Microservices

Small Autonomous Services that work together, modeled around business domain

• Subset of Service Oriented Architecture(SOA)
• Allow use of multiple DB & Tech Stacks for differeent features

Principal of MicroServices

1. Domain Driven Design

• Services Modelled around Business Domain
• Service decomose by business fomain
• Slice vertically
• Team expert in perticular domain

2. Automation CI/CD

• Create a platform for deploying service.
• Automate testing & deployment process for scaling with confidence.
• Setting up a Process help scaling service easily.
• Build->Test->UAT->Performance->Prod

3. Abstraction

• Hide Implementation Detail from other service
• Hide Database and allow service to communicate over service to access DB
• Abstarction allow changing in Tables easier.
• Easier to expose Info what is hidden than hide Info what is Shown

4. Independently Deployability

• Avoid Release Train
• Must not have aneed redeploy entire System if a service changes
• Add test for dependent service using Test cases as Pact to give confidence
• Co Exist End Point-Beta API release to give consumer time to upgrade.

5. Distributed/ Decentralized Power Across Team

• Give power to team
• Team must be autonomus
• A man with power slow down the process

6. Consumer First

• Make life easy for Developer
• Good Documentation
• Provide Swagger JSON

7. Isolate Failure

• Reduce Impact of failure of Service failure
• Reduce Cascading failure

8. Highly Observable

• Logs to monitor failure
• Error Logs & Failure
• Latency of calls between service
• ID to trace service flow

---

Migration to MicroService

Monolith Types

1. Single Process Monolith

   • All code in a Single process
   • All data in single DB

2. Modular Monolith

   • Code is broken into modules.
   • Teams work on split modules
   • All Modules are packaged together into a single process
   • All data in single DB

3. 3rd Party Monolith

   • Limited to no abaility to change core system
   • Might have access to storage & API

4. Distributed Monolith

• Process communicate over process
• Complex release System
• Downside of both distributed system &
• Large scoped deployment lerage
• High cost of change & risk
• Release train to abort before release

Issue with monolith

• Scalability

CAP Theory

• In a Partion we have to tradeoff Consistancy over Availability

Compariosn to MicroService

monolith	Microservice
Simplicity	Incremental Partial Development
Consistancy of DBMS	High Availability of Services
Inter module refactoring b/n module boundry	Force Preserve Modularity all the time
Fix Tech Stack and hard to change	Flexibility of Multiple platforms Tech stack/dbms/platform

Decomposition

Incremental Upgrade

• Extract a process to Micro Service
• Increamental extract & refactor feature/process
• Wrap new functionality around old functionality

Routing

• Find issue with Network beforehand and add HTTP Proxy
• HTTP Proxy divert call to Monolith and Side Service

Branch by Abstraction

• Add abstraction around functionality
• It Give swap point for changing feature
• Start work on a new service
• Switch Service
• Clean Up

Parallel Run

• Call current service & new service
• Varify new service output & check if it behave same as Existing implementation

Accessing Data

• MS need to access same data
• Direct access cause Coupling Problem
• Go through Monolith to access DB in this case
• Dependency can be inverted in case Monolith need data from Monolith

CAP Theory

• In a Partion we have to tradeoff Consistancy over Availability

---

Security

Prevention-> Detection-> Response- >Recovery

1. Attack Tree

• Brainstorm ways attack can happen and try to prevent them
• Find possible point of attack & Possibility of cracking

2. HTTPS

• Server gurantees
• Payload not manipulated
• Client not guarantee
• Make sure Certificate is Secure and Authentic

3. Client side Certificate

• Client guarantees

4. Add Authentication (OAuth, FormAuth)

• User access management

5. Confused Deputy

• Whoever asking a data is authenticated
• Token validation

6. Security Problem in Docker Image

• Check if image is up to date
• Check if docker image have vulnaribility

7. Patch the tech stack