Data Transform Accelerator Metadata Caching for Lower Latency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data transform operations in data analytics, big data, storage, and cryptographic applications are inefficient due to repeated metadata retrieval from host computing units, leading to increased latency and memory usage.
Innovation Solution
A data transform accelerator configures a data transform pipeline using metadata stored in its own memory or host memory, reducing latency by caching metadata and sharing it across multiple commands, thereby optimizing data transform operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If metadata is retrieved from host computing unit for each data transform command, then data transform operations can be performed, but latency increases and memory usage increases
Solution Approach 1:
The host computing unit pre-loads metadata into the data transform accelerator's memory before data transform operations are needed. This preliminary action ensures that metadata is already available when commands arrive, eliminating retrieval latency and improving data transform efficiency.
Solution Approach 2:
The system creates a copy of metadata in the data transform accelerator's memory separate from the host computing unit's memory. This copy allows the accelerator to access metadata locally without repeatedly querying the host, reducing latency and improving productivity.
2Loss of time
If metadata is stored in data transform accelerator memory, then metadata retrieval latency is reduced, but memory usage increases
Solution Approach 1:
The system implements local quality by storing only the necessary metadata locally in the data transform accelerator's memory rather than all data. This selective local storage reduces memory usage while still providing low-latency access to required metadata for data transform operations.
Data Source
AI summary
A method includes determining an address associated with a data transform command in a container data structure which is in the data transform accelerator. The data transform accelerator is in communication with a host computing unit. In response to a determination that the address is in the container data structure, the method includes accessing the data transform command based on the address. The data transform command is in the host computing unit. The method includes obtaining metadata based on information in the data transform command. The metadata is in the data transform accelerator or spread out in the host computing unit memory and in the memory of data transform accelerator. The method includes configuring a data transform pipeline based on the metadata. The metadata can be shared in its entirety or partially by multiple data transform commands grouped together.