Decentralized Network Relay Aggregation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large-scale pervasive computing systems face scalability and reliability issues due to network instability and congestion, particularly in scenarios where millions of devices need to communicate, as traditional IP Multicast and peer-to-peer protocols struggle with managing network states and traffic load.
Innovation Solution
A decentralized network using a diffusion-like protocol with epidemic techniques for message propagation, where relay devices aggregate and forward compressed data files to minimize bandwidth and ensure reliable communication, employing Bloom filters for efficient data aggregation and dissemination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional IP Multicast protocols are used for event notification, then reliable communication can be provided, but network state management becomes complex and scalability is limited
Solution Approach 1:
The patent extracts the network state management function from individual routers and consolidates it at a central controller. The controller maintains the forwarding table and manages multicast group states, while routers only perform simple forwarding based on instructions from the controller. This extraction resolves the complexity of distributed state management while maintaining communication reliability.
Solution Approach 2:
The central controller acts as an intermediary between source routers and destination routers. It receives event notifications, determines appropriate forwarding paths, and sends instructions to routers. This intermediary role simplifies the complexity of direct peer-to-peer state management while ensuring reliable message delivery.
2Adaptability or versatility
If peer-to-peer protocols are used for decentralized communication, then system autonomy is improved, but traffic load management becomes difficult under high load conditions
Solution Approach 1:
The central controller serves as an intermediary that manages traffic load by receiving event notifications from source routers, processing them, and distributing instructions to appropriate destination routers. This prevents direct peer-to-peer traffic congestion while maintaining the autonomy of individual routers in executing forwarding decisions.
Solution Approach 2:
The patent segments the communication function into three parts: event notification generation at source routers, centralized processing at the controller, and forwarding instruction distribution to destination routers. This segmentation allows each component to operate independently and optimally, managing traffic load effectively under high load conditions.
3Reliability
If all event notifications are forwarded to all participants, then message delivery reliability is improved, but network bandwidth consumption increases
Solution Approach 1:
The patent applies local quality by having each router forward event notifications only to routers in its vicinity rather than broadcasting to all participants. The central controller determines appropriate forwarding paths based on local network topology and group membership, ensuring reliable delivery while minimizing unnecessary bandwidth consumption across the entire network.
Solution Approach 2:
Instead of forwarding all event notifications to all participants (excessive action), the system forwards notifications only to relevant destination routers that need the information (partial action). The central controller identifies and forwards only the necessary portions of event notifications to appropriate destinations, reducing bandwidth consumption while maintaining delivery reliability.
Data Source
AI summary
Data collected by sensors (10, 30) is transmitted to data collection points (31, 61) over a network using an epidemic protocol in which data received by a relay (3) is forwarded to a randomly-selected group of further relays (4, 5, 6) unless the same data has been previously received by the relay (3) or the data is time-expired. Data received by a relay (3) from different sources (1, 2, 30) in the same time-frame is aggregated using a Bloom filter process so that in each time frame only a single Bloom filter message is transmitted by each relay.