Gwangwoo Park

An SIP-Based Location Management Framework in Opportunistic WiFi Networks

In this paper, we propose a novel session-initiation-protocol-based location management framework in opportunistic WiFi networks. In the proposed framework, a mobile terminal needs to update its attachment to the network only when it connects to a new WiFi access point to reduce its location update traffic. In addition, to find out the exact point of attachment, a timer-based session setup procedure is introduced, which can intelligently reduce the session setup latency by looking for a user in either WiFi or cellular systems. We develop analytical models for the location update cost and the session setup latency and investigate the optimization of the timer under the realistic WiFi hotspot residence time model.

Reducing Call Setup Latency in Mobile VoIP Systems

In mobile voice over IP (VoIP) systems, user mobility leads to the increased call setup latency to find out the current location of user, which has a negative impact on users call quality. In this letter, we propose a prefetching technique for user location to reduce the call setup latency. We develop an analytical model and investigate the impacts of mobility parameters and thresholds.

Analysis of SIP Transfer Delay in Multi-Rate Wireless Networks

In this letter, we analyze the SIP transfer delay over multi-rate wireless networks where different transmission modes based on adaptive modulation code are supported. We consider a Nakagami m fading channel at the physical layer and automatic repeat request at the data link layer. Extensive simulation results are given to validate the analytical results and to demonstrate the effects of message sizes and channel conditions.

Simulation study of bufferbloat problem on WiFi access point

In home networks, a variety of smart devices are connected to a WiFi access point (AP) and the oversized buffer of the AP causes high latency/delay jitter and reduced network throughput (known as the bufferbloat problem). To mitigate this problem, appropriate active queue management (AQM) algorithms should be implemented at the AP. In this paper, we conduct extensive simulation study on the performance of CoDel, which is one of the most effective AQMs. Simulation results demonstrate that CoDel with carefully configured parameters can provide comparable throughput and shorter round-trip time compared to the tail drop queue management algorithm.

Bloom-filter-aided redundancy elimination in opportunistic communications

Opportunistic communication is a promising solution to offload mobile data traffic by exchanging data among mobile devices without the aid of infrastructures. However, considerable redundant traffic can be transmitted in opportunistic communications. In this article, we introduce a protocol-independent redundancy elimination (RE) technique to reduce redundant data traffic at the packet level. Specifically, we propose a Bloom-filter-aided RE (BFRE) to further improve the performance of RE on opportunistic communication links. In BFRE, the receiver predicts the next consecutive data transmissions and delivers the compressed information to the sender by means of a Bloom filter. Then the sender can transmit small encoded packets, and the data traffic can be effectively reduced. Extensive simulation results demonstrate that BFRE can significantly reduce traffic and transmission time by 34 and 23 percent, respectively.

Mobility-aware DASH for cost-optimal mobile multimedia streaming services

For multimedia streaming in mobile environments, video bitrate should be carefully determined and different channel quality and pricing in heterogeneous wireless networks (e.g., cellular and WiFi networks) need to be considered. In this work, we propose a new bitrate adaptation (BA) technique called mobility-aware dynamic adaptive streaming over HTTP (MDASH), in which a cellular data usage is minimized through a Markov Decision Process (MDP) framework. Preliminary simulation results are shown that MDASH can reduce 23% of the cellular data usage without any playback interruptions while minimizing the temporal variations in the video quality.

Lightweight traffic redundancy elimination in software-defined wireless mesh networks

Wireless mesh network (WMN) is a promising technology for building cost-effective and easy-deployable wireless home networking infrastructure. To efficiently utilize limited radio resources in WMNs, packet transmissions (especially, redundant packet transmissions) should be carefully managed. To this end, we introduce a software-defined WMN architecture where the data plane and control plane are separated and packet transmissions over WMNs are controlled by a centralized entity. We also propose a lightweight traffic redundancy elimination (LTRE) scheme in software-defined WMNs to reduce redundant packet transmissions. Preliminary analysis results show that LTRE can significantly reduce the traffic overhead.

Contents-aware multicast in rate adaptive wireless networks

We propose a contents-aware multicast (CAM) scheme in rate adaptive wireless networks. In CAM, transmission rate is determined depending on the contents type. That is, base layer packets are transmitted by the lowest rate for higher reliability. On the contrary, enhancement layer packets are sent by higher transmission rates. Moreover, enhancement layer packets are combined with base layer packets and redundantly transmitted, which allow higher decoding probability in error-prone wireless networks.

A timer-based session setup procedure in cellular-WLAN integrated systems

The multiple registration scheme can reduce signaling overhead at the expense of increased session setup latency in cellular-WLAN integrated systems. In this paper, we propose a timer-based session setup procedure, which can reduce the session setup latency by looking for a user based on the estimation of the WLAN residence time. We investigate the optimization of the timer under realistic WLAN residence time model. Numerical results demonstrate that the timer-based session setup procedure can reduce the session setup latency by means of adaptive timer setting.

A Two-Stage Failover Mechanism for High Availability in Service Function Chaining

Service function chaining (SFC) enables specific network flows to be processed through an ordered set of service functions. In this paper, we propose a two-stage failover mechanism to provide high availability (HA) in SFC. In the proposed mechanism, a secondary service function path (SFP) is pre-computed, which is maximally disjointed with the primary SFP, and a backup service function instance (SFI) is prepared for each service function. For a failure event, the first local repair stage is examined by means of the preconfigured backup SFI. If the backup SFI is not available and/or any link is failed, the second end-to-end repair stage is triggered by replacing the primary SFP with the secondary SFP. Simulation results demonstrate that the proposed failover mechanism can reduce the recovery time by 50%-75% compared to the conventional approach.