Network-on-Chip Overlay for Multicast Data Flow Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In networks of computational nodes, such as those in a multicore processor, existing data management systems face inefficiencies in data transmission and resource management, particularly in executing complex computations, due to the reliance on unicast transmissions and complex resource coordination, which can lead to increased latency and resource overloading.
Innovation Solution
The implementation of a network overlay layer and graph that logically isolates the computation layer from the network layer, allowing for the maximization of multicast transmissions and efficient data flow by distributing programmable stream controllers across the nodes, which compiles and assigns instructions to optimize data transmission paths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If unicast transmissions are used for data transfer between computational nodes, then data can be transmitted reliably between source and destination nodes, but the number of transmission operations increases and network resources are overloaded
Solution Approach 1:
The patent merges multiple unicast transmission operations into a single multicast transmission operation. The system identifies when multiple destination nodes require the same data from a source node, and consolidates these separate transmissions into one multicast operation that delivers data to all destinations simultaneously, thereby reducing network traffic and improving efficiency while maintaining reliability
Solution Approach 2:
The patent implements a network overlay layer that provides universal data distribution capability across the network. This overlay layer enables multicast functionality that can serve multiple destination nodes with a single transmission operation, making the network infrastructure more versatile and efficient for handling multiple-to-many communication patterns
2Reliability
If complex resource coordination is implemented to manage data flow between computational nodes, then data can be routed correctly to multiple destinations, but system complexity and coordination overhead increase
Solution Approach 1:
The patent segments the network communication system into two distinct layers: the underlying network layer that handles basic packet forwarding, and the network overlay layer that handles high-level data distribution logic. This segmentation allows the complex routing coordination to be isolated to the overlay layer, while the base network layer remains simple and efficient
Solution Approach 2:
The patent introduces a network overlay layer as an intermediary between applications and the underlying network infrastructure. This overlay layer acts as a mediator that implements the complex multicast routing logic and resource coordination, shielding the simplicity of the base network layer while enabling sophisticated data distribution patterns
3Quantity of substance
If multiple unicast transmissions are issued to transmit larger amounts of data, then complete data can be delivered to destination nodes, but transmission latency increases
Solution Approach 1:
The patent combines multiple sequential unicast transmissions into a single parallel multicast transmission. When a source node needs to send data to multiple destinations, the system consolidates these into one multicast operation that delivers the complete data set to all destinations simultaneously, eliminating the sequential delays inherent in multiple unicast operations and significantly reducing overall transmission latency
Data Source
AI summary
Methods and systems for executing an application data flow graph using a network of computational nodes are disclosed. In specific examples, the network of computational nodes can be a network-on-chip for a multicore processor. One method includes transitioning first application data from a first source computational node to an intermediary computational node. The method can also include providing second application data, from a computation layer of the network of computational nodes, on the intermediary computational node. The method can also include multicasting the first application data in combination with the second application data from the intermediary computational node to at least two destination computational nodes. The first source computational node, the intermediary computational node, and the at least two destination computational nodes are all in the network of computational nodes.