An agent with one intentionally distant dependency, trace capture, region or cache control, and fallback behavior.
Education← BRAINS pathwayBRAINS · Masterclass
Networks, Latency, and Regions
Stop making the fast accelerator wait on avoidable distance.
Map every hop, set a latency budget, place services deliberately, and design degraded operation.
Network sequence, regional decision, before-and-after p50/p95 trace, and a demonstrated fallback.
Regional Latency Lab
Class requirement: create or sign in to your Alternate Clouds account, deploy the Regional Latency Lab template, and submit its live URL before the class is marked complete.
Create your Alternate Clouds account ↗Copyright and rights clearance
Use only work you own or are authorized to use.
Every submission needs a source-and-license manifest covering text, code, images, audio, video, fonts, datasets, models, weights, and AI reference inputs. Attribution alone is not permission. Questionable material stays private or is replaced before deployment or showcase.
Ask about rights clearance01Draw the request+
Map user, edge, API, queue, model, tools, storage, and response across synchronous and asynchronous work.
Applied exercise
Draw one complete task and identify its critical path.
02Region is a system choice+
Balance users, data, services, policy, accelerator availability, and recovery.
Applied exercise
Choose a region and document the dependency that could invalidate it.
03Moving large things+
Plan model weights, images, datasets, checkpoints, caching, prewarming, and egress.
Applied exercise
Measure one large transfer and decide whether to move, cache, or prewarm it.
04Failure on the wire+
Use timeouts, backoff, idempotency, circuit breakers, and visible degraded modes.
Applied exercise
Block one dependency and prove the fallback or honest failure state runs.
The complete course adds the teaching deck, Alternate Clouds lab, source-bounded guide, review rubric, and instructor feedback.
Find a class