Fragmented Packet Routing in Multi-Node Clusters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional approaches to handling fragmented packets in multi-node all-active clusters lead to costly operations and extraneous bandwidth usage due to asymmetrical routing and Virtual Fragmentation Reassembly (VFR).
Innovation Solution
The method involves creating a secondary flow table linked to a primary flow table to determine the primary flow owner of packet fragments, allowing for efficient transmission outside the cluster without the need for VFR, thereby conserving bandwidth and memory resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If asymmetrical routing and Virtual Fragmentation Reassembly (VFR) are used to handle fragmented packets in multi-node clusters, then traffic can be routed flexibly across different nodes, but bandwidth usage increases and operational costs rise
Solution Approach 1:
The patent applies preliminary action by pre-distributing flow table entries to appropriate nodes before fragmented packets arrive. When a packet is fragmented, the flow table already contains the routing information needed to direct all fragments to the same primary flow owner node, eliminating the need for VFR operations and reducing bandwidth consumption while maintaining routing flexibility
2Reliability
If Virtual Fragmentation Reassembly (VFR) is performed frequently to ensure proper packet delivery, then packet delivery reliability improves, but memory resources and operational costs increase
Solution Approach 1:
The patent performs preliminary action by pre-distributing flow table entries to nodes before fragmented packets arrive. This ensures that when fragments are received, the routing information is already in place to direct them to the correct primary flow owner, maintaining packet delivery reliability without requiring memory-intensive VFR operations
Solution Approach 2:
The patent extracts the VFR operation from the packet handling process by using pre-distributed flow tables to determine routing decisions. This removes the need to perform VFR operations that consume memory resources, while still ensuring reliable packet delivery through proper flow table management
3Measurement precision
If flow tables are synchronized across all nodes in real-time, then routing accuracy improves, but network overhead and complexity increase
Solution Approach 1:
The patent applies local quality by allowing each node to maintain its own flow table with locally relevant routing information. Nodes receive flow table entries distributed to them based on their role in the cluster, enabling accurate routing decisions locally without requiring complex real-time synchronization across all nodes
Data Source
AI summary
An efficient method to handle fragmented packets in multi-node all-active clusters. In one particular embodiment, a method includes receiving an initial fragment packet at a node in a cluster, creating a secondary flow table, linking the secondary flow table to a primary flow table, determining the primary flow owner of the initial fragment packet, and transmitting initial and succeeding fragment packets out of the cluster through, if possible, the primary flow owner.