Token-Based Admission Control for Equitable Resource Distribution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems face challenges in managing workload variations and preventing overload scenarios in multi-tier e-commerce systems, as existing solutions fail to account for differences in service requests and varying workloads, leading to inefficiencies and potential service degradation.
Innovation Solution
Implementing a token-based admission control mechanism that uses provisioned-capacity and burst-mode token buckets to manage work requests, allowing for dynamic adjustment of throughput capacity and prioritization of requests based on available tokens, thereby preventing overload and ensuring efficient resource allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a fixed or varying global threshold is used to compare request rates, then the system can prevent overload by refusing service when the threshold is crossed, but it cannot account for differences in work requirements for different service requests, leading to inefficient resource allocation
Solution Approach 1:
The patent assigns different token consumption rates to different service types (e.g., read operations consume 1 token, write operations consume 3 tokens). This local differentiation allows the system to account for varying work requirements of different service requests while maintaining a unified token bucket mechanism, thereby improving resource allocation efficiency without compromising overload prevention
Solution Approach 2:
The system dynamically adjusts the token consumption parameter based on service type and workload characteristics. By changing the token consumption rate parameter according to the specific service being requested, the system can more accurately reflect the actual resource cost of each operation, resolving the contradiction between reliable overload prevention and efficient resource allocation
2Reliability
If service is denied or requests are throttled to protect from crushing loads, then system overload is prevented, but customer satisfaction and service quality deteriorate
Solution Approach 1:
The token bucket mechanism performs preliminary admission control by checking token availability before service execution. Requests that would cause overload are rejected in advance with clear token-based reasoning, while legitimate requests are processed normally. This preliminary filtering protects system stability without unnecessarily denying service to valid requests, thereby maintaining customer satisfaction
Solution Approach 2:
The system provides feedback to clients about token bucket status and admission decisions. When requests are denied due to token exhaustion, the system can communicate this clearly to clients, allowing them to retry later when tokens are replenished. This feedback mechanism transforms arbitrary rejections into informed decisions, improving customer understanding and satisfaction while maintaining system stability
3Adaptability or versatility
If token buckets are used to manage admission control, then workload variations can be handled dynamically, but the system complexity increases due to multiple buckets and token distribution mechanisms
Solution Approach 1:
The patent combines multiple token buckets (provisioned-capacity bucket and burst-mode bucket) into a unified admission control framework. The service management component coordinates these buckets and manages token distribution automatically, transforming what would be complex separate mechanisms into an integrated system that handles both steady-state and burst workloads through a single coherent process
Solution Approach 2:
The token distribution mechanism operates autonomously based on predefined rules and current system state. The service management component automatically refills tokens, distributes them to work targets, and adjusts parameters without manual intervention. This self-service capability reduces operational complexity while maintaining high adaptability to workload variations
Data Source
AI summary
Methods and apparatus for equitable distribution of excess shared-resource throughput capacity are disclosed. A first and a second work target are configured to access a shared resource to implement accepted work requests. Admission control is managed at the work targets using respective token buckets. A first metric indicative of the work request arrival rates at the work targets during a time interval, and a second metric associated with the provisioned capacities of the work targets are determined. A number of tokens determined based on a throughput limit of the shared resource is distributed among the work targets to be used for admission control during a subsequent time interval. The number of tokens distributed to each work target is based on the first metric and/or the second metric.