GCP 설계 & 디자인

강사 : Jasen Baker

제목 : Architecting, Design and Process

 

기존 수업은 구글 상품을 개별로 다루었다면 , 이변 강좌는 전체적인 구글 디자인, 설계를 다룸

강좌가 다시 듣고 싶다면 : https://ko.coursera.org/learn/cloud-infrastructure-design-process?

Reliable Cloud Infrastructure: Design and Process | Coursera

- 관련 서적

Google SRE Book

https://landing.google.com/sre/books/

Google - Site Reliability Engineering

-  목표. 다른 환경에 같은 서비스 제공해주고 싶다면

더 작은 환경, 혹은 더 큰 환경에 적용하고 싶다면?

---> 디자인 챌린지를 받게 됨

 

- QwikLab : 구글 프러덕트를 테스트할 수 있는 곳

https://googlecloud.qwiklabs.com/

Qwiklabs

 

1. Defining the Service

- Defining the service

- Architecting

- Resiliency, scalability, disaster recovery

- Security

- Capacity planning, Cost optimization

- Deployment, monitoring, alerting, incident response

 

# Ideas will evolve ... iteratively

 

# Desigin Process

1. Begin simply. Iterate.

2. Plan for failure.

3. Measure stuff.

 

# State and Solution

Q. What is first thing for Desigining?

A. State

State : handles type of data

Worst Case ; They doesn't share teh state

 

개선 디자인; Sharing State after fronted ; 데이터가 많다면?

개선2 ; reliable이 생김, 그러나 latency 생김

* Threshold Issue

https://www.reddit.com/r/programming/comments/2bar4m/til_about_the_doherty_threshold_400ms_response/

TIL about the Doherty Threshold: < 400ms response time addicting; > 400ms painful. (old paper still very true today)

Google General Solution for large-scale cloud-based systems

# Defining a Level of Service

- SLI ; Service Level Indicator

- SLO ; Service Level Objective

- SLA ; Service Level Agreement

 

 

# Gather Requirements

1. Qualitive Requirements

2. Quntitative Requirements

3. Scaling Requirements

4. Size Requirements