Storage Tiering via Access Pattern Transmission
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In data replication systems, discrepancies in data storage management between primary and secondary storage devices occur due to differences in access patterns, leading to inefficiencies when the secondary device needs to assume the workload of the primary device in case of failover, as the replicated data may not be stored as efficiently or effectively.
Innovation Solution
A method is implemented where access pattern information, including access rate, frequency, and time of last use, is monitored and transmitted from the primary storage array to the secondary array, allowing the secondary array to manage data storage tiers to match the performance requirements of the primary array, ensuring efficient data management and handling of workloads during failover.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If access pattern information is fully transferred from primary to secondary storage array, then data management efficiency at secondary array is improved, but network bandwidth consumption and data transfer overhead increase
Solution Approach 1:
The patent extracts only the essential access pattern information (access frequency, access timing, access size) from the complete raw data access information. This selective extraction allows the secondary storage array to infer storage tiering decisions without receiving the full volume of access logs, thereby improving data management efficiency while minimizing network bandwidth consumption.
Solution Approach 2:
Instead of transferring complete access information and having the secondary array process it, the patent inverts the approach by having the primary array pre-process access patterns and transmit only the distilled results. This inversion reduces the data transfer burden while maintaining the ability to make informed storage management decisions at the secondary array.
2Reliability
If secondary storage array independently manages data tiers using access pattern information, then failover performance is improved, but the complexity of coordination between primary and secondary arrays increases
Solution Approach 1:
The patent implements preliminary action by having the primary storage array continuously monitor and analyze access patterns, and pre-calculate storage tiering recommendations before any failover event occurs. This preparation ensures that when failover happens, the secondary array can immediately assume optimal data management without complex real-time coordination, thereby improving failover performance while managing complexity.
Solution Approach 2:
The patent introduces access pattern information as an intermediary that mediates between the primary and secondary storage arrays. Instead of requiring direct complex coordination between the arrays, this information intermediary carries the necessary guidance from primary to secondary, simplifying the coordination mechanism while enabling independent tiering management at the secondary array.
3Measurement precision
If all raw data access information is transmitted to secondary array, then storage tiering accuracy is improved, but transmission time and network overhead increase
Solution Approach 1:
The patent applies parameter changes by transforming the complete set of raw access parameters into a condensed set of essential parameters (access frequency, access timing patterns, access size distributions). This parameter transformation maintains the precision needed for accurate storage tiering decisions while dramatically reducing the amount of data that needs to be transmitted, thereby resolving the contradiction between accuracy and transmission time.
Data Source
AI summary
Synchronization of data layouts and resource utilizations at one or more remote replica sites with the workload and data access statistics at the primary site allows for an efficient and effective workload support transfer in the event of site failover from a primary site to a remote site. Relevant data access information about workload being supported at the primary site is monitored and access pattern information is generated that provides relevant information about frequency and/or rate of access of data on the primary site. The access pattern information is generated and transmitted by a host and/application running on the host. The access pattern information is exported and imported to enable independent decisions to be made a remote site, storing replicated data from the primary site, on how best to utilize its available resources to match the performance requirements currently being supported by the primary site.


