FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3099681974> ?p ?o ?g. }

• W3099681974 abstract "This dissertation addresses the problem of dynamic graph partitioning in a streaming manner in the cloud. This problem applies to real-world graph applications such as PageRank, Social Networks, Shortest Path and so on. The scale of graphs of these applications has increased to such a degree that a single machine is not capable of efficiently processing large graphs. Thereby, efficient graph partitioning and wise resource allocation are necessary for these large graph applications.At the beginning of this study, this dissertationevaluates two existing streaming graph partitioning algorithms in the cloud. After having completed the empirical study of these algorithms in the cloud machines, we identified the following research problems: 1) There are no existing streaming graph partitioning methods to find an optimised number of machines and scale the resources, as per the demands of an ever-increasing graph dataset. 2) How can we minimise the number of edge-cuts while balancing the load in a streaming manner in the cloud environment? 3) How can we use dynamic graph partitioning in a streaming manner to reduce the edge-cuts and the load imbalance during the partitioning? We also address the scaling of the resources as per the demands of dynamic graph data. Streaming graph partitioning is a variant of traditional graph partitioning which accepts graph input in a one-pass manner. This partitioning technique was introduced to overcome a memory bottleneck issue in traditional graph partitioning. In streaming graph partitioning, it is necessary to utilise the resources as per the demands of graph data stream. This thesis proposes an auto-scaling algorithm to determine the required number of machines, based on the upcoming stream data rate and the service time at the worker machines. The proposed method helps to minimise the cost and provide the best use of cloud resources by allocating the number and types of machines wisely. Once the optimised resources and costs are fixed, this study looks into the problem of graph partitioning in a streaming manner, with the aim of minimising inter-machine communication and reducing the computational load imbalance as much as possible. In order to achieve these goals, we propose a window-based, streaming graph partitioning algorithm. The proposed method utilises sliding window technology with a partitioning strategy and the load balancing method as well.After exploring the streaming graph partitioning with the static datasets, we studied the problem of the dynamic behaviour of graph datasets. This problem was how to partition dynamic graph data while minimising the edge-cut and keeping the computational load imbalance to a minimum. In addition to this partitioning technique, we also proposed an auto-scaling algorithm which adaptively scales in and out the machines, as per the demands of the computational cost." @default.
• W3099681974 created "2020-11-23" @default.
• W3099681974 creator A5023568746 @default.
• W3099681974 date "2020-01-01" @default.
• W3099681974 modified "2023-09-27" @default.
• W3099681974 title "Dynamic graph partitioning in streaming manner" @default.
• W3099681974 hasPublicationYear "2020" @default.
• W3099681974 type Work @default.
• W3099681974 sameAs 3099681974 @default.
• W3099681974 citedByCount "0" @default.
• W3099681974 crossrefType "dissertation" @default.
• W3099681974 hasAuthorship W3099681974A5023568746 @default.
• W3099681974 hasConcept C11413529 @default.
• W3099681974 hasConcept C120314980 @default.
• W3099681974 hasConcept C132525143 @default.
• W3099681974 hasConcept C13670688 @default.
• W3099681974 hasConcept C149635348 @default.
• W3099681974 hasConcept C173608175 @default.
• W3099681974 hasConcept C2780513914 @default.
• W3099681974 hasConcept C41008148 @default.
• W3099681974 hasConcept C48903430 @default.
• W3099681974 hasConcept C80444323 @default.
• W3099681974 hasConceptScore W3099681974C11413529 @default.
• W3099681974 hasConceptScore W3099681974C120314980 @default.
• W3099681974 hasConceptScore W3099681974C132525143 @default.
• W3099681974 hasConceptScore W3099681974C13670688 @default.
• W3099681974 hasConceptScore W3099681974C149635348 @default.
• W3099681974 hasConceptScore W3099681974C173608175 @default.
• W3099681974 hasConceptScore W3099681974C2780513914 @default.
• W3099681974 hasConceptScore W3099681974C41008148 @default.
• W3099681974 hasConceptScore W3099681974C48903430 @default.
• W3099681974 hasConceptScore W3099681974C80444323 @default.
• W3099681974 hasLocation W30996819741 @default.
• W3099681974 hasOpenAccess W3099681974 @default.
• W3099681974 hasPrimaryLocation W30996819741 @default.
• W3099681974 hasRelatedWork W1899807892 @default.
• W3099681974 hasRelatedWork W2005227805 @default.
• W3099681974 hasRelatedWork W2040670613 @default.
• W3099681974 hasRelatedWork W2112044792 @default.
• W3099681974 hasRelatedWork W2150437713 @default.
• W3099681974 hasRelatedWork W2164274587 @default.
• W3099681974 hasRelatedWork W2374139601 @default.
• W3099681974 hasRelatedWork W2518701676 @default.
• W3099681974 hasRelatedWork W2557812552 @default.
• W3099681974 hasRelatedWork W2574663572 @default.
• W3099681974 hasRelatedWork W2728568407 @default.
• W3099681974 hasRelatedWork W2786498279 @default.
• W3099681974 hasRelatedWork W2795817788 @default.
• W3099681974 hasRelatedWork W2810381754 @default.
• W3099681974 hasRelatedWork W2960721979 @default.
• W3099681974 hasRelatedWork W2967314080 @default.
• W3099681974 hasRelatedWork W2991446828 @default.
• W3099681974 hasRelatedWork W3047020862 @default.
• W3099681974 hasRelatedWork W3099389600 @default.
• W3099681974 hasRelatedWork W3173367869 @default.
• W3099681974 isParatext "false" @default.
• W3099681974 isRetracted "false" @default.
• W3099681974 magId "3099681974" @default.
• W3099681974 workType "dissertation" @default.