MIMO Workload Model for Dynamic Distributed Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Dynamic distributed electronic storage systems pose challenges due to uncertainties in connectivity, device location, and dynamic topology, making it difficult to predict and manage storage resources effectively using conventional methods, which are inadequate for systems with multiple uncertainties and rapid changes.
Innovation Solution
A multiple-input multiple-output (MIMO) workload model is generated to manage storage resources in dynamic systems, using cost functions based on power consumption, performance, and reliability, allowing for real-time adaptation and optimization of data dispersion or aggregation through an information dispersion algorithm.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional workload modeling methods are used in dynamic distributed storage systems, then system management can be simplified, but the models cannot accurately predict system behavior due to uncertainties in connectivity, device location, and dynamic topology
Solution Approach 1:
The patent applies dynamics by transitioning from static workload models to dynamic workload models that continuously adapt to changing system conditions. The system monitors connectivity, device location, and topology changes in real-time, updating workload predictions accordingly. This allows the model to maintain accuracy despite the dynamic nature of mobile computing environments.
Solution Approach 2:
The patent implements feedback mechanisms where system performance data and workload patterns are continuously collected and fed back into the workload model. This feedback loop enables the model to learn from actual system behavior and refine its predictions, addressing the inaccuracies that would otherwise result from using static models in dynamic environments.
2Reliability
If all content is copied to all devices to ensure uniform storage access, then storage availability is improved, but power consumption and communication bandwidth usage increase excessively
Solution Approach 1:
The patent applies local quality by customizing storage distribution strategies for different devices based on their individual characteristics, usage patterns, and connectivity conditions. Rather than uniform copying, the system determines optimal content distribution for each device locally, considering factors like storage capacity, power availability, and access requirements to balance reliability with energy consumption.
Solution Approach 2:
The patent utilizes parameter changes by dynamically adjusting replication factors, storage locations, and data distribution strategies based on changing system conditions. The workload model parameters are updated in real-time to reflect current connectivity and device states, enabling the system to optimize between storage availability and power consumption without fixed rigid rules.
3Reliability
If data is dispersed across multiple devices to improve storage redundancy, then system reliability is improved, but communication channel bandwidth is substantially utilized causing latency
Solution Approach 1:
The patent applies partial action by selectively replicating data only where necessary and optimal, rather than universally dispersing all content to all devices. The workload model identifies critical data and determines appropriate replication levels based on connectivity conditions and device capabilities, achieving sufficient redundancy without excessive communication overhead that would cause latency.
Data Source
AI summary
A method, apparatus and computer program product are provided for generating a multiple-input, multiple output (MIMO) workload model of a distributed storage environment on a plurality of predictive controller devices. The MIMO models can then be utilized by the predictive controller to manage storage resources on a distributed storage system.