Content Placement Across SSD and HDD Tiers by Predicted Popularity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional digital storage systems fail to optimize data placement based on storage media type, leading to underutilization of high-throughput SSDs and inefficient data serving, especially when popular content is stored on slower HDDs.
Innovation Solution
A method and system that predicts data popularity and allocates content on different storage media types (SSDs and HDDs) based on throughput, placing popular content on high-throughput media and less popular content on lower-throughput media, using cluster hardware and popularity information to optimize data storage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If data is stored on HDDs to maximize storage capacity, then storage capacity is improved, but data throughput and access speed deteriorate
Solution Approach 1:
The patent applies local quality by assigning different storage media types (SSD vs. HDD) to different data segments based on their access patterns. Frequently accessed data segments are placed on SSDs for high-speed access, while less frequently accessed segments are stored on HDDs for cost-effective capacity, thereby optimizing both throughput and capacity locally across different data portions.
Solution Approach 2:
The storage system segments data into different categories (frequently accessed vs. infrequently accessed) and assigns them to different storage media types. This segmentation allows the system to simultaneously utilize the high-speed characteristics of SSDs for hot data and the high-capacity characteristics of HDDs for cold data, resolving the contradiction between speed and capacity.
2Productivity
If data is placed on storage clusters with large available space, then storage utilization is improved, but data access speed deteriorates
Solution Approach 1:
The system implements feedback mechanisms that monitor data access patterns and storage cluster performance. Based on this feedback, the system dynamically adjusts data placement decisions, moving frequently accessed data to high-speed SSD clusters and less accessed data to capacity-oriented HDD clusters, thereby optimizing both utilization and access speed through continuous adaptation.
Solution Approach 2:
The patent introduces dynamic data placement that adapts to changing access patterns. Rather than static allocation, the system continuously evaluates data popularity and repositions data segments between different storage media types as needed, ensuring that storage utilization remains high while access speed is optimized for current data demands.
3Speed
If SSDs are used for all data storage to maximize throughput, then data throughput is improved, but storage cost and capacity efficiency deteriorate
Solution Approach 1:
The system changes the parameter of storage media selection based on data access characteristics. Instead of using a single storage media type for all data, the system varies the storage medium parameter (SSD or HDD) according to each data segment's access frequency and importance, thereby achieving high throughput for critical data while maintaining capacity efficiency for the overall system.
Data Source
AI summary
The disclosed computer-implemented method includes accessing cluster hardware information that identifies at least two different types of storage media within a cluster and provides an indication of a respective amount of data throughput for each identified type of storage media. The method next includes accessing popularity information for digital content that is to be stored in the cluster. The popularity information indicates how often the digital content is predicted to be accessed over a specified future period of time. The method also includes allocating the digital content on the different types of storage media within the cluster according to the popularity information. Accordingly, digital content predicted to have higher popularity is placed on storage media types with higher throughput amounts, and digital content predicted to have lower popularity is placed on storage media types with lower throughput amounts. Various other methods, systems, and computer-readable media are also disclosed.