A Dependency-Aware Task Stealing Framework for Mobile Crowd Computing

Sanjay Segu Nagesh, Niroshinie Fernando (), Seng W. Loke, Azadeh Ghari Neiat and Pubudu N. Pathirana
Additional contact information
Sanjay Segu Nagesh: School of Information Technology, Deakin University, Geelong 3216, Australia
Niroshinie Fernando: School of Information Technology, Deakin University, Geelong 3216, Australia
Seng W. Loke: School of Information Technology, Deakin University, Geelong 3216, Australia
Azadeh Ghari Neiat: School of Electrical Engineering and Computer Science, The University of Queensland, Brisbane 4072, Australia
Pubudu N. Pathirana: School of Engineering, Deakin University, Geelong 3216, Australia

Future Internet, 2025, vol. 17, issue 10, 1-33

Abstract: Mobile crowd computing (MCdC) leverages the collective computational resources of nearby mobile devices to execute complex tasks without relying on remote cloud infrastructure. However, existing MCdC systems struggle with device heterogeneity and complex application dependencies, often leading to inefficient resource utilization and poor scalability. This paper presents Honeybee-Tx, a novel dependency-aware work stealing framework designed for heterogeneous mobile device clusters. The framework introduces three key contributions: (1) capability-aware job selection that matches computational tasks to device capabilities through lightweight profiling and dynamic scoring, (2) static dependency-aware work stealing that respects predefined task dependencies while maintaining decentralized execution, and (3) staged result transfers that minimize communication overhead by selectively transmitting intermediate results. We evaluate Honeybee-Tx using two applications: Human Activity Recognition (HAR) for sensor analytics and multi-camera video processing for compute-intensive workflows. The experimental results on five heterogeneous Android devices (OnePlus 5T, Pixel 6 Pro, and Pixel 7) demonstrate performance improvements over monolithic execution. For HAR workloads, Honeybee-Tx achieves up to 4.72× speed-up while reducing per-device energy consumption by 63% (from 1.5% to 0.56% battery usage). For video processing tasks, the framework delivers 2.06× speed-up compared to monolithic execution, with 51.4% energy reduction and 71.6% memory savings, while generating 42% less network traffic than non-dependency-aware approaches. These results demonstrate that Honeybee-Tx successfully addresses key challenges in heterogeneous MCdC environments, enabling efficient execution of dependency-aware applications across diverse mobile device capabilities. The framework provides a practical foundation for collaborative mobile computing applications in scenarios where cloud connectivity is limited or unavailable.

Keywords: pervasive computing; mobile edge computing; distributed edge computing; mobile crowd computing; dependency-aware task stealing; reliability assessment; capability-aware task stealing; middleware framework (search for similar items in EconPapers)
JEL-codes: O3 (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
https://www.mdpi.com/1999-5903/17/10/446/pdf (application/pdf)
https://www.mdpi.com/1999-5903/17/10/446/ (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:gam:jftint:v:17:y:2025:i:10:p:446-:d:1761063

Access Statistics for this article

Future Internet is currently edited by Ms. Grace You

More articles in Future Internet from MDPI
Bibliographic data for series maintained by MDPI Indexing Manager ().