Selective Data Replication via I/O Driver Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data replication techniques in computer systems lack efficient mechanisms for selective and granular control over data replication, particularly in networked storage environments, leading to unnecessary data replication and potential performance issues.
Innovation Solution
A method and system for performing data replication that involves a host determining, based on predefined criteria, whether to replicate data to a second logical device on a second data storage system, with the ability to disable replication for specific types of data, applications, or file portions, and scheduling replication according to frequency or significant data changes, using an I/O driver to manage the replication process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data replication is performed for all data to ensure data safety and availability, then reliability is improved, but network bandwidth consumption increases and performance deteriorates
Solution Approach 1:
The patent segments data into different types (user data, metadata, paging data sets, dump data sets, work files, temporary files) and applies different replication policies to each segment. This allows selective replication of only critical data portions, reducing unnecessary network bandwidth consumption while maintaining data safety for important data types.
Solution Approach 2:
The patent applies different replication qualities to different data locations within the same logical device. By using mapping structures, it enables fine-grained control where specific portions of data can have replication enabled or disabled independently, allowing critical data to be replicated with high reliability while non-critical data uses minimal bandwidth.
2Loss of energy
If selective data replication is implemented to reduce unnecessary data transfer, then network bandwidth efficiency is improved, but device complexity increases
Solution Approach 1:
The patent introduces an I/O driver as an intermediary component between the host and storage system. This driver automatically determines which data should be replicated based on predefined criteria and data classification, handling the complexity of selective replication logic centrally rather than requiring complex modifications throughout the entire storage system.
Solution Approach 2:
The patent changes the parameter of replication control from a binary on/off state to a multi-level classification system. By defining different data types with different replication priorities, the system manages complexity through structured parameterization rather than ad-hoc decision-making logic.
3Reliability
If all data is replicated to remote storage systems, then data availability is improved, but replication time and processing overhead increase
Solution Approach 1:
The patent applies partial replication action by selecting only certain data types for replication based on their importance. Critical data such as user data and metadata are replicated to ensure availability, while less critical data like temporary files and paging data sets are excluded, thereby reducing total replication time while maintaining necessary data availability.
Solution Approach 2:
The patent implements different replication frequencies for different data types. By monitoring data changes and applying replication schedules, it replicates critical data more frequently or immediately when changed, while less critical data is replicated periodically or on-demand, optimizing the balance between data availability and replication time.
Data Source
AI summary
Described are data replication techniques that may include receiving, on a host, a write operation that writes first data to a target location on a first logical device having storage provisioned on a first data storage system; determining, on the host and in accordance with one or more criteria, whether the first data is to be replicated on a second logical device on a second data storage system; and issuing the write operation from the host to the first data storage system, wherein the write operation denotes whether to disable replication of the first data to the second logical device on the second data storage system. Such techniques may include partitioning a single write operation into multiple write operations and setting an indicator in each of the multiple writes denoting whether remote replication of the write operation is enabled or disabled.