Distributed Quota Estimation via Parent Server Mediator
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Service provider systems face challenges in accurately tracking and enforcing quotas across multiple request processing servers without requiring real-time communication between them, which can lead to overconsumption of resources.
Innovation Solution
Implementing a method where each request processing server estimates the total number of requests by using a fill factor, synchronizing with a parent server to gather data, and updating the fill factor using specific functions to determine whether to approve or reject requests based on the quota, allowing for near-actual quota estimation without extensive resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If real-time communication between multiple request processing servers is implemented to accurately track quota consumption, then measurement precision of quota consumption is improved, but device complexity and communication overhead increase
Solution Approach 1:
A parent server is introduced as an intermediary that receives snapshots from child servers and provides estimated quota consumption data back to them. This mediator eliminates the need for direct real-time communication between child servers, reducing system complexity while maintaining accurate quota tracking through the parent server's coordination
Solution Approach 2:
Instead of requiring all servers to maintain complete real-time knowledge of quota consumption across the entire system, each child server works with a copied/estimated view of the total consumption derived from snapshots and fill factors. This copying approach reduces communication overhead while providing sufficient accuracy for quota enforcement
2Measurement precision
If frequent synchronization between servers is performed to maintain accurate quota estimates, then measurement precision of distributed consumption is improved, but use of energy and communication resources increases
Solution Approach 1:
Child servers perform snapshots of their request counts at periodic intervals rather than continuously, and the parent server updates fill factors periodically based on these snapshots. This periodic action maintains adequate estimation accuracy while significantly reducing the energy and communication resources required compared to continuous synchronization
Solution Approach 2:
The system uses a fill factor that may be intentionally biased or updated at less-than-optimal frequency, accepting some estimation inaccuracy in exchange for reduced synchronization overhead. This partial action approach balances resource consumption with acceptable quota tracking precision
3Device complexity
If each server independently tracks quota consumption without sharing data, then device complexity is reduced, but reliability of quota enforcement deteriorates due to potential overconsumption
Solution Approach 1:
The parent server receives periodic snapshots from child servers and calculates fill factors that represent the distribution of request loads across servers. This feedback mechanism allows the system to maintain reliable quota enforcement by comparing individual server consumption against the estimated total derived from aggregated feedback, preventing overconsumption without requiring complex inter-server communication
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for estimating distributed consumption of a quota. One of the methods includes receiving, by a first request processing server, a request having one or more specified characteristics; identifying a quota; determining, by the first request processing server, an actual number of requests having the one or more characteristics received by the first request processing server since a most recent synchronization with a parent server; generating, by the first request processing server, a current estimated total number of requests having the one or more characteristics that have been received; determining, by the first request processing server, that the quota has not been reached; and approving, by the first request processing server, the received request.