Sung-Kwan Youm

An adaptive resource allocation mechanism including fast and reliable handoff in IP-based 3G wireless networks

The aim of the all-IP architecture in third-generation wireless networks is to allow operators to deploy IP technology in delivering third-generation services, that is, an architecture based on packet technologies and IP telephony for simultaneous real-time and non-real-time services. We present a new handoff management scheme that utilizes simultaneous multiple bindings in IP-based 3G wireless networks. The proposed handoff scheme minimizes packet losses and incurs negligible delays during handoff. We also present an adaptive resource allocation scheme that allocates connection resources to incoming calls depending on the network conditions. The resources accumulated by bandwidth degradation for on-going calls are utilized to minimize dropping probability of handoff calls and blocking probability of new calls.

Multi-level service differentiation scheme for the IEEE 802.15.4 sensor networks

In the sensor networks, the data packets transmitted by the different devices in the home networking and the industrial application maintain the different levels of importance. In this paper, we propose two mechanisms for the IEEE 802.15.4 sensor networks to provide the multi-level differentiated services which are required by each and every device. The mathematical model based on the discrete-time Markov chain is presented and is analyzed to measure the performances of the proposed mechanisms. The numerical results show the effect of the variation of the contention window size and the backoff exponent for the service differentiation on the 802.15.4 sensor networks. From the results, we derive that the contention window size is more affective than the backoff exponent on the service differentiation while the backoff exponent is more affective than the contention window size on the average delay of every device. Simulation results are also given to verify the accuracy of the numerical model.

Adaptive resource allocation mechanism during handoff for mobile Internet telephony

We present an overview of recent research efforts in developing adaptive delivery models for Internet multimedia applications, which dynamically adjust the transmission rate according to network conditions. The proposed mechanism utilizes service priority in multimedia applications. Simulation results show that the average rate of quality degradation increases as utilization increases. However, the degradation rate should be weighted against the increased handoff blocking probability, as priority is assigned to each channel within the permitted range that meets the traffic needs of each channel. The handoff dropping probability is reduced, in situations where resources were highly utilized, by the proposed mechanism.

Traffic Engineering using a Heuristic Multi-Path Routing Strategy in MPLS Network

Offline routing algorithms are used for network administrator to route flows on optimized paths obtained through the traffic matrix. Online routing algorithms, however, find paths for dynamic LSP setup requests under multiple constraints such as bandwidth, delay and hop count. This paper presents new hybrid optimization algorithm (HOA) for dynamic routing of bandwidth guaranteed flows. The newly proposed algorithm is to decide an appropriate one among route set which is predetermined through offline routing algorithm. The simulation results show that the developed algorithm far outweighs previously developed algorithms on metrics like total bandwidth accepted and rejection ratio

Input register architectures of high speed router for supporting the PHB of differentiated services

In this paper, we presents a new router architecture for supporting the Differentiated Services(DiffServ) in the input-registered router, which has multiple Virtual Output Registers(VORs) for buffering packets ahead of the non-blocking switching fabric. A queuing discipline function block needed on the DiffServ router is added in the input-registered router in order to support Per-Hop Behaviors(PHBs) of DiffServ. Also a new matching algorithm, First Scheduled First Matching (FSFM), is considered to match input packets and output ports in the proposed VOR architecture. The simulation results by using ARENA show that the proposed architecture offer packet loss rate and delay close to the results of output-queued router with N time speed-up switch fabric in various PHBs on DiffServ.

Performance analysis of reliable multicast protocol using transparent proxy servers over hybrid networks

We consider the reliable multicast transport protocols used in hybrid networks that include wired and wireless networks and transparent proxy servers in it, which cache multicast packets at the base station and perform local retransmissions across the wireless link. We present two extreme reliable multicast transport protocols, sender-initiated and receiver-initiated, and two analytic performance models supported and unsupported by transparent proxy servers in each reliable multicast protocol, and analyze the throughput of each model. Numerical results show that in any case the protocol with transparent proxy servers effect overall performance more than the one without transparent proxy servers, and furthermore the receiver initiated reliable multicast supported by transparent proxy servers gives better performance of total throughput than the one sender initiated reliable multicast. We provide efficiency criterion of the number of transparent proxy servers for each protocols for varying wireless loss probabilities.

Mitigating the Hidden Terminal Problem and Improving Spatial Reuse in Wireless Ad Hoc Network

The 802.11 DCF was designed for the single-hop wireless LAN scenario, where one central control system such as AP controls many nodes and mitigated the hidden terminal problem by exchanging the RTS/CTS message. However, in Mobile Ad Hoc Network (MANET), RTS/CTS exchange mechanism is not sufficient to solve hidden terminal problem and also this mechanism reserved a space more than needs so that spatial reuse is reduced. In this Paper, we proposed new Distributed MAC protocol to mitigate the hidden terminal problem and to increase the spatial reuse. we classify the neighbor nodes into two states – free state and constraint state – based on the interference tolerance of a receiving node.

Multiple hierarchical protection schemes for differentiated services in GMPLS networks

To deliver reliable differentiated services in optical networks, GMFLS requires a set of procedures to provide protection for the traffic carried on the light paths. Here we propose multiple hierarchical protection schemes that improve the utilization of network resources-optical label switched paths (LSPs) such as lambda-LSP (L-LSP), waveband-LSP (WB-LSP), fiber-LSP (F-LSP). In the priority based LSPs, each LSP has its own recovery mechanism that F-LSP has 1+1 protection, WB-LSP has 1:1 protection, and L-LSP has 1:n protection. And the differentiated services including EF (explicit forwarding), AF (assured forwarding) and DF(default forwarding) can be allocated into hierarchical LSP respectively. We also propose the structure of the OXC (optical crossconnect) capable of setting up hierarchical LSPs or releasing LSPs by GMPLS control plane. The performance and effectiveness of our proposed multiple hierarchical protection schemes are demonstrated by the results of simulations on the mesh network.

Reliable fault management using hierarchical path protection mechanism for differentiated services in GMPLS networks

To deliver reliable differentiated services in optical networks, GMPLS requires a set of procedures to provide protection for the traffic carried on the light paths. In this paper we propose a hierarchical path protection mechanism that is simple, scalable, fast, and efficient. We describe in detail our design considerations, the communication of fault information to appropriate forwarding or/and switching elements, and the fault detection protocol. In particular, we propose a new notification protocol for efficient and fast distribution of fault notification messages. A new Hierarchical Path Protection (HPP) improves the utilization of network resources- Label Switched Paths (LSPs) such as Electrical LSP (E-LSP), Lambda-LSP (L-LSP), WaveBand-LSP (WB-LSP), Fiber-LSP (F-LSP), bandwidth etc.- that are necessary when recovery mechanism occurred in the GMPLS networks. The HPP is used in the multiple-layer networks made by priority number of paths. We assume that there are three kinds of priorities on LSPs by their traffic characteristics of multimedia services, real time, mission-critical and best-effort services in the GMPLS networks. When the faults of LSPs are detected, primary LSPs classified by the priority of multimedia traffic choose recovery mechanism according to their priorities. In the priority based LSPs, each LSP has its own recovery mechanism that the LSP of priority 1 (real-time traffic) has the dedicated protection scheme, the LSP of priority 2 (mission-critical traffic) has the preplanned restoration scheme and the LSP of priority 3 (best-effort traffic) has the dynamic restoration scheme. After recovery mechanism is accomplished, the information for setting up on LSP based on the given priority is transferred to the ingress Optical Cross Connect (OXC) for the purpose of making new LSP. The OXC has the structure capable of setting up new LSPs or releasing LSPs via a signaling to maintain the consistent recovery, regardless of fault count, based on the priority scheme. The HPP has some advantages expa