Two-Dimensional Scheduler for Bursty Data Traffic Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques for managing data traffic in multiple queues often lead to overflow issues, restricting data bursts and increasing system costs due to the need for larger queues and increased congestion, as they react to congestion rather than prevent it and rely on costly in-band flow control messages.
Innovation Solution
A two-dimensional scheduler regulates data flow by adhering to quality of service agreements, ensuring random inbound traffic meets deterministic guarantees, and preventing queue overflow by dynamically managing permissions for input ports to forward data based on queue status information, ensuring fair share and minimum bandwidth allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional flow control techniques are used to prevent queue overflow, then queue overflow is prevented, but data bursts are restricted and system costs increase
Solution Approach 1:
The patent applies preliminary action by proactively regulating data flow into queues before congestion occurs. The scheduling system monitors queue status information and dynamically controls input port permissions in advance, preventing overflow conditions rather than reacting to them after they occur. This allows the system to maintain reliability while preserving data burst capabilities.
2Quantity of substance
If larger queues are used to handle data bursts, then data burst capacity is increased, but system cost and hardware size increase
Solution Approach 1:
The patent applies dynamics by implementing a dynamic scheduling system that adjusts data flow regulation based on real-time queue status. Rather than using static large queues, the system dynamically controls when and how data enters queues, allowing small queues to handle variable traffic patterns efficiently. This reduces hardware requirements while maintaining data burst handling capability.
3Ease of operation
If conventional flow control messages are transmitted, then flow control is achieved, but overall congestion increases
Solution Approach 1:
The patent applies the extraction principle by removing flow control messages from the data traffic stream entirely. Instead of using in-band flow control messages that add to congestion, the system extracts flow control functionality into a separate scheduling mechanism that operates out-of-band. This eliminates the harmful effect of message-induced congestion while maintaining effective flow control.
4Reliability
If conventional systems react to congestion, then congestion response is provided, but congestion prevention is not achieved
Solution Approach 1:
The patent transforms the reactive congestion response approach into a proactive congestion prevention approach. The scheduling system continuously monitors queue status and regulates data flow in advance, preventing congestion before it occurs. This eliminates the time loss associated with congestion events while maintaining system reliability through continuous proactive management.
Data Source
AI summary
A system includes multiple input ports that forward received data (e.g., data packets) to each of multiple queues. Data received at the input ports of the system can be somewhat random or “bursty” at times. That is, the input ports can receive data at a variable bit rate or unspecified bit rate from an internal system source or an external source such as an FTP (File Transfer Protocol) server or SCSI disk array. The queues output data at a constant bit rate. A two-dimensional scheduler associated with the system forces random inbound server traffic from the input ports to adhere to a QoS (Quality of Service) agreement such that the random nature of the inbound traffic does not negatively affect the deterministic guarantees of existing server traffic output from the queues. In other words, techniques herein ensure adherence to QoS requirements among the data flows, without overflowing the queues.