Microservice Capacity Estimation Using Flow Network Bottlenecks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Accurately estimating the capacity of a microservice system is difficult due to complex relationships between microservices and the challenge of simulating productive traffic levels, which can lead to processing bottlenecks and inefficient resource allocation.
Innovation Solution
A method is employed to determine the sequence of microservices required for different workflows, generate a directed weighted graph, split vertices based on capacity, and use flow network theory to identify maximum flow and residual capacity, allowing for resource reallocation to optimize system capacity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If load tests are performed during development to estimate single microservice capacity, then capacity estimation for individual microservices is achieved, but accurate estimation of overall microservice system capacity remains difficult
Solution Approach 1:
The patent segments the microservice system into individual microservices, representing each as a separate node in a graph. By analyzing individual service capacities and their interconnections, the system breaks down the complex problem of overall system capacity estimation into manageable components that can be evaluated separately and then integrated.
Solution Approach 2:
The patent introduces a flow network graph as an intermediary model between individual microservice capacities and overall system capacity. This graph representation serves as a mediator that captures the complex relationships and dependencies between services, enabling systematic analysis of system-wide capacity through mathematical flow analysis.
2Reliability
If resources are allocated based on maximum anticipated workload, then system capacity is sufficient to handle peak demand, but resource costs increase compared to elastic allocation
Solution Approach 1:
The patent enables dynamic resource allocation by continuously analyzing flow network data to identify bottleneck services. Resources can be dynamically adjusted based on actual system performance and identified constraints, allowing the system to maintain reliability during peak demand while reducing resource waste during lower utilization periods.
Solution Approach 2:
The patent changes the allocation parameters from static maximum workload estimates to dynamic values based on flow network analysis. By identifying which services are actual bottlenecks versus which have excess capacity, resource allocation parameters can be optimized to match actual system needs rather than worst-case scenarios.
3Measurement precision
If productive traffic levels are simulated to estimate system capacity, then realistic capacity assessment is possible, but the complexity of monitoring and analyzing health status of each microservice increases
Solution Approach 1:
The patent creates a simplified flow network graph copy that mirrors the microservice architecture but uses abstracted capacity values rather than requiring detailed monitoring of each service's internal health status. This graphical model allows capacity estimation without the complexity of direct productive traffic simulation and comprehensive health monitoring.
Data Source
AI summary
Systems and methods include determination, for each of a plurality of different types of requests, of a sequence in which services are executed in response thereto, generation of a directed graph based on the sequences s, where each vertex of the directed graph represents a service, generation of a flow network graph by splitting each vertex of the directed graph into two vertices with a directed edge between, and associated with a capacity of the service represented by the vertex, determination of a maximum flow through the service system based on the flow network graph, determination of a residual capacity of each service based on the maximum flow and its capacity of each service, determination of services associated with a zero residual capacity, and increasing of computing resources available to the determined services.