, even though the required rate is not powers of two of the basic rate. Based on the dynamic programming algorithm, the adaptive grouping approach forms several calls into a group. Then it allocates a subtree to the group and adaptively shares the subtree codes for these calls in the concept of time-sharing of slots during a group cycle time. Therefore, the waste rate and code blocking are thus reduced obviously while using a single rake combiner. Since the delay problem may be occurred in such a time-sharing approach, we propose two schemes of cycle interleaving methods to reduce delay. Numerical results indicate that the proposed adaptive grouping approach reduces significantly the waste rate and thus increases the system utilization. Moreover, the proposed cycle interleaving scheme reduces data delay significantly.

Single Code-Based Dynamic Grouping with Cycle Interleaving Algorithm for Reducing Waste Rate in WCDMA Cellular Networks

3G must offer high data rates since it should support real-time multimedia services; one performance enhancement, the use of the OVSF code tree, has adopted in 3G WCDMA networks. Unfortunately, this technique allows the link capacity to be set at the base rate times powers of two. This results in wasting bandwidth while the required rate is not powers of two of the basic rate. Several multi-code assignment mechanisms have been proposed to reduce the waste rate, but incur some drawbacks, including high complexity of handling multiple codes and increasing cost of using more rake combiners. Our solution is a dynamic grouping code assignment that allows any rate to be achieved with a single code for any possible rate of traffic. The dynamic grouping approach first forms several calls into a group. It then allocates a subtree to the group and dynamically shares the subtree codes based on time-sharing of slots within a group cycle time. The waste rate and code blocking is thus reduced significantly. Since transmission delay and jitter may occur in such a time-sharing approach, two schemes of cycle interleaving are proposed to minimize delay and jitter. Numerical results demonstrate that the proposed approach reduces the waste rate and increases the system utilization obviously, and the proposed cycle interleaving schemes minimizes delay and jitter significantly.

Fair and efficient packet scheduling algorithms for multiple classes of service under QoS guarantee in UMTS

Packet scheduling is the essential technique to provide QoS for multiple classes of service. Provided a minimum amount of bandwidth guarantee and a fair shared of residual bandwidth under a certain amount of power are the two major objectives of the packet scheduling in 3G networks. This paper presents two scheduling algorithms, MWF^2Q+ and MDRR, for multiple classes of service in the forward link of the UMTS network. These scheduling algorithms can allocate bandwidth in proportion to weights of flows sharing the channel under a certain amount of transmission power, and assign OVSF code to backlogged flows on a frame-by-frame basis. The MWF^2Q+ algorithm has better fairness properties while the MDRR algorithm requires less computational complexity. The efficiency and fairness properties are verified by our performance analysis. Our simulation results show that these two algorithms support multiple traffic sources with heterogeneous rate guarantees while fully utilizing the system bandwidth. The impact of self-similar traffic is also addressed in our simulations.

MDP-based OVSF code assignment scheme and call admission control for wideband-CDMA communications

Since the orthogonal characteristic of the orthogonal variable spreading factor (OVSF) code tree in wideband CDMA (WCDMA) systems, code blocking increases as traffle load or required rate increases. This causes inefficient utilization of channelization codes. Hence, how to efficiently manage the resource of channelization codes of OVSF code tree in WCDMA is an important issue and has been studied extensively. Therefore, in this paper we focus on how to assign channelization code efficiently. Additionally, most researches did not consider the analysis of tree state with dynamic traffic load and lack of systematic call admission control (CAC) mechanism. Therefore, in this paper, we first propose the Markov decision process (MDP) based analysis to assign channelization codes efficiently. Next, we extend the MDP-based approach as the call admission control mechanism to maximize system revenue while reducing blocking. Numerical results indicate that the proposed MDP approach yields the best fractional reward loss and code blocking reward loss as compare to that of others, including the random, left most, and crowded first schemes, and the MDP-based CAC outperforms the capacity-based CAC significantly.

Integrating service discovery technologies in OSGi platform

This paper describes the service discovery and interaction for home network devices using heterogeneous standards and protocols. OSGi was proposed to allow several kinds of services coming from different providers to be loaded and run on a gateway. We present a residential gateway based on the OSGi architecture for a smart home network. We combine the SLP SA/DA, the UPnP control point and the SIP UA into the gateway to achieve automated device discovery, registry, and management. Application examples are introduced and the implementation results show that our gateway can provide automatic heterogeneous service or device discovery, registry, and management.

Markov-Based OVSF Code Assignment Scheme and Call Admission Control for Wideband-CDMA Communication Systems

For the reason of the orthogonal characteristic of the Orthogonal Variable Spreading Factor (OVSF) code tree in Wideband CDMA (WCDMA) systems, code blocking increases as traffic load (i.e. Erlang load) or the required rate increases. This causes inefficient utilization of channelization codes. Hence, how to efficiently manage the resource of channelization codes of the OVSF code tree in WCDMA systems is an important issue and has been studied extensively. There are two aspects to achieve efficiency including code assignment and code reassignment. In the aspect of code assignment, an efficient code assignment scheme reduces code blocking probability significantly. In the aspect of code reassignment, code reassignment results in several drawbacks, such as large overhead of computation, high complexity of codes moving, and long call setup time for a new request call, etc. Therefore, in this paper we focus on the first aspect of how to efficiently assign the channelization codes. Additionally, most researches did not consider the analysis of tree state with dynamic traffic load and their analysis lack of systematic call admission control (CAC) mechanism. Therefore, in this paper, we first propose the Markov decision process (MDP) based analysis to assign channelization codes efficiently. Next, we extend the MDP-based approach as the call admission control mechanism to maximize the system revenue while reducing blocking probability. Furthermore, a bit string masking algorithm is proposed to reduce the time complexity of tree managing and searching for available channelization codes. Numerical results indicate that the proposed MDP approach yields the best fractional reward loss, code blocking reward loss, and code blocking ratio as compared to that of other schemes, including the random, left most, and crowded first schemes.

A novel location management in IP-based cellular networks

Mobile IP provides a simple and scalable global mobility solution that supports users to access the Internet at anytime and anywhere. However, as the number or the velocity of the mobile IP users grows, the signaling overhead associated with the mobility management also grows. One of the solutions, cellular IP, a micro-mobility protocol, provides a seamless mobility support for semi-soft handoff and paging in limited geographical areas. In this paper, we propose a new micro-mobility management, which is an extension of cellular IP. Two special features are added in our scheme: page area group and pointer forwarding. Analytical model is developed to evaluate the reduction of signaling overhead of the proposed scheme. Simulation results are also conducted to verify our analysis. Our numerical results show that the proposed scheme can reduce the signaling overhead effectively.

Fair and efficient scheduling for UMTS forward link

This paper presents two scheduling algorithms, MWF2Q+ and MDRR, for multiple classes of service over the same spectrum in the forward link of the UMTS network. These scheduling algorithms can allocate bandwidth in proportion to weights of flows sharing the channel, and assign OVSF code to backlogged flows on a frame-by-frame basis. The MWF2Q+ algorithm has better fairness properties while the MDRR algorithm requires less computational complexity and space complexity. The fairness properties of these scheduling algorithms are analysed in this paper. Our simulation results show that these two algorithms support multiple traffic sources with heterogeneous rate guarantees while fully utilizing the system bandwidth. The impact of self-similar traffic is also addressed in our simulations. Copyright

Routing in ATM networks with multiple classes of QoS

In this paper we study the path-constrained-cost-optimized routing problem, which tries to minimize path cost while satisfying the QoS requirements of a connection. In previous research, each link of the network is associated with a set of QoS metrics, which represents the QoS metrics that can be provided by this link. However, in a real network, ATM switches are able to provide a finite number of QoS classes, instead of just one QoS class. Therefore, in this paper, we studied the path-constrained-cost-optimized routing problem where each link of the network is associated with several sets of QoS metrics. Our solution to this problem consists of two tasks: QoS decomposition and QoS-constrained least cost routing. We propose a greedy algorithm to decompose the end-to-end QoS constraint to local QoS constraints such that the path cost can be minimized. For QoS-constrained least cost routing, we first study three approaches for defining link costs. We then propose two routing algorithms for finding QoS-constrained paths.