Cascade Switch Latency Compensation for Storage Frame Synchronization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Storage systems, such as RAID systems, experience performance imbalances due to varying frame arrival times caused by differences in storage positions, leading to uneven processing times and potential command sinking issues, which are difficult to manage manually or through complex programming, especially in multi-cascaded setups.
Innovation Solution
A storage apparatus with cascade-connected switches and storages, where at least one switch includes memory for storing latency values corresponding to destination addresses, allowing frames to be transmitted after a delayed time based on these values, ensuring equal latency across the system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If frames are transmitted directly without delay adjustment, then transmission speed is fast, but performance imbalance occurs due to different arrival times at storages
Solution Approach 1:
The switch performs preliminary action by calculating and storing latency values for each destination storage before frame transmission occurs. When a frame is received, the switch reads the pre-calculated latency value from the table and applies the appropriate delay, ensuring that frames arrive at all storages simultaneously despite different physical distances. This preliminary preparation of latency data enables the system to maintain both high transmission speed and performance balance.
2Reliability
If manual setting or complex programming is used to control storage performance, then performance balance can be achieved, but time and labor expenses increase
Solution Approach 1:
The switch performs self-service by automatically calculating latency values based on its position in the cascade connection and storing these values in an internal table. The switch independently determines the appropriate delay for each destination storage without requiring manual configuration or complex external programming. This self-service capability eliminates time and labor expenses while maintaining performance balance across the storage system.
3Quantity of substance
If many storages are multi-cascaded to increase capacity, then storage capacity increases, but latency variations increase causing larger performance imbalance
Solution Approach 1:
The switch applies local quality by creating destination-specific latency values tailored to each individual storage's position in the cascade connection. Rather than using a uniform delay for all storages, the switch stores and applies different latency values in its table based on the specific path and distance to each destination storage. This localized approach allows the system to accommodate many multi-cascaded storages while maintaining performance balance by compensating for each storage's unique distance from the computer.
Data Source
AI summary
A storage apparatus connectable to a computer for transmitting a frame includes a plurality of switches connected by a cascade connection, and a plurality of storages connected to the plurality of switches, wherein at least one of the plurality of switches includes a memory for storing latency value corresponding to a destination address of the frame, the destination address indicating one of the plurality of storages and the computer of a destination of the frame, the latency value indicating an amount of delaying to begin a transmission of the frame, and a port for executing a process including receiving the frame, reading out the latency value corresponding to the destination address included in the received frame from the memory, and transmitting the frame after a time corresponding to the latency value elapsed.