Storage Data Rate Control for Predictable Burst Transfers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data storage systems face challenges in providing deterministic and predictable data access behavior due to bursty and unpredictable data transfer behavior caused by complex background management tasks and varying media characteristics, leading to inefficiencies and increased complexity in IT systems.
Innovation Solution
Implementing storage devices that control data transfers at specified data rates and peak burst sizes, allowing for customized data streams tailored to host characteristics, enabling efficient coexistence of multiple applications and hosts, and optimizing resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If data transfers are performed at peak rates without control, then data transfer speed is improved, but resource allocation becomes unpredictable and network congestion increases
Solution Approach 1:
The patent implements periodic action by alternating between peak data transfer rates and idle periods to maintain a constant bit rate. The storage device controller schedules data transfers in periodic intervals, allowing the system to achieve high average throughput while ensuring predictable and deterministic data access behavior. This periodic scheduling prevents network congestion by distributing data transfers over time rather than concentrating them in bursts.
2Productivity
If multiple applications and hosts share storage devices with varying bandwidth requirements, then resource utilization is improved, but data access predictability deteriorates
Solution Approach 1:
The patent applies segmentation by dividing the shared storage device resources into separate data streams, each assigned to a specific host or application. The storage device controller maintains independent scheduling for each data stream, allowing multiple applications and hosts to share the storage device while preserving deterministic data access behavior for each individual stream. This segmentation enables fine-grained control over bandwidth allocation.
Solution Approach 2:
The patent implements local quality by customizing data streams according to each host's specific characteristics and requirements. The storage device controller applies different scheduling parameters, peak burst sizes, and constant bit rates to different data streams based on the specific needs of each host or application. This allows optimized resource allocation for each local context while maintaining overall system efficiency.
3Reliability
If data transfers are controlled with fine granular bandwidth allocation, then Quality of Service is improved, but device complexity increases
Solution Approach 1:
The patent applies parameter changes by controlling data transfers through adjustment of key parameters such as constant bit rate, peak burst size, and scheduling intervals. The storage device controller modifies these parameters dynamically to achieve fine-grained bandwidth allocation and improved Quality of Service. By focusing control on a limited set of critical parameters rather than micro-managing individual data packets, the system achieves sophisticated resource management without excessive device complexity.
Data Source
AI summary
A method may include receiving, at a storage device, a command for a data transfer between the storage device and a host, determining a specified data rate for the data transfer, and performing the data transfer between the storage device and the host based on the command, wherein the storage device may control the data transfer based on the specified data rate. The data transfer may include a peak portion and an idle portion. The method may further include controlling, at the storage device, a peak portion and an idle portion of the data transfer based on the specified data rate. The method may further include controlling, at the storage device, the data transfer based on a peak burst size. The specified data rate may be received from the host and/or determined by the storage device by monitoring one or more parameters of a data transfer.