rio: extract 3 claims from Whitt queueing models paper

- What: square-root staffing law, Halfin-Whitt QED regime, variance pooling economies of scale
- Why: Ward Whitt (Columbia) practitioner guide on queueing theory — foundational mechanism design results applicable to any multi-server pipeline
- Connections: relates to foundations/critical-systems variance and resilience claims

Pentagon-Agent: Rio <2EA8DBCB-A29B-43E8-B726-45E571A1F3C8>

domains/internet-finance/pooling demand across servers reduces required excess capacity because total variance grows as the square root of n while demand grows as n.md

@@ -0,0 +1,36 @@
+---
+type: claim
+domain: mechanisms
+description: "Economies of scale in service systems are not about bulk purchasing but about variance pooling: doubling servers less than doubles required buffer, giving large systems a structural cost advantage."
+confidence: proven
+source: "Rio; Ward Whitt (Columbia University), 'What You Should Know About Queueing Models' (2019)"
+created: 2026-03-12
+secondary_domains: [internet-finance]
+depends_on: ["square-root staffing sets optimal server count at base load plus beta times its square root making excess capacity scale sublinearly with demand"]
+challenged_by: []
+---
+
+# pooling demand across servers reduces required excess capacity because total variance grows as the square root of n while demand grows as n
+
+When independent demand streams are pooled into a single multi-server system, the system's total demand variance grows as n (number of jobs) but the standard deviation — the quantity that drives queuing delay — grows only as √n. Since the safety margin in the square-root staffing formula is β√R, doubling throughput demand R only multiplies the buffer by √2, not by 2.
+
+This is the mechanism behind economies of scale in any queuing system: not cheaper inputs, but mathematical variance reduction from pooling. Two systems of size n/2 each need combined buffer 2·β√(n/2) = β√(2n) ≈ 1.41·β√n, whereas one pooled system of size n needs only β√n. Pooling eliminates ~29% of required buffer at the 2× scale.
+
+The effect compounds: at 100× scale, the pooled system needs 10× less excess capacity than 100 separate small systems. This creates a natural structural advantage for centralized or highly integrated service architectures over distributed ones when service homogeneity allows pooling.
+
+## Evidence
+- Follows directly from the central limit theorem applied to arrival processes: sum of n independent Poisson(λ) streams is Poisson(nλ), with SD = √(nλ), so the coefficient of variation = 1/√n decreasing in n
+- Whitt (2019) makes this explicit: "larger systems need proportionally fewer excess servers" (Section on economies of scale)
+- Applied example: a contact center with 100 agents pooled together outperforms 10 centers of 10 agents each on service quality at equal total headcount
+
+## Challenges
+Pooling requires demand to be homogeneous or service to be fungible. Specialized workers, geographic constraints, or heterogeneous task types limit how much pooling is achievable in practice.
+
+---
+
+Relevant Notes:
+- [[square-root staffing sets optimal server count at base load plus beta times its square root making excess capacity scale sublinearly with demand]] — provides the formula whose β√R term encodes the pooling benefit
+- [[the Halfin-Whitt QED regime simultaneously achieves near-full server utilization and bounded delay because utilization approaches one at rate proportional to one over root n]] — the QED regime is where pooled systems operate at peak efficiency
+
+Topics:
+- [[_map]]

domains/internet-finance/square-root staffing sets optimal server count at base load plus beta times its square root making excess capacity scale sublinearly with demand.md

@@ -0,0 +1,32 @@
+---
+type: claim
+domain: mechanisms
+description: "The square-root staffing law gives a tractable formula for any multi-server system: safety margin grows as √R not R, so costs rise slower than throughput."
+confidence: proven
+source: "Rio; Ward Whitt (Columbia University), 'What You Should Know About Queueing Models' (2019)"
+created: 2026-03-12
+secondary_domains: [internet-finance]
+depends_on: []
+challenged_by: []
+---
+
+# square-root staffing sets optimal server count at base load plus beta times its square root making excess capacity scale sublinearly with demand
+
+Multi-server queuing systems achieve the best balance of service quality and capacity cost by provisioning **R + β√R** servers, where R is the number of servers required at full utilization (i.e., traffic intensity) and β is a quality-of-service parameter. The term β√R is the safety margin — the buffer that absorbs demand variance without letting queues explode.
+
+This result, derived from Halfin-Whitt heavy-traffic analysis of the M/M/n queue, is a mathematical theorem rather than a heuristic. The key implication is that the safety margin grows as the square root of base load, not linearly with it. A system handling 4× the demand needs only 2× the excess capacity buffer, not 4×. That sublinear scaling is what makes large pooled systems cheaper per unit of throughput than small ones.
+
+The β parameter encodes the service-level target: higher β means shorter expected wait times but more idle capacity. Practitioners can select β from published Erlang C tables or the Halfin-Whitt approximation, given an arrival rate λ, mean service time 1/μ, and target delay quantile.
+
+## Evidence
+- Whitt (2019) derives the square-root staffing rule formally in Section 3, showing it emerges from the heavy-traffic limiting regime of the M/M/n queue
+- The Erlang C formula is the exact calculation for the same quantity; square-root staffing is the closed-form approximation valid at scale
+- Practical validation: call center staffing models have used this formula operationally for decades (Whitt 2019 is itself a practitioner guide, written for applied use)
+
+---
+
+Relevant Notes:
+- [[optimization for efficiency without regard for resilience creates systemic fragility because interconnected systems transmit and amplify local failures into cascading breakdowns]] — complementary: square-root staffing provides the minimum resilience margin, but this claim clarifies why the margin must not be zero
+
+Topics:
+- [[_map]]

domains/internet-finance/the Halfin-Whitt QED regime simultaneously achieves near-full server utilization and bounded delay because utilization approaches one at rate proportional to one over root n.md

@@ -0,0 +1,40 @@
+---
+type: claim
+domain: mechanisms
+description: "The QED (Quality-and-Efficiency-Driven) regime proves high utilization and manageable delay are not in tension for large n, contradicting the intuition that busy systems must have long queues."
+confidence: proven
+source: "Rio; Ward Whitt (Columbia University), 'What You Should Know About Queueing Models' (2019)"
+created: 2026-03-12
+secondary_domains: [internet-finance]
+depends_on: ["square-root staffing sets optimal server count at base load plus beta times its square root making excess capacity scale sublinearly with demand"]
+challenged_by: []
+---
+
+# the Halfin-Whitt QED regime simultaneously achieves near-full server utilization and bounded delay because utilization approaches one at rate proportional to one over root n
+
+For a system of n servers, the Halfin-Whitt (1981) heavy-traffic theorem shows that as n → ∞, if the offered load is set to n − β√n for a fixed β > 0, then:
+1. Utilization approaches 1 (full efficiency) at rate Θ(1/√n)
+2. The probability of delay and expected wait time converge to nonzero but bounded constants
+
+This is the QED (Quality-and-Efficiency-Driven) regime — the unique operating point where a system is simultaneously nearly fully utilized AND provides acceptable service quality. Outside the QED regime, a system is either:
+- **Under-loaded** (QD regime): good quality but wasteful, utilization far from 1
+- **Over-loaded** (ED regime): high utilization but unbounded delays as queues grow without limit
+
+The practical implication: the correct provisioning target is not peak-load headroom (wasteful) nor average-load capacity (triggers queue explosion during variance spikes), but the QED point defined by the square-root staffing formula. This is neither intuitive nor obvious — it requires the mathematical framework of heavy-traffic limits to see that the sweet spot exists.
+
+## Evidence
+- Halfin and Whitt (1981) proved the convergence result for M/M/n queues; Whitt (2019) summarizes it for practitioners
+- The result extends to G/G/n (general arrival and service distributions) in the heavy-traffic limit, making it broadly applicable beyond Poisson arrival assumptions
+- Empirical validation comes from decades of call-center operational research applying these formulas to real staffing decisions
+
+## Challenges
+The QED regime requires accurate estimates of arrival rate λ and service time distribution. In practice, non-stationarity (time-varying λ) means systems must track demand dynamically — the static formula gives a snapshot, not a control law.
+
+---
+
+Relevant Notes:
+- [[square-root staffing sets optimal server count at base load plus beta times its square root making excess capacity scale sublinearly with demand]] — the staffing rule that targets the QED regime
+- [[complex systems drive themselves to the critical state without external tuning because energy input and dissipation naturally select for the critical slope]] — the QED regime is an engineered analog: the critical state is chosen deliberately, not self-organized
+
+Topics:
+- [[_map]]

inbox/archive/2019-00-00-whitt-what-you-should-know-about-queueing-models.md

@@ -6,7 +6,14 @@ url: https://www.columbia.edu/~ww2040/shorter041907.pdf
 date: 2019-04-19
 domain: internet-finance
 format: paper
-status: unprocessed
+status: processed
+processed_by: rio
+processed_date: 2026-03-12
+claims_extracted:
+  - "square-root staffing sets optimal server count at base load plus beta times its square root making excess capacity scale sublinearly with demand"
+  - "the Halfin-Whitt QED regime simultaneously achieves near-full server utilization and bounded delay because utilization approaches one at rate proportional to one over root n"
+  - "pooling demand across servers reduces required excess capacity because total variance grows as the square root of n while demand grows as n"
+enrichments: []
 tags: [pipeline-architecture, operations-research, queueing-theory, square-root-staffing, Halfin-Whitt]
 ---