An Adaptive Proximity Route Selection Method in DHT-Based Peer-to-Peer Systems
Ji-Young Song, Sae-Young Han, Sung-Yong Park
Abstract
Open-access reader
Ji-Young Song, Sae-Young Han, Sung-Yong Park
Abstract
Open-access reader
다양한 네트워크로 구성된 인터넷 환경에서 오버레이 홉 수를 최소화하는 Chord를 비롯한 기존의 분산 해싱 테이블 기반의 피어-투-피어 시스템에서는 목적 노드까지의 실질적인 질의 라우팅 시간을 줄일 수 없다. 따라서 제안하는 적응적 근접경로 선택기법에서 각 노드는 라우팅테이블의 엔트리 노드 중 지연시간을 최소로 하는 노드를 선택하여 질의 메시지를 전달하도록 한다. 이를 위하여 각 엔트리의 목적지까지의 지연시간을 큐 라우팅 알고리듬과 지수적 최근-가중치 평균을 이용해 예측하여 검색 테이블로 저장하고, 노드들 간 메시지 전달 시 이 정보를 교환하여 추가 부하 없이 검색 테이블에 반영한다. 시뮬레이션을 통해 제안하는 기법이 전채 검색 시간과 흡 간 지연시간에 있어서 근접성을 고려하지 않은 Chord나 CFS의 서버 선택 알고리즘에 비해 월등히 좋은 성능을 보임을 확인할 수 있다. In the Internet of various networks, it is difficult to reduce real routing time by just minimizing their hop count. We propose an adaptive proximity route selection method in DHT-based peer-to-peer systems, in which nodes select the nぉe with smallest lookup latency among their routing table entries as a next routing node. Using Q-Routing algorithm and exponential recency-weighted average, each node estimates the total latency and establishes a lookup table. Moreover, without additional overhead, nodes exchange their lookup tables to update their routing tables. Several simulations measuring the lookup latencies and hop-to-hop latency show that our method outperforms the original Chord method as well as CFS' server selection method.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
다양한 네트워크로 구성된 인터넷 환경에서 오버레이 홉 수를 최소화하는 Chord를 비롯한 기존의 분산 해싱 테이블 기반의 피어-투-피어 시스템에서는 목적 노드까지의 실질적인 질의 라우팅 시간을 줄일 수 없다. 따라서 제안하는 적응적 근접경로 선택기법에서 각 노드는 라우팅테이블의 엔트리 노드 중 지연시간을 최소로 하는 노드를 선택하여 질의 메시지를 전달하도록 한다. 이를 위하여 각 엔트리의 목적지까지의 지연시간을 큐 라우팅 알고리듬과 지수적 최근-가중치 평균을 이용해 예측하여 검색 테이블로 저장하고, 노드들 간 메시지 전달 시 이 정보를 교환하여 추가 부하 없이 검색 테이블에 반영한다. 시뮬레이션을 통해 제안하는 기법이 전채 검색 시간과 흡 간 지연시간에 있어서 근접성을 고려하지 않은 Chord나 CFS의 서버 선택 알고리즘에 비해 월등히 좋은 성능을 보임을 확인할 수 있다. In the Internet of various networks, it is difficult to reduce real routing time by just minimizing their hop count. We propose an adaptive proximity route selection method in DHT-based peer-to-peer systems, in which nodes select the nぉe with smallest lookup latency among their routing table entries as a next routing node. Using Q-Routing algorithm and exponential recency-weighted average, each node estimates the total latency and establishes a lookup table. Moreover, without additional overhead, nodes exchange their lookup tables to update their routing tables. Several simulations measuring the lookup latencies and hop-to-hop latency show that our method outperforms the original Chord method as well as CFS' server selection method.
Key concepts: Chord (peer-to-peer), Latency (audio), Computer science, Routing table, Computer network, Peer-to-peer, Hop (telecommunications), Distributed hash table