Oracle's IORM enforces I/O limits across heterogeneous storage with 5% accuracy

Disaggregated storage clusters serving latency-sensitive databases and opaque block workloads face two governance challenges: maintaining consistent performance across heterogeneous hardware generations and enforcing global I/O limits when access patterns skew to a subset of nodes. The I/O Resource Manager (IORM), deployed in Oracle Exadata Exascale, addresses these with three mechanisms. First, a hardware-aware cost modeler normalizes I/O accounting using datasheet-derived fixed costs, making limits invariant across hardware generations. Second, a quantum-based rate limiter with bounded carry-forward credits accommodates database micro-bursts while enforcing long-term service-level objectives (SLOs). Third, a distributed adaptive feedback controller redistributes unused entitlements across the cluster to resolve topological access skew.

In production tests on an 8-node cluster running up to 100 concurrent tenant volumes, IORM converges within 5% of provisioned limits under extreme sequential skew. It scales without inter-tenant interference and recovers full throughput within 15 seconds of a storage-node failure. Beyond design, the paper shares operational lessons from deployment, including how the system handles hardware heterogeneity transparently. Accepted at IEEE IC2E 2026, this work provides a practical framework for cloud engineers dealing with mixed-generation storage infrastructure.