Research collaboration showcases advances in Microservice Resilience at IEEE ASYU 2026
Researchers from Norsk Regnesentral and Özyeğin University presented the LOCO study, exploring how Kubernetes-based microservices perform under high workloads and unexpected disruptions, contributing to ENFIELD’s ongoing collaboration within the Adaptive AI pillar.
The ENFIELD project was represented at the IEEE Innovations in Intelligent Systems and Applications Conference (ASYU 2026), held from 9 to 11 September 2026 at Özyeğin University in Istanbul, Türkiye. During the conference, Ismail Ari presented the research paper “LOCO: Testing Resilience of Microservices with Kubernetes under High Load and Chaos”, co-authored by Osman Serhat Yılmaz, Ata Baranta, Ismail Ari, Sandeep Pirbhulal, and Habtamu Abie.
The paper presents research into the resilience, scalability, and reliability of microservice-based applications, an increasingly important area as modern digital systems move away from traditional monolithic architectures towards distributed, cloud-native solutions.
As Kubernetes and container-based technologies become essential for managing complex applications, understanding how these systems respond to heavy workloads, unexpected failures, and network disruptions is critical to ensuring their reliability and operational performance.
To address these challenges, the LOCO study evaluated two widely used microservice benchmarks, Boutique and Sock Shop, under different stress conditions. The research involved nearly 300 experiments conducted over approximately 150 hours, examining performance across different Kubernetes distributions, networking configurations, and automated scaling strategies.
Using controlled chaos engineering experiments, the researchers simulated real-world disruptions, including service failures, network delays, and network partitions, to assess how different configurations respond to challenging operational conditions.
The findings highlight that there is no one-size-fits-all solution for achieving microservice resilience. In particular, Sock Shop proved challenging to scale due to its stateful architecture, while the combination of K3s and Calico networking provided a relatively stable baseline until higher workloads were introduced. These results reinforce the importance of carefully selecting and managing infrastructure configurations according to specific performance, scalability, availability, and reliability requirements.
International research collaboration through ENFIELD
This scientific contribution is also an important outcome of the ongoing collaboration between Norsk Regnesentral (Norwegian Computing Center) and Özyeğin University’s Department of Computer Engineering, established through ENFIELD’s Third-party Exchange Scheme (TES).
As a selected researcher under ENFIELD’s TES Open Call, Ismail Ari has been collaborating with Norsk Regnesentral within the project’s Adaptive AI pillar, contributing to joint research activities and scientific knowledge exchange.
Although the LOCO study focuses specifically on cloud-native infrastructure rather than AI algorithms, its findings are relevant to the development of reliable and adaptable computing environments capable of supporting increasingly complex distributed and AI-enabled applications.
The presentation at IEEE ASYU 2026 reflects the continuity of this collaboration and demonstrates how ENFIELD’s research network facilitates scientific cooperation beyond individual exchange activities.
Through joint publications, sustained researcher engagement, and the exchange of technical expertise, ENFIELD continues to strengthen international research partnerships and support advances in resilient, adaptive, and trustworthy digital technologies.
This achievement reinforces ENFIELD’s broader mission of fostering scientific excellence, encouraging collaboration between European and international research communities, and generating lasting impact through its research and innovation activities.
Scientific publication
Research institutions: Norsk Regnesentral (Norwegian Computing Center) and Özyeğin University, Department of Computer Engineering.
ENFIELD research pillar: Adaptive AI.
Paper information
Title: LOCO: Testing Resilience of Microservices with Kubernetes under High Load and Chaos.
Authors: Yılmaz, O. S., Baranta, A., Ari, I., Pirbhulal, S., & Abie, H. (2026).
Read more: Chaos Engineering LOCO IEEE ASYU 2026
