Use this URL to cite or link to this record in EThOS:
Title: Node selection in distributed overlays
Author: Latif, L.
Awarding Body: University College London (University of London)
Current Institution: University College London (University of London)
Date of Award: 2013
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
With the proliferation of latency aware services such as live video streaming, Internetbased financial trading and the popularity of distributed overlays such as BitTorrent there is a growing need for latency-aware distributed overlays. To make such overlays viable, efficient resource discovery services are needed. Anycast is a routing protocol that sends packets to nodes that are a member of a particular group, with the work presented the Anycast protocol in the distributed overlay domain. Structured and unstructured distributed networks have become a popular way to disseminate data without the need for a fixed infrastructure, however there is a need to provide quality of service (QoS). To meet the demands of applications, an overlay needs to maintain accurate Anycast group membership data, locality information and have minimal protocol overhead. Three protocols are proposed to meet these goals. The Distributed Overlay Anycast Table (DOAT) brings the notion of locality to a structured overlay, while introducing Bloom filters as an efficient data structure to present an overlay that can accurately return a node that is participating in a particular group. The Gossip Overlay Anycast Table (GOAT) is a scalable location-aware unstructured overlay that can provide the probabilistic Anycast routing. Through the use of an efficient discovery protocol and the use of Bloom filters, GOAT is able to provide the advantages of a structured overlay, while mitigating the performance issues typically found in unstructured overlays. The N-casting overlay is an unstructured overlay with the ability to send queries to multiple members of an group, uses a hierarchical decomposition of the Internet and an elegant data structure that offers predictable compression of overlay membership. N-casting shows that unstructured overlays can be scalable and sustain high performance in environments that exhibit realistic membership churn. DOAT, GOAT and N-casting present viable services that implemented at the application layer provide location aware node discovery in QoS-enabled applications.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available