Please use this identifier to cite or link to this item: https://hdl.handle.net/20.500.11851/5507
Title: A holistic abstraction to ensure trusted scaling and memory speed trusted analytics
Authors: Agca M. A.
Keywords: Directed acyclic graph
Distributed ledger
Distributed system
Event abstraction
Markovian chain monte carlo
Memory centric file system
Permissioned block chaining
Permissionless block chaining
Stream processing
Trusted computing
Trusted computing base
Publisher: Institute of Electrical and Electronics Engineers Inc.
Source: 6th Annual International Conference on Computational Science and Computational Intelligence, CSCI 2019, 5 December 2019 through 7 December 2019, , 159441
Abstract: In this study, a trusted holistic abstraction is proposed and analytically discussed using universal scalability law and Markovian chain Monte-Carlo method. Moreover, a feedback mechanism is modeled to explain the elasticity performance of the proposed distributed system. The system extends the data locality to the edges in a trusted manner and ensures trust while scaling the whole system and increasing the number of nodes. By the help of such a trusted solution, lineage information of the data at the edges enable fault-recovery from an available checkpoint, while maximizing the trustworthiness of the overall system. Innovative distributed data structures, make databases fresh for all scaled nodes by unifying the memory resources; minimize the need to trusted third parties via trusted distributed data structures, which uses checksums of the datum periodically. Hence, multi-layer neural networks and hierarchical tree structures, has confidential data, can be updated and trained dynamically. Searching speed and performance of an object or set of objects in massive systems is maximized while keeping the trustworthiness of the total system. Initial results indicate that the trust cost worth to pay to scale and to keep the performance of the whole system. The System also shows good elasticity in the case of sudden provisioning/de-provisioning of control nodes. The proposed system also has satisfactory resource-allocation capability with efficient clustering thanks to the introduction of distributed ledger-based transaction management and lineage data recording for dynamic management of DAG structures, has sub-modular and disjoint cluster sets. Initial results of micro-blog analytics indicate promising performance of unified batch/interactive/ad-hoc querying with the holistic abstraction. © 2019 IEEE.
URI: https://doi.org/10.1109/CSCI49370.2019.00267
https://hdl.handle.net/20.500.11851/5507
ISBN: 9781728155845
Appears in Collections:Bilgisayar Mühendisliği Bölümü / Department of Computer Engineering
Scopus İndeksli Yayınlar Koleksiyonu / Scopus Indexed Publications Collection

Show full item record



CORE Recommender

WEB OF SCIENCETM
Citations

4
checked on Sep 24, 2022

Page view(s)

52
checked on Mar 25, 2024

Google ScholarTM

Check




Altmetric


Items in GCRIS Repository are protected by copyright, with all rights reserved, unless otherwise indicated.