Deferring Audio Encryption to Reduce Latency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speech interface devices face performance issues due to computationally-intensive encryption processes competing with time-sensitive operations, leading to increased latency and resource congestion, which negatively impacts user experience and efficiency.
Innovation Solution
Implementing a technique to defer audio data encryption until local processing resources are not heavily utilized by other computationally-intensive operations, allowing for the storage of audio data in volatile memory until resources are available, and then encrypting it in non-volatile memory, thus avoiding resource competition and reducing latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If audio data encryption is performed immediately upon capture, then data security is improved, but processing latency increases and local resources are congested
Solution Approach 1:
The system performs preliminary actions by capturing and buffering audio data in volatile memory before encryption occurs. This allows the audio capture and initial processing to be completed without waiting for encryption, thereby reducing latency while maintaining security through subsequent encryption when resources are available.
Solution Approach 2:
The encryption process is made dynamic by adjusting its execution timing based on local resource availability. The system dynamically schedules encryption operations to occur when computational resources are free, rather than performing encryption immediately or continuously, thus balancing security requirements with processing latency constraints.
2Reliability
If audio data encryption is performed immediately upon capture, then data security is improved, but computing resources are congested
Solution Approach 1:
The system performs preliminary actions by capturing and buffering audio data in volatile memory before encryption occurs. This allows the audio capture and initial processing to be completed without waiting for encryption, thereby reducing latency while maintaining security through subsequent encryption when resources are available.
Solution Approach 2:
The encryption process is made dynamic by adjusting its execution timing based on local resource availability. The system dynamically schedules encryption operations to occur when computational resources are free, rather than performing encryption immediately or continuously, thus balancing security requirements with processing latency constraints.
3Speed
If audio data is stored in volatile memory before encryption, then processing speed is improved, but data retention time is reduced
Solution Approach 1:
The system performs preliminary actions by capturing and buffering audio data in volatile memory before encryption occurs. This allows the audio capture and initial processing to be completed without waiting for encryption, thereby reducing latency while maintaining security through subsequent encryption when resources are available.
Data Source
AI summary
A speech interface device is configured to defer encryption of audio data on-device until a time when the encryption operation is not competing with other computationally-intensive operations for responding to the audio data. For example, audio data based on sound captured in an environment of the speech interface device can be stored in volatile memory of the speech interface device, without encrypting it, until a set of processing operations (e.g., ASR processing, NLU processing, audio event processing, etc.) performed based on the audio data have stopped. Based on a determination that these processing operations for responding to the audio data have stopped, the logic may encrypt the audio data to generate encrypted data, and the encrypted data can be stored in non-volatile memory of the speech interface device for uploading to a remote system when a connection is available.