Bufferbloat has drawn much attention in the network community for its negative impact on TCP delay performance and user QoE. Recently, it has been more commonly noted in wireless access networks, in part, due to over-provisioned buffer space. Previous works that focused only on bufferbloat prevention have suffered from either deployment or fairness problems when coexisting with conventional TCP flows. In this paper, we address the bufferbloat problem in resource-competitive environments such as Wi-Fi, and design a receiver-side countermeasure for easy deployment that does not require any modification at the sender or intermediate routers. Exploiting TCP and AQM dynamics, our scheme competes for shared resource in a fair manner with conventional TCP flow control methods and prevents bufferbloat. We implement our proposed scheme in commercial smart devices and verify its performance through real experiments in LTE and Wi-Fi networks.

Abstract—Internet forums provide a varied and graceful stage for those who have hobbies or interests in common and want to share their contents of interest with each other. Traditionally they use Client-Server (CS) model. A distinct characteristic of a Internet forum is that some registered nodes1 periodically visit the forum. It implies that those nodes are regular visitors and can be considered relatively stable for some duration of time. The efficient use of these relatively stable nodes is very useful in a P2P network design, especially when the forum server is with limited bandwidth. In this paper, we propose a hybrid P2P network structure which is targeted for Internet forums. In the proposed system, the forum server maintains a list of registered nodes and classifies some nodes that are regularly visiting the forum with high prob-ability as stable nodes. The P2P network created by these stable nodes is triggered when the forum server is highly demanded due to a special event. To resolve the bandwidth shortage problem, the forum server selects stable nodes in advance, according to each node’s activity history. For the stable P2P node selection, we consider two types of mechanisms; a time period mechanism and a valid date window (VDW) mechanism, according to the considered time period. Then, we confirm that our proposed P2P model outperforms the traditional CS model through simulations. I.

With the multi-fold development in communication technology, there has been a rapid growth in consumption of multimedia services using mobile devices. In particular, lots of research has been done to provide multimedia services using a 3G/WLAN dual mode to mobile device users. All users download contents from content providers through cellular networks and can share their contents with their neighbors. This may lead to a situation where the cost of receiving multimedia contents is reduced owing to contents sharing. In other words, users effectively reduce the cost of receiving contents as they share some contents with neighbors. However, we believe that selfish users (or users with limited resources) are not willing to cooperate or share their contents with other users. In this paper, we propose an incentive scheme that gives the benefit of contents sharing to users in a 3G/WLAN dual mode supporting network. There are two types of users; premium users and ordinary users. Premium users can get high quality contents at the expense of spending their own money and resources of power and bandwidth, while ordinary users receive plain quality contents freely or with paying little amount of money. Our incentive mechanism designs the content provider to offer each premium user a discounted price for downloading high quality contents. Through our approach, each premium user will receive some incentive of a discounted price for receiving high quality contents in proportion to the contribution to the content provider. By doing so, the content provider can increase its total utility too. Our simulation results confirm that our proposed incentive scheme performs well.

In this paper, we propose a novel P2P VoD streaming system, named PECAN where each peer adjusts its cache capacity adaptively to meet the server's upload bandwidth constraint and achieve the fairness. For doing so, we first propose a new cache replacement algorithm that designs the number of caches for a segment to be proportional to its popularity. Second, we mathematically prove that if the cache capacity of a peer is proportional to its segment request rate, the fairness between upload and download amounts at each peer can be achieved. Third, we propose a method that determines each peer's cache capacity adaptively according to the constraint of the server's upload bandwidth. Through simulations, we show that PECAN meets the server upload bandwidth constraint, and achieves the fairness well at each peer.