LOAD BALANCER DINAMIS PADA KUBERNETES
<p align="justify"> Internet users globally have increased by 82% in five years. As a result, business needs for services also increased. Increased in service requirements drive container and orchestration infrastructure technology to grow. Kubernetes is a useful tool for orchestrati...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/26196 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | <p align="justify"> Internet users globally have increased by 82% in five years. As a result, business needs for services also increased. Increased in service requirements drive container and orchestration infrastructure technology to grow. Kubernetes is a useful tool for orchestration of containers based service. Kubernetes can automate deployment, scaling, and operation of containers based service. One component of Kubernetes is load balancer. More efficient load balancer implementation can support a higher quality service. Load balancers available on Kubernetes are static. It assumes the state of the system. Dynamic load balancer instead bases the division of tasks on load information. Dynamic load balancer is expected to improve the performance of Kubernetes. In addition, this dynamic load balancer is also expected to remain capable of handling system failures. The load information used is the maximum utilization of CPU, memory, network, and disk of each worker. The use of these values is based on related studies. The load information should be collected from any Kubernetes server by the dynamic load balancer. The process of gathering information is done through gRPC. The load information that has been collected is then processed using a Basic LI algorithm that supports dynamic task division processing with stale information. Dynamic load balancer testing is done with loaded CPU, memory, network, and disk loads. In addition, Kubernetes static load balancer is also compared to see performance comparison. The test results show that dynamic load balancer successfully improve system performance in the case of a very wide variety of load testing and memory load. <p align="justify"> |
|---|