Network Load Balancing via Discrete State Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network load balancers face challenges in effectively balancing workload for clusters of caching devices, as they fail to accurately assess the readiness of caching devices to handle workload due to simplistic health checks and resource measurements, leading to inefficiencies and vulnerabilities.
Innovation Solution
A state model is introduced that reduces the load-handling capability of network devices into discrete states, allowing load balancers to select devices based on their current state, represented as SNMP variables or HTTP response codes, without needing detailed metrics, thereby optimizing load distribution and minimizing information exchange.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If existing network load balancers use simplistic health checks and resource measurements, then device complexity is reduced, but measurement precision deteriorates leading to inaccurate assessment of device readiness
Solution Approach 1:
The patent introduces a state model as an intermediary layer between the load balancer and the detailed device metrics. This state model aggregates complex device states into discrete, manageable states (such as ready, partially ready, not ready) that the load balancer can easily interpret. The state model acts as a mediator that translates detailed device metrics into simplified state information, resolving the contradiction by maintaining measurement precision through the state model while keeping the load balancer simple.
Solution Approach 2:
The state model is implemented as a self-updating mechanism that automatically monitors device metrics and adjusts device states without requiring complex load balancer intervention. The load balancer simply queries the current device states rather than implementing complex assessment algorithms itself. This self-service approach allows the system to maintain high measurement precision through continuous monitoring while keeping the load balancer architecture simple.
2Measurement precision
If load balancers exchange detailed metrics information, then measurement precision improves, but loss of information increases due to excessive information exchange overhead
Solution Approach 1:
The patent extracts only the essential state information needed for load balancing decisions from the full set of device metrics. Instead of exchanging all detailed metrics information, the system extracts and communicates only the relevant device states (ready, partially ready, not ready). This extraction approach maintains measurement precision for decision-making while minimizing information exchange overhead by filtering out unnecessary details.
Solution Approach 2:
Rather than having the load balancer collect and process detailed metrics from devices, the patent inverts the approach by having devices (or state model) proactively report their discrete states to the load balancer. This inversion reduces the information exchange burden on the load balancer while maintaining accurate workload assessment, as devices communicate only their current state rather than all possible metrics.
3Ease of operation
If load balancers use discrete state models instead of detailed metrics, then device complexity is reduced and ease of operation improves, but measurement precision deteriorates
Solution Approach 1:
The patent applies parameter changes by transforming continuous device metrics into discrete state parameters. The state model defines specific thresholds and conditions that map continuous metric values to discrete states (such as ready, partially ready, not ready). This parameter transformation maintains ease of operation through simple discrete state comparison while preserving measurement precision by using well-defined thresholds that accurately reflect device readiness conditions.
Data Source
AI summary
In one embodiment, an electronic device receives a request; obtains a current state from each of a plurality of electronic devices; and selects one of the plurality of electronic devices to service the request based on the current state of each of the plurality of electronic devices. The current state of each of the plurality of electronic devices is one of a plurality of states in a state model. Each of the plurality of states in the state model indicates a discrete level of workload for the plurality of electronic devices.