FReAK: fault injection based resilience assessment of Kubernetes
Sreejith, Harshitha
This item's files can only be accessed by the System Administrators group.
Permalink
https://hdl.handle.net/2142/124728
Description
Title
FReAK: fault injection based resilience assessment of Kubernetes
Author(s)
Sreejith, Harshitha
Issue Date
2024-05-03
Director of Research (if dissertation) or Advisor (if thesis)
Iyer, Ravishankar K.
Department of Study
Electrical & Computer Eng
Discipline
Electrical & Computer Engr
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
M.S.
Degree Level
Thesis
Keyword(s)
Fault injection
resilience
Kubernetes
control plane
failure
container orchestration
Abstract
Some factors contributing to cluster outages and failures, resulting in revenue loss and diminished trust among cloud providers and users, are rooted within container orchestrators. This underscores the necessity for thorough resilience testing. Resilience assessment of Kubernetes, the most widely adopted container orchestrator, using nearly 12000 fault injections in the control plane of Kubernetes performed using FReAK, an automated fault injector, is presented in this thesis. Subsequently, the vulnerable functionalities of the components are uncovered. Additionally, a Bayesian model for fault localization and higher fault manifestation is designed. The results of the fault injections show that about 60% of the activated faults do not affect the control plane, but critical failures that make at least one control plane component unavailable were observed in 11% of the fault injections. Moreover, the API server turned out to be the most critical. The results also indicated insufficiency in input validation and insufficiency of heartbeats in indicating the availability of the components. Testing the Kubernetes control plane through such tests can aid in the development of more robust resilience strategies.
Use this login method if you
don't
have an
@illinois.edu
email address.
(Oops, I do have one)
IDEALS migrated to a new platform on June 23, 2022. If you created
your account prior to this date, you will have to reset your password
using the forgot-password link below.