Input Data Categorization for Secure Memory Routing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing semiconductor memory systems struggle to efficiently manage and secure the transmission of input data based on its categorization as public or private, leading to potential breaches and non-compliance with regulatory guidelines.
Innovation Solution
A system and method for categorizing input data as private or public based on its source, using machine learning algorithms to determine appropriate storage and transmission paths, ensuring secure storage of private data and sharing of public data, while utilizing various memory types to balance performance and power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If input data is transmitted without categorization, then data transmission speed is improved, but data security and compliance are worsened
Solution Approach 1:
The system performs preliminary categorization of input data into public and private categories before transmission. This advance classification allows data to be transmitted efficiently while ensuring that private data receives appropriate security handling, thus resolving the contradiction between transmission speed and security compliance.
Solution Approach 2:
The data transmission system is segmented into different paths based on data categorization. Public data follows a standard transmission path for speed, while private data is routed through secure channels. This segmentation allows simultaneous optimization of both transmission speed and security compliance for different data types.
2Reliability
If machine learning algorithms are used to categorize data, then data security and compliance are improved, but processing time is worsened
Solution Approach 1:
The machine learning model performs categorization as a preliminary action before data transmission or storage operations. By classifying data upfront into public and private categories, the system establishes security protocols in advance, ensuring compliance without repeatedly processing data during subsequent operations.
Solution Approach 2:
The machine learning model autonomously categorizes incoming data without requiring manual intervention or repeated security checks. This self-service categorization reduces processing overhead by establishing data classification once, allowing subsequent operations to proceed efficiently while maintaining security and compliance requirements.
3Reliability
If private data is stored in secured memory, then data security is improved, but memory resource utilization is worsened
Solution Approach 1:
The memory system is segmented into secured and unsecured storage regions. Private data is automatically directed to secured memory while public data utilizes unsecured memory resources. This segmentation ensures that security-critical data receives enhanced protection while maximizing overall memory resource utilization by not unnecessarily securing all data.
Solution Approach 2:
Different security qualities are applied locally to different data types based on their classification. Private data receives high-security storage in secured memory, while public data uses standard unsecured memory. This local quality approach optimizes security where needed while preserving memory resource efficiency for non-critical data.
Data Source
AI summary
Systems, devices, and methods related to generating instructive actions based on categorization of input data are described. In an example, a method can include receiving, from an edge device and at a processing resource of a device, a plurality of input data associated with a plurality of sources communicatively coupled to the edge device and categorizing each piece of the plurality of input data as private or public based on an associated one of the plurality of sources. The categorizing can include writing each piece of data with metadata that indicates that it is private or public and/or selecting a first data path indicated as private or a second data path indicated as public. The method can include writing each piece of the plurality of input data categorized as private to a dedicated buffer or a dedicated address space of a memory resource.