2.4 KiB
| type | domain | description | confidence | source | created |
|---|---|---|---|---|---|
| claim | internet-finance | Larger service systems need proportionally fewer excess servers due to square-root scaling of variance | proven | Ward Whitt, What You Should Know About Queueing Models (2019) | 2026-03-11 |
Multi-server queueing systems exhibit economies of scale because safety margin grows sublinearly with system size
Queueing theory proves that larger service systems are more efficient per unit of capacity. If a system with R servers needs β√R excess servers for quality-of-service, then doubling the base load to 2R requires only β√(2R) ≈ 1.41β√R excess servers, not 2β√R.
The safety margin grows as the square root of system size, not linearly. This creates natural economies of scale: the proportional overhead for handling variance decreases as systems grow. A system with 100 servers needs ~10% overhead (assuming β=1), while a system with 10,000 servers needs only ~1% overhead.
This explains why:
- Large call centers are more efficient than small ones
- Cloud providers achieve better utilization than on-premise infrastructure
- Centralized service systems outperform distributed ones on pure efficiency metrics
- Pipeline architectures benefit from batching and pooling
The implication for Teleo: as processing volume grows, the relative cost of maintaining service quality decreases. Early-stage over-provisioning is proportionally more expensive than it will be at scale.
Evidence
Ward Whitt presents this as a fundamental result from multi-server queueing analysis. The square-root staffing principle directly implies sublinear scaling of overhead. The Halfin-Whitt regime formalizes this: utilization approaches 1 at rate Θ(1/√n), meaning the gap between capacity and load shrinks proportionally as systems grow.
This is observable in practice across industries: Amazon's fulfillment centers, telecom networks, and financial trading systems all exhibit this scaling behavior.
Additional Evidence (confirm)
Source: 2025-04-25-bournassenko-queueing-theory-cicd-pipelines | Added: 2026-03-16
M/M/c queue analysis demonstrates that the marginal improvement of worker N+1 decreases as N grows, providing mathematical proof that safety margins scale sublinearly. This is a fundamental property of multi-server queues, not just an empirical observation.
Relevant Notes:
- domains/internet-finance/_map
Topics:
- core/mechanisms/_map
- foundations/teleological-economics/_map