Dynamic Network Load Balancing via Roundtrip Heuristic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network load balancing solutions lack a centralized mechanism to dynamically adjust load distribution among servers in a cluster, leading to inefficiencies due to the absence of a single point for aggregating dynamic data, making it difficult to manage load imbalances effectively.
Innovation Solution
A load balancing system that determines relative server loads based on packet handling times and transfers load between servers by constructing a logical tree of nodes for recursive and parallel operations, allowing for dynamic redistribution of network traffic based on processing times and IP address subsets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a single-tier front-end server model is used for load balancing, then load distribution can be managed centrally, but the system cost increases requiring high-end servers or hardware load balancers
Solution Approach 1:
The patent extracts the load balancing function from a dedicated front-end server and distributes it across all target servers in the cluster. Each server runs load balancing software that participates in cluster-wide load distribution, eliminating the need for expensive hardware load balancers while maintaining centralized management capabilities through coordinated communication between servers.
Solution Approach 2:
The patent makes each target server multi-functional by enabling them to perform both their primary application functions and load balancing functions simultaneously. This universal approach allows commodity servers to serve dual purposes, reducing the need for specialized expensive hardware while maintaining effective centralized load management across the cluster.
2Quantity of substance
If a distributed software-based load balancing solution is used, then system cost decreases using commodity switches and software, but the ability to aggregate dynamic cluster data is lost
Solution Approach 1:
The patent merges the load balancing control functions across all distributed servers while maintaining their individual operational independence. Through coordinated communication and shared data structures, the system achieves unified cluster-wide visibility and dynamic load adjustment capabilities without requiring a centralized controller, thus combining the benefits of both distributed architecture and centralized management.
Solution Approach 2:
The patent implements feedback mechanisms where each server continuously monitors its own load status and communicates with other servers to aggregate cluster-wide dynamic data. This distributed feedback system enables real-time load adjustment based on current cluster conditions, maintaining effective centralized management capabilities through coordinated information exchange without losing the distributed architecture's cost benefits.
3Adaptability or versatility
If load balancing is distributed across target servers, then deployment flexibility increases, but dynamic load adjustment becomes difficult without centralized data aggregation
Solution Approach 1:
The patent implements dynamic load balancing where each server continuously adjusts its load distribution based on real-time cluster conditions. The system maintains deployment flexibility of distributed architecture while enabling dynamic load adjustment through coordinated communication between servers, allowing each server to adapt its behavior based on aggregated cluster state information from other servers.
Solution Approach 2:
The patent establishes preliminary communication channels and data sharing mechanisms between distributed servers before load balancing operations begin. This pre-configured infrastructure enables servers to quickly aggregate dynamic cluster data and perform load adjustments without centralized coordination overhead, maintaining both deployment flexibility and ease of dynamic load management.
Data Source
AI summary
Described is a technology by which a relative load of network traffic handling is determined between servers of a cluster, based upon time values that correspond to the time taken by each server to handle a packet. Load may then be transferred between the servers based upon the relative load, for example by having a less-loaded server take some of the responsibility for processing incoming traffic from a more-loaded server. For example, the processing time of a server may be determined by when a receiving server receives a request packet, and when that server sends a return packet. A round trip time for a request and return communication may also be established. A logical tree of nodes representing the servers may be constructed to select pairs of servers for balancing with one another, with the selection algorithm operating recursively, in parallel, and/or repeatedly, until the cluster is balanced.