Service Chain Load Balancing via Path Weight Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for load balancing in cloud environments face challenges in efficiently managing distributed service instances across multiple regions or availability zones, leading to increased overhead, cost, latency, and throughput issues due to the lack of a global optimal approach.
Innovation Solution
A system and method that utilize metadata from cloud environments to identify optimal paths for service chains, allowing for globally optimal load balancing by determining path weights based on latency, resource overhead, and resource cost, and selecting the most efficient service chains to direct network traffic.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If service instances are distributed across multiple regions or availability zones, then service availability and reliability are improved, but overhead, cost, latency, and throughput are worsened
Solution Approach 1:
The patent segments the service chain into multiple service nodes, each capable of independent load balancing decisions. This segmentation allows the system to distribute services across multiple regions while maintaining localized control, thereby improving reliability without proportionally increasing overhead. Each service node operates semi-autonomously, making routing decisions based on local metadata and global path information.
Solution Approach 2:
The controller performs preliminary actions by pre-calculating optimal paths through the distributed service chain before traffic arrives. By using metadata from cloud APIs to determine placement, resource allocation, and connectivity information in advance, the system prepares routing decisions that minimize latency and overhead when actual traffic flows through the distributed instances.
2Productivity
If load balancing is performed at every hop level or service node level, then local load distribution is improved, but global optimal multihop processing is worsened
Solution Approach 1:
The patent merges local load balancing decisions with global path optimization by having service nodes exchange metadata and path information. The controller combines global view of the service chain with local service node decisions, allowing each node to make load balancing choices that contribute to overall optimal multihop processing rather than just local optimization.
Solution Approach 2:
Service nodes provide feedback to the controller about their current state, resource allocation, and routing decisions. The controller uses this feedback along with metadata from cloud APIs to dynamically adjust path weights and routing decisions, ensuring that local load balancing actions align with global optimal multihop processing objectives.
3Device complexity
If traditional load balancing approaches are used without cloud awareness, then system simplicity is maintained, but efficiency in managing distributed instances is worsened
Solution Approach 1:
The controller acts as an intermediary between the load balancing functionality and the cloud infrastructure. It retrieves metadata from cloud APIs about service instance placement, resource allocation, and connectivity, then uses this information to make intelligent routing decisions. This intermediary approach adds cloud awareness and efficiency without requiring complex modifications to each service node or the underlying cloud infrastructure.
Data Source
AI summary
The present disclosure is directed towards systems and methods of service chain load balancing. A controller intermediary to a client and computing infrastructure identifies a plurality of service chains. Each of the plurality of service chains include a path having an instance of a first service provided by the computing infrastructure and an instance of a second service provided by the computing infrastructure. The controller determines a path weight for each of the plurality of service chains. The path weight indicates a level of efficiency of delivering services in accordance with the service chain. The controller selects, based on a load balancing function and the path weight for each of the plurality of service chains, a service chain from the plurality of service chains to direct network traffic from a client.