Cloud OS
Phase 1 · Simulation environment

Virtual Cloud Infrastructure,
Reimagined.

A modern cloud infrastructure simulator that demonstrates intelligent resource allocation, scheduling, monitoring and auto-scaling through a premium operating system experience.

Explore features
Cloud OS · overview
CPU & Memory
Allocation
atlas-edge-0162%
helios-compute-a81%
nimbus-worker-0137%
pulsar-ml-0188%
12
Running VMs
17
Active tasks
58%
Fleet CPU
Scheduler · Least Loaded
Overview

Cloud infrastructure, observed end to end.

A single simulated control plane for virtual machines, workload scheduling, live telemetry and elastic capacity — modelled on how real fleets actually behave.

0
Running VMs
5 regions
0%
CPU utilisation
fleet average
0%
Memory
of 384 GB
0
Active tasks
9 queued
0 Mbps
Network
egress now
0 GB
Storage
provisioned
Core features

Powerful infrastructure.
Simple experience.

Virtual machine management

Provision, pause and retire simulated nodes with live capacity, region and image metadata.

Resource scheduling

Four dispatch policies decide where each workload lands, and why.

Container-based compute

Workloads are modelled as isolated container units with declared CPU and memory requests.

Auto scaling

Capacity expands and contracts against sustained pressure thresholds.

Performance analytics

Compare policies across utilisation, throughput and fairness over time.

Real-time monitoring

Continuous CPU, memory, storage and network telemetry across the fleet.

Architecture

Every request, traced through the stack.

Workloads flow from the interface down to the runtime and back up as telemetry — the same loop a production cloud runs.

1
React interface
Cloud OS shell
2
Control plane
Simulated orchestration
3
Scheduler
Placement policies
4
Container runtime
Isolated workloads
5
Monitoring
Telemetry stream
6
Analytics
Historical insight
Scheduling

Four ways to place a workload.

Each policy trades fairness, speed and density differently. The simulator makes those trade-offs visible.

Round Robin

Equal turns, predictable rhythm

O(1)

Tasks are dispatched to virtual machines in a repeating cycle, giving every node an identical share of incoming workload regardless of its current pressure.

Advantages
  • Perfectly fair distribution
  • Constant-time decisions
  • Trivial to reason about
Trade-offs
  • Ignores real node load
  • Weak with uneven task sizes

Least Loaded

Always feed the quietest node

O(n)

Every dispatch inspects live CPU and memory pressure across the fleet and places the task on the machine with the most available headroom.

Advantages
  • Excellent utilization balance
  • Adapts to live pressure
  • Reduces hotspots
Trade-offs
  • Requires continuous telemetry
  • Costlier scheduling pass

Best Fit

The tightest possible packing

O(n log n)

Selects the node whose remaining capacity most closely matches the request, minimising leftover fragments across the cluster.

Advantages
  • Minimal resource fragmentation
  • High density packing
  • Lower fleet cost
Trade-offs
  • Leaves fragile slivers of capacity
  • Slower placement search

First Fit

The fastest good-enough answer

O(n) worst case

Scans the fleet in order and commits to the first machine that can host the workload, trading optimality for scheduling speed.

Advantages
  • Very fast placement
  • Low scheduler overhead
  • Great for burst traffic
Trade-offs
  • Front nodes saturate first
  • Suboptimal long-run balance
Analytics

Telemetry you can actually read.

Charts animate into view as the fleet reports CPU, memory, network and task outcomes.

CPU & memory · 24h
Task distribution
Network throughput
Weekly usage
Policy performance
Why this platform

Seven ideas, one coherent system.

Container-based compute
Cloud computing fundamentals
Scheduling algorithms
Real-time monitoring
Elastic auto scaling
Performance analytics
A modern, premium interface

Experience modern cloud infrastructure.

Learn more