Intelligent NIC Failover via Hash Engine Traffic Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In load-balanced networking environments, failover of network connections to reliable NICs can lead to NIC overloading and reduced system performance due to increased network traffic, which existing solutions fail to adequately address.
Innovation Solution
A method involving a hash engine that tracks network statistics to intelligently offload connections from a failed NIC to an underutilized failover NIC, using a hardware offload engine and delegated connection tables to manage network traffic without overloading the remaining NICs, thereby maintaining system performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If network connections are redirected from an unreliable NIC to other reliable NICs during failover, then network connection reliability is improved, but the remaining reliable NICs may become overloaded and their performance deteriorates
Solution Approach 1:
The patent segments network connections into different categories (new connections vs. failed-over connections) and assigns them to different NICs based on current load conditions. The load balancer monitors NIC utilization and dynamically directs segments of traffic to appropriate NICs, preventing any single NIC from becoming overloaded while maintaining connection reliability during failover events.
Solution Approach 2:
The patent implements dynamic failover management where the system continuously monitors NIC health and performance metrics. When a NIC fails, the load balancer dynamically adjusts traffic distribution in real-time, routing failed-over connections to currently healthy NICs while preventing overload. This dynamic adaptation ensures both reliability maintenance and performance preservation.
2Reliability
If additional network traffic is routed through remaining reliable NICs during failover, then connection availability is improved, but network processing performance of the computing device deteriorates
Solution Approach 1:
The load balancer segments traffic processing by directing different types of network traffic to different NICs. New connections are routed to NICs with lower current utilization, while failed-over connections are intelligently distributed to avoid overwhelming any single NIC. This segmentation prevents excessive network processing on any one interface, maintaining energy efficiency while ensuring connection availability.
Solution Approach 2:
The load balancer acts as an intermediary between network connections and NICs, intelligently managing traffic distribution. It monitors NIC health status and performance metrics, then mediates traffic routing decisions to optimize both availability and processing efficiency. This intermediary layer prevents direct overload of NICs during failover events.
Data Source
AI summary
A hash table in the network device driver maintains data on the traffic characteristics for each network interface (“NIC”) within a computing device. If one of the NICs in the computing device becomes unreliable, the cost function in the hash engine allows the software driver to initiate network traffic redistribution among the remaining reliable NICs in the computing device. Using this hash engine, the software driver is able to intelligently redirect each of the network connections on an unreliable NIC to a reliable NIC within the computing device, in a way that optimizes the distribution of network traffic across the remaining reliable NICs. Alternatively, if a connection is moved from an old NIC to a new NIC, the software driver can detect the moved connection and offload the moved connection to a hardware offload engine on the new NIC. With this approach, issues such as network interface overloading and computing device performance degradation may be more easily avoided when failing over network connections, thereby improving overall system performance relative to prior art techniques.