viewof defaCloudDatacentres = Inputs.range([1, 5], {label: "Data centres", step: 1, value: 2})
viewof defaCloudVms = Inputs.range([2, 80], {label: "Virtual machines", step: 2, value: 20})
viewof defaCloudRequests = Inputs.range([100, 5000], {label: "Requests per minute", step: 100, value: 1400})
viewof defaCloudLatency = Inputs.range([5, 250], {label: "Network round-trip latency (ms)", step: 5, value: 60})
viewof defaCloudPolicy = Inputs.radio(["spread", "pack"], {label: "VM allocation policy", value: "spread"})Cloud Placement Simulation Lab
Cloud Placement Simulation Lab
animation
cloud-computing
platforms
Try Change VM allocation policy, then press Play.
Observe Press Play and compare Modelled utilisation, response time, and infrastructure cost.
Explain Describe how allocation policy changes response time and infrastructure cost when you press Play.
Technical boundaries. This model uses identical VMs, fixed round-trip latency, deterministic request cost, and one synthetic cost rate. It has no cold starts, data-gravity or egress fee and is not CloudSim itself.
Cloud Placement Simulation Lab
Use this small browser model to explore the placement trade-off. Change the number of data centres, virtual machines, load, latency, and allocation policy. Compare the estimated response and cost, then test the same idea in a discrete-event simulator and a real pilot. This model uses fixed assumptions and is not CloudSim itself.
defaCloudSimulation = {
const perVmCapacity = 95;
const utilisation = defaCloudRequests / (defaCloudVms * perVmCapacity);
const queuePenalty = utilisation < 0.7 ? 8 * utilisation : 8 * 0.7 + 90 * Math.pow(utilisation - 0.7, 2);
const placementPenalty = defaCloudPolicy === "spread" ? 4 : Math.max(0, 16 - defaCloudDatacentres * 2);
const response = defaCloudLatency + 12 + queuePenalty + placementPenalty;
const hourlyCost = defaCloudVms * 0.045 + defaCloudDatacentres * 0.12 + (defaCloudPolicy === "spread" ? defaCloudVms * 0.004 : 0);
const overloaded = utilisation > 1;
return html`<div class="callout callout-style-default callout-note"><div class="callout-body-container"><div class="callout-body">
<p><strong>Modelled utilisation:</strong> ${(utilisation * 100).toFixed(0)}%; response time <strong>${response.toFixed(1)} ms</strong>; infrastructure cost <strong>£${hourlyCost.toFixed(2)}/h</strong>.</p>
<p>${overloaded ? "Demand exceeds nominal VM capacity: add capacity, shed work, or test a queue/failure policy." : "Nominal capacity remains: test a data-centre or VM failure before accepting the placement."}</p>
<p><small>Assumptions: identical VMs, 95 requests/minute per VM, fixed round-trip latency, deterministic request cost, no cold starts, no data-gravity or egress fee, and one synthetic cost rate. Validation requires a timestamped simulator event log, parameter file and seed, sensitivity runs, and a real pilot measuring p50/p95 latency, errors, utilisation, and billed cost.</small></p>
</div></div></div>`;
}