You're reading from The Machine Learning Solutions Architect Handbook Practical strategies and best practices on the ML lifecycle, system design, MLOps, and generative AI

Product type Paperback

Published in Apr 2024

Publisher Packt

ISBN-13 9781805122500

Length 602 pages

Edition 2nd Edition

Languages

Python

Tools

MLOps

Concepts

Machine Learning

Author (1):

David Ping

View More author details

Table of Contents (19) Chapters

Preface

1. Navigating the ML Lifecycle with ML Solutions Architecture FREE CHAPTER

2. Exploring ML Business Use Cases

3. Exploring ML Algorithms

4. Data Management for ML

5. Exploring Open-Source ML Libraries

6. Kubernetes Container Orchestration Infrastructure Management

7. Open-Source ML Platforms

8. Building a Data Science Environment Using AWS ML Services

9. Designing an Enterprise ML Architecture with AWS ML Services

10. Advanced ML Engineering

11. Building ML Solutions with AWS AI Services

12. AI Risk Management

13. Bias, Explainability, Privacy, and Adversarial Attacks

14. Charting the Course of Your ML Journey

15. Navigating the Generative AI Project Lifecycle

16. Designing Generative AI Platforms and Solutions

17. Other Books You May Enjoy

18. Index

Networking on Kubernetes

Kubernetes operates a flat private network among all the resources in a Kubernetes cluster. Within a cluster, all Pods can communicate with each other cluster-wide without a network address translation (NAT). Kubernetes gives each Pod its own cluster private IP address, which is the same IP address seen by the Pod itself and what others see it as. All containers inside a single Pod can reach each container’s port on the localhost. All nodes in a cluster have their individually assigned IP addresses as well and can communicate with all Pods without a NAT. The following figure shows the different IP assignments for Pods and nodes, and communication flows from different resources:

Figure 6.6: IP assignments and communication flow

Sometimes, you might need a set of Pods running the same application container (the container for an nginx application) for high availability and load balancing, for example. Instead of calling each Pod by its private...