Constructive Dissonance in the Cloud: Adaptive Out-of-Phase Scheduling for Periodic Tasks
(2023) International Conference on Cloud Networking- Abstract
- In the realm of cloud computing, the prevalence of intra-phase congestion due to concurrent periodic requests from a multitude of clients frequently results in heightened resource consumption, elongated queuing times, and prolonged response durations. To tackle these challenges, this paper presents a groundbreaking 'Adaptive Out-of-Phase Scheduling' technique, utilizing a server-side PID controller to induce constructive dissonance, strategically phase-shifting clients by delaying subsequent requests based on server feedback, leading to dispersed out-of-phase transitions. The hallmark of our method is its uncomplicatedness, obviating the necessity for intricate execution analysis or elaborate optimal scheduling decisions. Empirical... (More)
- In the realm of cloud computing, the prevalence of intra-phase congestion due to concurrent periodic requests from a multitude of clients frequently results in heightened resource consumption, elongated queuing times, and prolonged response durations. To tackle these challenges, this paper presents a groundbreaking 'Adaptive Out-of-Phase Scheduling' technique, utilizing a server-side PID controller to induce constructive dissonance, strategically phase-shifting clients by delaying subsequent requests based on server feedback, leading to dispersed out-of-phase transitions. The hallmark of our method is its uncomplicatedness, obviating the necessity for intricate execution analysis or elaborate optimal scheduling decisions. Empirical evaluation on an operational web server demonstrates the efficacy of this technique in mitigating resource utilization, reducing response time by half, and significantly decreasing response time variability. Our technique offers an efficient and practical solution to intra-phase congestion in periodic tasks, optimizing both system performance and resource efficacy in diverse cloud computing environments. Our method is scalable, versatile, and aptly adapted for real-world applications, with potential future research exploring fine-tuning and trade-offs in various system setups. By proficiently mitigating the ramifications of simultaneous periodic tasks in an efficient and streamlined fashion, our approach is primed to advance the evolution of more responsive and optimized cloud computing infrastructures, yielding advantages across a wide range of applications and services. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/75b3339a-d5e6-427b-9cb7-948d65d623a7
- author
- Tärneberg, William LU and Skarin, Per LU
- organization
- publishing date
- 2023-11-02
- type
- Chapter in Book/Report/Conference proceeding
- publication status
- in press
- subject
- keywords
- Cloud computing, Scheduling, Response time, Intra-period congestion, Self-adaptive
- host publication
- International Conference on Cloud Networking
- publisher
- IEEE - Institute of Electrical and Electronics Engineers Inc.
- conference name
- International Conference on Cloud Networking
- conference location
- New York, NY, United States
- conference dates
- 2023-11-01 - 2023-11-03
- project
- WASP: Wallenberg AI, Autonomous Systems and Software Program at Lund University
- ELLIIT LU P01: WP6 Vehicular communication
- Cyber Security for Next Generation Factory (SEC4FACTORY)
- Intelligent Management of next generation MobIle NEtworks aNd serviCEs (IMMINENCE)
- language
- English
- LU publication?
- yes
- id
- 75b3339a-d5e6-427b-9cb7-948d65d623a7
- date added to LUP
- 2023-11-07 14:27:19
- date last changed
- 2023-11-21 13:00:59
@inproceedings{75b3339a-d5e6-427b-9cb7-948d65d623a7, abstract = {{In the realm of cloud computing, the prevalence of intra-phase congestion due to concurrent periodic requests from a multitude of clients frequently results in heightened resource consumption, elongated queuing times, and prolonged response durations. To tackle these challenges, this paper presents a groundbreaking 'Adaptive Out-of-Phase Scheduling' technique, utilizing a server-side PID controller to induce constructive dissonance, strategically phase-shifting clients by delaying subsequent requests based on server feedback, leading to dispersed out-of-phase transitions. The hallmark of our method is its uncomplicatedness, obviating the necessity for intricate execution analysis or elaborate optimal scheduling decisions. Empirical evaluation on an operational web server demonstrates the efficacy of this technique in mitigating resource utilization, reducing response time by half, and significantly decreasing response time variability. Our technique offers an efficient and practical solution to intra-phase congestion in periodic tasks, optimizing both system performance and resource efficacy in diverse cloud computing environments. Our method is scalable, versatile, and aptly adapted for real-world applications, with potential future research exploring fine-tuning and trade-offs in various system setups. By proficiently mitigating the ramifications of simultaneous periodic tasks in an efficient and streamlined fashion, our approach is primed to advance the evolution of more responsive and optimized cloud computing infrastructures, yielding advantages across a wide range of applications and services.}}, author = {{Tärneberg, William and Skarin, Per}}, booktitle = {{International Conference on Cloud Networking}}, keywords = {{Cloud computing; Scheduling; Response time; Intra-period congestion; Self-adaptive}}, language = {{eng}}, month = {{11}}, publisher = {{IEEE - Institute of Electrical and Electronics Engineers Inc.}}, title = {{Constructive Dissonance in the Cloud: Adaptive Out-of-Phase Scheduling for Periodic Tasks}}, url = {{https://lup.lub.lu.se/search/files/164504688/Dissonance_in_the_Cloud_Adaptive_Out_of_Phase_Scheduling_for_Periodic_Tasks-5.pdf}}, year = {{2023}}, }