Per-Processor Data Queuing for Multi-Tier Network Capacity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing networked data processing systems face inefficiencies in utilizing transmission capacity due to connection-level prioritization, leading to underutilization and saturation issues, as prioritization is limited to connection establishment and does not account for per-processor data queuing or dynamic adjustment of network capacity allocation.
Innovation Solution
A method for prioritizing data on a per-processor basis in a multi-tier network system, where client data is identified by priority and queued accordingly, and connection capacities are dynamically reassigned between priority classes based on traffic loads to optimize network transmission capacity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If connection-level prioritization is used to service high priority clients first, then client service quality is improved, but network transmission capacity utilization deteriorates due to saturation and underutilization issues
Solution Approach 1:
The patent segments the network processing architecture by introducing per-processor data queues that are separate from connection-level queues. Each processor has its own set of data queues that can independently manage data packets, allowing fine-grained control over data transmission while preventing any single processor from becoming a bottleneck that saturates the entire network capacity.
Solution Approach 2:
The patent implements dynamic priority assignment for data packets based on real-time network conditions and client priority levels. Instead of static connection-level prioritization, the system dynamically adjusts which data packets receive priority processing at each processor queue, enabling the network to adapt to changing traffic patterns and maintain optimal capacity utilization.
2Reliability
If network capacity is partitioned into fixed allocations for different traffic priority classes, then service level agreement prioritization is improved, but overall network capacity utilization deteriorates due to idle capacity in lower priority classes
Solution Approach 1:
The patent replaces fixed capacity partitioning with dynamic capacity allocation. The system continuously monitors network traffic conditions and adjusts the capacity allocation to different priority classes in real-time. When lower priority classes have idle capacity, that capacity can be dynamically reassigned to handle bursts of higher priority traffic, ensuring both SLA compliance and optimal overall utilization.
Solution Approach 2:
The patent changes the parameter of capacity allocation from a fixed state to a variable state that can be adjusted based on network conditions. By modifying capacity allocation parameters dynamically rather than maintaining fixed partitions, the system achieves both prioritization guarantees and efficient resource utilization.
3Speed
If connection prioritization is enforced at the connection level, then high priority connections are serviced first upon arrival, but all connections are treated equally after establishment causing priority discrimination loss
Solution Approach 1:
The patent segments the prioritization function from the connection level to the data packet level. Instead of assigning priority once at connection establishment, the system assigns priority to individual data packets based on the originating client's priority level. This segmentation allows priority discrimination to be maintained throughout the entire data transmission process, not just during connection setup.
Solution Approach 2:
The patent performs preliminary priority assignment at data queue insertion points within each processor, before data packets are processed or transmitted. This preliminary action ensures that priority discrimination is established early in the data handling process and maintained throughout subsequent processing stages.
Data Source
AI summary
A method, computer program product, and a data processing system for data prioritization in a multi-tier network system is provided. A server having a plurality of processors receives data from a client. A priority of the client is then identified. Responsive to identifying the priority, the data is queued in a queue of a first plurality of queues associated with a first processor of the plurality of processors. The queue is one of a plurality of queues associated with the first processor and is associated with the priority.