Throughput, Capacity & Queueing

Primes about moving load through shared, finite-capacity systems: constraints on throughput (bottleneck, channel capacity, queueing, contention), strategies for managing demand against capacity (load balancing, caching, redundancy, scalability), and the statistical shape of extreme demand spikes (thundering herd, dragon king theory).

15 primes in this family — primes that sit near one another in abstraction space (k-means over structural-signature embeddings). Each is shown with its short description.

• Bottleneck — The single limiting stage that caps an entire system's throughput.
• Caching — Store for faster retrieval.
• Central Limit Theorem — Summing many independent finite-variance contributions yields a Gaussian envelope that forgets the shapes of its parts.
• Channel Capacity — Any information-bearing medium has a hard upper bound on reliable throughput that effort cannot exceed.
• Client Server Model — An asymmetric request-response relation in which one party holds a capability and many others initiate to consume it.
• Correlated Capacity Demand — When demands on a shared finite resource are tail-correlated rather than independent, capacity sized for independent peaks fails at the rare joint exceedance.
• Dragon King Theory — The largest events in some complex systems come from a distinct mechanism — synchronization or bifurcation near criticality — leaving them above the power-law tail and partly predictable.
• Interference and Contention — Competing demands for shared bottleneck degrade throughput.
• Load Balancing — Distributing work across resources so none is overloaded.
• Mediator Availability Constraint — Expert guidance scarcity limits one-to-one learning support.
• Queueing — Organizes tasks into a waiting line based on arrival and service rates.
• Redundancy — Duplicate critical components.
• Scalability — Handle growth.
• Thundering Herd — Many independent waiters are released by a single shared signal and collide at once on a finite resource, so the failure lives in the correlated timing of their release rather than in the population size or the resource capacity.
• Unevenness Waste — Variance in arrivals or service through finite shared capacity imposes its own cost — additive to the cost of the mean and rising nonlinearly near the capacity ceiling — so capacity sized to mean demand is not capacity adequate to demand.