New Algorithm for Conserved Capacity Allocation Across Locations and Service Classes
A recent paper on arXiv, designated as 2608.07747, introduces a sophisticated two-tier algorithm aimed at optimizing budget distribution across various locations and two distinct service categories. The first tier emphasizes intra-class capacity adjustments to address imbalances, while the second tier facilitates inter-class capacity exchanges. This innovative strategy preserves the budget, prevents negative allocations, and ensures quick stabilization of resource distributions, operating at a computational cost of O(KN), where K represents service classes and N denotes locations. This framework is particularly beneficial for content delivery networks (CDNs) facing high demand spikes and enhances throughput management for diverse tenants.
Key facts
- The paper is published on arXiv with ID 2608.07747.
- The algorithm is two-level: intra-class redistribution and inter-class lending.
- It conserves the budget exactly and preserves non-negativity.
- It reaches a stable allocation in one iteration under stationary demand.
- The cost per cycle is O(KN) for K classes and N locations.
- It is evaluated for defending a CDN's per-domain budget under volumetric attack.
- Applications include request-rate caps, licensed throughput caps, and egress budgets.
- The paper is authored by researchers (names not provided in the source).
Entities
Institutions
- arXiv