Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

Optimizing Operators for Temporal and Spatiotemporal Data

Alseghayer, Rakan (2023) Optimizing Operators for Temporal and Spatiotemporal Data. Doctoral Dissertation, University of Pittsburgh. (Unpublished)

Primary Text

Download (1MB) | Preview


Mobile devices and IoT technologies have become highly available, which led to the development of smart solutions and applications. A common characteristic of these solutions is operating on temporal and/or spatiotemporal data that is often in the form of data streams. Motivated by two important classes of applications, health monitoring and contact tracing, this dissertation optimizes the spatiotemporal operators in the core of these applications. The former requires efficient data streams and/or timeseries correlations for monitoring temporal events, and the latter optimizes temporal aggregation joins for spatiotemporal data (trajectories). The broader contributions of this dissertation are two novel frameworks that offer effective implementations of such applications, demonstrated experimentally with real and synthetic data.

In the context of health monitoring (e.g., server farms), we develop the Detection of Correlated Data Streams (DCS) framework. It is a real-time monitoring framework of large volumes of data streams that are produced at high velocity. Typically, pairs of most recent data streams need to be correlated within a specified delay target in order for their analysis to lead to actionable results. We address this need by: (i) segmenting data streams into micro-batches; and (ii) leveraging incremental sliding window computation, priority scheduling, and caching techniques, to avoid unnecessary re-computations and I/O. Furthermore, we devise and evaluate exploration strategies that effectively steer the processing of data stream correlations based on the monitoring objective.

In the context of contact tracing (CT), we propose the Privately Detecting Indoors Exposure Risk (PriDIER) distributed framework for detecting indoor contacts between two individuals and measuring the individual risk of infection for respiratory transmitted diseases (e.g., COVID-19). PriDIER carries out the CT queries locally on the individual users’ devices to protect their privacy and utilizes a data movement protocol to achieve scalability and reduce energy consumption at the users’ devices. In realizing PriDIER we develop e-Racoon, a novel in-memory structure to optimize temporal aggregation joins for trajectories. The e-Raccon structure enables efficient trajectory joins with the duration of contacts cumulatively between an individual and a single other individual, while considering the exposure across other users.


Social Networking:
Share |


Item Type: University of Pittsburgh ETD
Status: Unpublished
CreatorsEmailPitt UsernameORCID
Alseghayer, Rakanraa88@pitt.eduraa88
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Committee ChairChrysanthis, Panos
Committee MemberLabrinidis,
Committee MemberPruhs,
Committee MemberMohamed,
Committee MemberConstantinos,
Date: 1 September 2023
Date Type: Publication
Defense Date: 19 July 2023
Approval Date: 1 September 2023
Submission Date: 3 August 2023
Access Restriction: No restriction; Release the ETD for access worldwide immediately.
Number of Pages: 108
Institution: University of Pittsburgh
Schools and Programs: Dietrich School of Arts and Sciences > Computer Science
Degree: PhD - Doctor of Philosophy
Thesis Type: Doctoral Dissertation
Refereed: Yes
Uncontrolled Keywords: Data Streams, Data Exploration, Correlation, Search, Subsequence, Timeseries, Search, Spatiotemporal, In-memory Access Method, Trajectory, Trajectory Join, Aggregate Join, Contact Tracing, Moving Objects, Mobile Computing
Date Deposited: 01 Sep 2023 19:18
Last Modified: 01 Sep 2023 19:18


Monthly Views for the past 3 years

Plum Analytics

Actions (login required)

View Item View Item