REST API

Abstract

---

• REpresentational State Transfer
• A way to architect an interface that enables communication between Server & Client
• Usually implemented in HTTP
• Not a Network Protocol, but a set of rules - 6 Rules
• HTTP Status Code is used to tell the client the complete status of the request
• Alternatives are GraphQL & gRPC

2 Cores

---

REpresentational

• Represent resources in a way(URI) for clients to consume

State Transfer

• This architecture is Stateless Network Protocol so a state is required to be exchanged for client-server to have a continuous communication

6 Rules

---

Client-Server

Stateless

• Stateless Network Protocol

Uniform Interface

• Each resource is represented by a different URI
• With 4 main actions for clients to tell servers what it want to do. The actions can be categories into CRUD, except Create - POST, the rest is Idempotent

Create - POST
Read - GET
Update - PUT
Delete - DELETE

Cacheable

• Support cacheing because Stateless Network Protocol, Get requests are usually cached

Layer System

• We can place a load balancer in front of the client. This makes it harder for the client to fully understand the underlying architecture
• We can also add an extra layer of security, such as an API Gateway or a Web Application Firewall to enhance protection
• It also becomes easier to scale the architecture; for example, you can add Lambda functions with these layers acting as abstraction layers

Code on Demand

Best Practices

---

The endpoint/resource should be noun instead verb

• Actions verbs are represented by the request type

Use Pagination when server returns a large amount of data

• The server should accept the pagination parameters from clients
• If not specified, server should have default parameter values

Versioning

• Allow backward-compatibility when we introduce breaking changes, so consumers get enough time to move to the next version
• We can version the api by prefix the version in-front of the Resource like /v1/products, /v2/products

Benefits

---

Simple & Standardised

• Because the server does not need to programmed to keep track of any state information

Scalable

• Each request to a stateless API can be handled by any server, regardless of the state of the client or the previous requests that the client has made

High Performance

• Because Cacheable

Reliable

• Server doesn’t to need to worry about losing state information in the event of a failure

Cons

---

Requires In-Depth Documentation

Can’t Get Customised Resources

• We get all the data from an URI, then have to perform client-side manipulation to get what we actually want - waste Bandwidth

Terminologies

---

RESTful

• A service that uses REST API to communicate

Resource

• A subset of the services a server provides to the client