Search

Reputation aggregation methods are used in unstructured peer-to-peer (P2P) networks to evaluate the trustworthiness of participating peers and to combat malicious peer's behaviors. In reputation aggregation methods, each peer collects local scores by each transaction and calculates global scores by aggregating local scores. Using the global scores, individual peer can interact with reliable peers. However, existing reputation aggregation methods do not consider the reputation scores of the newly joining peers. In such conditions, file requests concentrate on the existing high reliability peers which leads to download failures. To avoid this problem, NP-Trust distributes the file request to all safe peers including the new peers by providing the newly joining peers with initial files and giving chances to be evaluated. Through computer simulations, the proposed method is compared with Gossip Trust and DFR-Trust in the following criteria: file request variation and transaction success rate. Compared to the DFR-Trust, the proposed method decreased the concentration of the file request by 60% and increased the transaction success rate by 4% and showed that the proposed scheme is efficient in the above two criteria. [ABSTRACT FROM PUBLISHER]

Some of P2P live streaming service encourage sending chunks of peers by changing quality of the video. In this case, peers connect each other and call each other as neighbors. However, the existing neighbor selection method has an issue. If peers that have high capacity for sending chunks has only low capacity peers as neighbor, the peer receives worse video quality than the peer might receive. We call this issue as peer's isolation. In this paper, We propose NSBTB, the neighbor selection method considering the capacity of the peer in order to eliminate isolation to reduce the difference between quality of high capacity peers. In NSBTB, each peer will notice own capacity to neighbors. And each peer selects a neighbor that has more capacity based on neighbor's capacity information. The simulation results show that this method is useful when there are less total amount of capacity of participating peers than all amount of the full video rate on the service. [ABSTRACT FROM PUBLISHER]

Many Internet users use P2P live streaming services which deliver live video contents to many users. In the services, the video content is sliced into small pieces called chunks. The peers not only receive but also send the chunks to watch the video contents. Source's load for delivering is reduced. In an Offer Select method, each peer offers sendable chunks to it's neighbors. The neighbors select receiving chunks from the sendable chunks. However, some chunk diffusions are low. Each peer sends chunk independently. It happens that peers offer same chunks to a neighbor. In this paper, we propose a chunk schedule method. Each peer receives chunks which the nearby peers hardly possess. Our method reduces dispersion of chunk diffusion. Simulation results show that peers receive more chunks, which enables peers to watch higher quality. [ABSTRACT FROM PUBLISHER]

In P2P live streaming services, a video streaming data is divided into chunks and peers collect the chunks to watch the live video. Peers receive chunks from a video server or other peers, and also send chunks to other peers. Generally, a peer cannot communicate with all of other peers and a peer can communicate with only a part of peers, which is called neighbors. How to select the neighbors has a big impact on video qualities of peers, because peers have various sending capacities. In an existing neighbor selection method, a peer selects neighbors which sent many chunks to itself. However, when neighbors of a peer with high sending capacity are only peers with low sending capacity, the peer is not able to receive enough amount of chunks and gets worse video quality than an average quality of peers with high sending capacities. In this paper, we propose a neighbor selection method based on sending capacity to decrease a variance of video qualities and to increase a chunk receiving rate for peers with high capacities. Simulation results show that the proposed method decreases a variance of video qualities for peers whose sending capacities are equivalent. [ABSTRACT FROM PUBLISHER]