Storage Cache Overflow Prevention via Dynamic Bandwidth Routing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In a remote copy technology where multiple primary storage systems share a network, the cache of one primary storage system can overflow even if another has sufficient free capacity, leading to reduced system usability, and there is no consideration for systems with different degrees of importance.
Innovation Solution
A computer system with multiple first and second storage systems, a network for data transfer, a network routing device, and a management computer that controls data routing based on the used capacity of each storage system, ensuring that the cache of less important systems does not overflow and prioritizing network use for more important systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the primary storage system limits data write requests when the cache is likely to be overflown, then the cache overflow is prevented, but the usability of the system lowers when one host computer writes a large amount of data
Solution Approach 1:
The patent segments the cache management by introducing separate differential data areas for each primary storage system in the remote copy group. The router divides the network bandwidth into multiple channels corresponding to each primary storage system, enabling independent cache capacity management and data transmission control for each system rather than treating them as a single unified resource.
Solution Approach 2:
The patent implements dynamic bandwidth allocation where the router dynamically adjusts the data transmission amount for each primary storage system based on real-time cache capacity status. The management server continuously monitors cache used capacities and dynamically modifies routing parameters to optimize both cache overflow prevention and system usability under varying load conditions.
2Reliability
If the primary storage system refuses data write requests to prevent cache overflow, then data loss is prevented, but all host computers are affected even when other storage systems have sufficient cache capacity
Solution Approach 1:
The patent segments the remote copy data transmission into separate channels for each primary storage system through the router. Each primary storage system has its own differential data area and dedicated bandwidth allocation, allowing the system to accept data write requests from hosts and route them to storage systems with sufficient cache capacity rather than refusing all writes when any single system's cache is full.
Solution Approach 2:
The router acts as an intermediary between host computers and multiple primary storage systems. It receives data write requests, monitors the cache capacity status of each storage system, and intelligently routes data to appropriate storage systems with sufficient cache capacity. The management server serves as another intermediary that coordinates the overall cache management and bandwidth allocation across the remote copy group.
3Use of energy by moving object
If multiple primary storage systems share a network for remote copy, then network resource utilization improves, but cache overflow cannot be prevented when one system has insufficient capacity despite other systems having sufficient capacity
Solution Approach 1:
The patent segments the shared network bandwidth into multiple dedicated channels, with each channel assigned to a specific primary storage system. The router implements this segmentation by creating separate data transmission paths and controlling the data amount for each system independently. This allows efficient network resource utilization while maintaining reliable cache capacity management for each storage system through isolated bandwidth control.
Data Source
AI summary
A computer system having a work computer, first storage systems, and second storage systems also has a network, a network routing device, and a management computer. Upon receipt of the update data from the work computer, the first storage system stores the update data in a first storage volume and stores the update data in a second storage volume until the update data is transferred to the second storage system. The management computer obtains an used capacity of the second storage volume of each of the first storage systems and controls the routing of the data by the network routing device based on the obtained used capacity to prevent the cache of each of the storage systems from being overflown.