Dynamic Admission Control for Block Storage Overload
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems face challenges in managing workload variations and preventing overload in multi-tier e-commerce systems, as defining a global threshold for service requests is difficult due to varying and unpredictable request rates, leading to suboptimal performance and customer satisfaction issues.
Innovation Solution
Implementing a workload management system that uses admission control mechanisms, such as token buckets, to dynamically adjust resource allocation and throttle requests, allowing for temporary increases in workload by borrowing capacity from less busy resources, while maintaining overall system responsiveness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a fixed global threshold is used for admission control, then system overload can be prevented, but system responsiveness and customer satisfaction deteriorate due to inability to handle temporal and spatial workload variations
Solution Approach 1:
The patent implements dynamic admission control by replacing fixed global thresholds with per-device thresholds that adapt to changing workload conditions. The system continuously monitors workload characteristics and adjusts admission control parameters in real-time, allowing the system to respond dynamically to temporal and spatial variations in request patterns while maintaining reliable overload prevention.
Solution Approach 2:
The patent segments the monolithic global admission control into device-specific admission control mechanisms. Each storage device has its own admission control threshold and workload assessment, allowing independent optimization for different devices based on their individual performance characteristics and current workload states, thereby improving overall system responsiveness.
2Productivity
If per-storage-device admission control is implemented, then system responsiveness improves, but device complexity increases due to multiple thresholds and monitoring mechanisms
Solution Approach 1:
The patent implements a universal admission control framework that applies the same fundamental mechanism across all storage devices. The core admission control logic, workload assessment methodology, and threshold adjustment algorithm are standardized and reused for each device, reducing complexity through consistency while still allowing device-specific parameter optimization.
Solution Approach 2:
The system implements self-service admission control where each storage device autonomously monitors its own workload characteristics and adjusts its own admission control threshold based on observed performance patterns. This distributed self-adjustment reduces the need for complex centralized control mechanisms while maintaining responsive admission control for each device.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The estimated rate of work requests expected during a time period at a first block storage device, implemented at a particular server of a storage service, exceeds a provisioned rate of the first device. At a client-side component of the storage service, a different storage server is identified, at which the rate of work requests directed during the time period to a second block storage device is anticipated to be less than the provisioned rate of the second device. At least one admission control parameter of the first device is modified to enable the first storage server to accept work requests at a rate that exceeds the provisioned rate of the first device.