Dynamic service prioritization with predicted intervals for QoS-sensitive service migrations in MEC

Service migrations in Multi-access Edge Compute (MEC) infrastructure are complex due to the geographically distributed infrastructure, limited availability of resources at the edge and heterogenous applications hosted in the infrastructure. Several existing studies have investigated service migratio...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Service oriented computing and applications Ročník 19; číslo 3; s. 237 - 249
Hlavní autori: Velrajan, Saravanan, Sharmila, V. Ceronmani
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: London Springer London 01.09.2025
Springer Nature B.V
Predmet:
Algorithms
Augmented reality
Autonomous vehicles
Bandwidths
Closed loops
Combinatorial analysis
Computer Appl. in Administrative Data Processing
Computer Science
Computer Systems Organization and Communication Networks
e-Commerce/e-business
Energy consumption
Geographical distribution
Infrastructure
Internet of Things
IT in Business
Machine learning
Management of Computing and Information Systems
Optimization
Original Research Paper
Particle swarm optimization
Real time
Software Engineering/Programming and Operating Systems
Streaming media
User behavior
User experience
Video transmission
Violations
United States > US
Service migrations
Particle Swarm optimisation
Service assurance
Quality of service
Edge computing
ISSN:1863-2386, 1863-2394
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:Service migrations in Multi-access Edge Compute (MEC) infrastructure are complex due to the geographically distributed infrastructure, limited availability of resources at the edge and heterogenous applications hosted in the infrastructure. Several existing studies have investigated service migrations in MEC to handle user mobility and energy savings. However, the intricate challenge of migrating thousands of user sessions in real-time during application overload conditions without impacting the users’ QoS has received limited attention. Ad-hoc service migrations in MEC pose a significant risk of service disruptions, affecting users' SLAs and degrading service performance for latency-sensitive applications. We propose a novel Time-aware Closed-loop Adaptive Particle Swarm Optimization (T-CLAPSO) algorithm that prioritises services to migrate in real-time based on application load and predicted service migration interval, minimising the impact on users’ QoS. Unlike existing studies that rely solely on application load-based heuristics, T-CLAPSO uniquely models the application performance and service migration time constraints as a combinatorial optimisation problem to prioritise service migrations while maintaining users’ QoS. T-CLAPSO algorithm reduces SLA violations during service migrations by 22% compared to the state-of-the-art Exponentially Weighted Moving Average with three-parameter smoothing (EWMA3) and 38% compared to the standard PSO algorithms, improving the user experience for QoS-sensitive applications such as video streaming, IoT Smart meter and Surveillance.
Bibliografia:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1863-2386
1863-2394
DOI:10.1007/s11761-024-00405-y