Cloud Computing Concepts, Service Models & Hardware Sizing
CSC3119 · Lab

Cloud ML Sizing, Spot Cost & Latency Architecture Simulator

Pick hardware, region and pricing for the stroke service and for nightly retraining — and feel the difference between what's cheap and what's safe.

  1. 1 · Predict0/1
  2. 2 · Experiment0/2
  3. 3 · Explain0/1
  4. 4 · Code it0/1
  5. 5 · Challenge
  6. 0/4 tasks done

Predict first

Commit to a guess before you touch anything. Being wrong here is how it sticks.

A $2.40/h GPU retrains for 20 hours on Spot at a 75% discount. Total cost?

Experiment

Move the controls and watch the outputs. Goals tick off automatically when you reach them.

Hardware, region & price

Monthly cost
$29.20
Latency per request
66.0 ms
Hours billed
730 (always on)

Right-size: tabular models run happily on CPUs; GPUs earn their cost on large neural networks.

Explain it

Put what you saw into words. This is what the exam will ask you to do.

When do you choose CPU vs GPU, and why is Spot right for training but wrong for the triage API?

Code it

Write real code against real data. Python and SQL run right here in your browser and hidden tests check your work; the AI reviews it like an examiner.

Pythoncore

Pick the cheapest safe instance

Write two functions:

  • monthly_cost(hourly, hours, spot_discount=0.0) → cost in USD, rounded to 2 dp.
  • choose(options, max_latency_ms, workload) → the name of the cheapest option whose compute_ms + network_ms ≤ max_latency_ms. For workload == "serving" ignore Spot options entirely (they can be reclaimed); for "training" Spot is allowed. Serving runs 730 h/month, training 120 h. Return None if nothing qualifies.
Ctrl/⌘ + Enter to run

Challenge

Fixed variations, a new problem every day, and unlimited AI challenges built from your current settings.

Multi-Stage Docker Image Footprint Reduction

A single-stage Docker image containing build tools, GCC, and pip cache measures 1,280 MB. Rebuilding with multi-stage compilation and python:3.11-slim produces a production image of 190 MB. How many MB are trimmed from the image?