Memory Unit Automatic Multicast Distribution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multicore processors and embedded systems, existing communication methods require the transmitter to know the memory or network addresses of receivers, making message distribution laborious and inefficient, especially when multiple receivers are involved, and rely on software layers for message transmission and registration, which can be cumbersome and resource-intensive.
Innovation Solution
A memory unit that automatically transmits data to multiple receivers without requiring the transmitter to know their addresses, allowing for automatic message distribution within and between processing units, reducing the need for software layers and registration requests, and enabling multicast communication through hardware-based routing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If the transmitter writes data to a memory location and the system uses conventional communication methods, then the message can be transmitted to receivers, but the transmitter must know the memory or network addresses of all receivers, making the process laborious and inefficient
Solution Approach 1:
The memory unit automatically performs the distribution of data to multiple receivers without requiring the transmitter to know receiver addresses. The memory unit monitors its own memory locations and autonomously copies data to registered receivers, making the system self-managing and eliminating the need for complex address management by the transmitter.
Solution Approach 2:
The memory unit acts as an intermediary between the transmitter and multiple receivers. Instead of the transmitter directly communicating with each receiver, the memory unit receives data from the transmitter and automatically distributes it to all registered receivers, simplifying the transmission process for the transmitter.
2Adaptability or versatility
If software layers are used for message distribution and registration, then communication between processor cores can be achieved, but the software layer becomes resource-intensive and cumbersome
Solution Approach 1:
The patent replaces the software-based message distribution mechanism with a hardware-based implementation in the memory unit. The memory unit's hardware circuitry automatically handles data copying and distribution, eliminating the need for software layers, operating system interventions, and associated computing resources.
Solution Approach 2:
The memory unit autonomously manages the registration and distribution processes through its own hardware mechanisms, eliminating the need for external software management. The memory unit independently monitors writes to its memory locations and automatically distributes data to registered receivers without software intervention.
3Reliability
If data is transmitted to multiple receivers through conventional methods, then all receivers can receive the message, but the process requires knowing all receiver addresses and is relatively laborious
Solution Approach 1:
Receivers register with the memory unit in advance before data transmission occurs. The memory unit maintains a registration table that stores the addresses of all registered receivers. When data is written to the memory unit, it automatically distributes the data to all pre-registered receivers, eliminating the need for the transmitter to know receiver addresses and enabling simultaneous delivery to all receivers.
Solution Approach 2:
The memory unit is divided into multiple independent memory locations, each capable of being monitored and distributed independently. This segmentation allows the memory unit to efficiently manage multiple receivers and their respective data subscriptions without interference, enabling parallel data distribution to multiple receivers simultaneously.
Data Source
AI summary
A memory unit, which has a plurality of memory locations for accommodating data and which is designed to copy the content of a first memory location of the memory unit, when this first memory location is written, automatically into a first memory location at least of one other memory unit, the first memory location of the at least one other memory unit being readable and writable independently of the first memory location of the memory unit; and to a data network having at least two such memory units, a transmitter and at least one receiver, the transmitter being designed to write a datum to be sent into the first memory location of a first of the at least two memory units, and the at least one receiver being designed to read and to process the datum from the first memory location of a second memory unit of the at least two memory units.