.png){kind=icon}
.png){kind=icon}
Return
Hypergraph-based adaptive resource scheduling for fault-resilient multi-tenant cloud allocation
.png){kind=icon}
OA
AI
J
DOI：10.1186/s13677-026-00925-2.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Multi-tenant cloud computing is crucial for resource sharing and adaptive resource utilization; however, it faces challenges such as fault propagation, resource contention, and workload imbalance. In multi-tenant cloud environments, shared resource dependencies cause cascading failures that affect multiple tenants simultaneously. Existing research often isolates faults within individual tenants without mapping inter-tenant resource propagation, making it challenging to prevent large-scale failures causing Service Level Objectives (SLO) violations. To address these challenges, a novel Hyper-Swarm Stochastic Resilience and Equi-Priority Cloud Orchestration (HSRECO) is proposed. The approach integrates the Hypergraph-Driven Stochastic Causal Fault Resilience Strategy to predict and prevent cascading failures by mapping inter-tenant dependencies and identifying fault-prone areas. Moreover, cloud infrastructure faces resource contention when handling multifaceted workloads (batch, real-time), but existing models segregate workloads without dynamically allocating resources, causing resource contention, task prioritization failure, and Service-Level Agreements (SLA) violations. Thus, EquiOpti Swarm-based Cloud Allocation (EOS-CA) is introduced to optimize workload handling by ensuring equitable resource distribution. The results show a significant improvement in cloud performance, with 89.43% throughput, 97% resource utilization, 11·3s response time, minimal SLA violations, and enhanced system resilience.
Keywords:
Resource fault propagation
Cascading failures
Multi-tenant cloud resilience
Fault prediction
Journal
J
IF:
4.3
Papers:
724
Citations:
2.2K