Benchmark Report | Cloud World Model AI

Cloud Architecture Simulation Benchmark

AWS vs GCP vs Azure vs OCI: Multi-Zone Burst Traffic Benchmark

10-Minute Simulation · 2× Burst Load · Four Major Cloud Providers

Powered by Cloud World Model — simulate cloud architecture without the cloud bill

AWS — us-east-1 — 2× m5.large + RDS r5.large + ALB

GCP — us-central1 — 2× n2-standard-4 + Cloud SQL + CLB

Azure — East US — 2× D4s_v3 + SQL S2 + Standard LB

OCI — us-ashburn-1 — 2× E4.Flex + MySQL HeatWave + LB

Scenario Configuration

Parameter<br>AWS (us-east-1)<br>GCP (us-central1)<br>Azure (East US)<br>OCI (us-ashburn-1)

Compute<br>2&times; EC2 m5.large (2 vCPU / 8 GB each)<br>2&times; n2-standard-4 (4 vCPU / 16 GB each)<br>2&times; Standard_D4s_v3 (4 vCPU / 16 GiB each)<br>2&times; VM.Standard.E4.Flex (1 OCPU / 8 GB each)

Database<br>RDS db.r5.large (2 vCPU / 16 GB, MySQL)<br>Cloud SQL db-n1-standard-2 (2 vCPU / 7.5 GB)<br>SQL Database Standard S2 (50 DTUs)<br>MySQL HeatWave DB System (2 OCPU / 16 GB)

Load Balancer<br>Application Load Balancer (ALB)<br>Cloud Load Balancing (CLB)<br>Standard Load Balancer<br>OCI Load Balancer

Traffic Pattern<br>Ramp from 25 k RPS baseline &rarr; 50 k RPS (2&times; burst) over 6 steps; sustained for 4 steps

Pricing Source<br>$0.45/hr baseline (official June 2026)<br>$0.60/hr baseline (official June 2026)<br>$0.51/hr baseline (official June 2026)<br>$0.25/hr baseline (official June 2026)

Performance & Cost

P95 Latency — 10-Minute Simulation

Annotated: burst traffic begins at minute 4, autoscale triggers at minute 6

Cost/hr at Baseline / 1.5&times; / 2&times; Burst

OCI delivers the lowest hourly cost across all three traffic levels

KPI Scorecards

Key Takeaways

Simulation Findings — 4 Quotable Conclusions

OCI costs 44% less than AWS at peak ($0.25/hr vs $0.45/hr) — VM.Standard.E4.Flex + MySQL HeatWave deliver the lowest per-hour bill across all three traffic levels, making OCI the clear cost winner in this tier.

GCP has the fastest autoscale and lowest error rate (24 s scale-out; 14.7% peak error rate) — n2-standard-4&rsquo;s generous vCPU headroom absorbs the 2&times; burst with fewer overload events than any other provider.

Azure wins on P95 tail latency at peak (155 ms at 50 k RPS) — Standard_D4s_v3&rsquo;s balanced compute-to-memory ratio keeps tail latency below all rivals even as traffic doubles; a strong choice where SLO headroom matters.

All four providers stay under 172 ms P95 at 2&times; burst — the gap between the best (Azure 155 ms) and worst (GCP 171 ms) is only 16 ms. For most SLOs the differentiator is cost or scale speed, not raw latency.

Methodology

About this report: All metrics were produced by the<br>Cloud World Model<br>simulation engine — a deterministic, capacity-aware model that calculates<br>CPU utilisation, P50/P95 latency, throughput, error rates, and cost from first-principles<br>resource profiles without provisioning any real cloud infrastructure.<br>Official hourly rates from AWS, GCP, Azure, and OCI pricing calculators<br>(June 2026, us-east-1 / us-central1 / East US / us-ashburn-1) are embedded<br>directly in the engine and validated weekly.<br>See cloudworldmodel.ai/fidelity<br>for the full accuracy specification (cost &plusmn;10%, latency &plusmn;15%).

Simulate your own architecture — free

Drop in your own resources, run burst traffic patterns, and compare AWS, GCP, Azure, OCI, and DigitalOcean side-by-side — no cloud account required.

Try Cloud World Model &rarr;