1. First request goes to API gateway
2. API gateway ensures https, authentication and authorization

1. Authentication
2. Authorization
3. SSL Termination
4. DDoS protection/ Throttling

1. service 1: /home, /my-cart
2. service 2: /product

1. An adapter (in gateway) can make the following calls
   1. Static files (HTML, JS, CSS) - A component can exist in the gateway for this (instead of having a service) called Static Content
   2. Personalized Recommendations
   3. Trending Products - This does not change so frequently so this can be cached in Cache of the API gateway (we can give threshold time for the Cache)
   4. Discount Offers

1. Router:
   1. If one service give low quality video (for mobile) this is chosen for mobile which is fast
   2. If another service gives high quality video (for desktop/laptop) this is chosen for other computers which is slow
2. Load Balancer:
   1. If a service is in multiple copies (because the load is high) - round robin can be used (or some algorithm like consistent hashing)

1. If backend services are not ready for latest protocols like http-2 and Websocket, API gateway can convert from new protocol to existing protocol
2. All requests can be monitored for performance and latencies
   1. Dashboards exist
3. We can expose to market an API service
   1. Based on service usage, we can charge

1. Apache
2. HAProxy
3. Nginx
4. Spring Cloud Gateway

1. AWS API Gateway
2. Azure API Gateway
3. Google Cloud Endpoints
4. Apigee

Example:

 public static void main(String[] args) {
 	new Thread(new Runnable() { // task 1 - first thread
 		@Override
 		public void run() {
 			processTax(user1);
 		}
 	}).start();
 	
 	new Thread(new Runnable() { // task 2 - second thread
 		@Override
 		public void run() {
 			processTax(user2);
 		}
 	}).start();
 	
 	heavyCalculations(); // task 3 - run by main thread
 }

1. If we have a 4 core processor, we can run the three tasks in parallel
   1. The tasks are not dependent on each other

Parallelism is about doing lot of things at once - Rob Pike

Example: ExecutorService

 public static void main(String[] args) {
 	ExecutorService es = Executors.newFixedThreadPool(4); // thread pool
 	es.submit(() -> processTax(user1));
 	es.submit(() -> processTax(user2));
 	
 	heavyCalculations();
 }

Ways of enabling parallelism: Java

1. Threads - raw threads
2. ThreadPool
   1. ExecutorService
   2. ForkJoinPool (newer)
   3. Custom ThreadPools (eg: Web Servers)
3. Note: Requires > 1 CPU cores (pre-requisite for parallelism)

Example:

 public static void main(String[] args) throws InterruptedException {
 	new Thread(() -> {
 		if (ticketsAvailable > 0) { // task accessing shared variable
 			bookTicket();
 			ticketsAvailable--;
 		}
 	}).start();
 	
 	new Thread(() -> { // task accessing shared variable
 		if (ticketsAvaialable > 0) {
 			bookTicket();
 			ticketsAvailable--;
 		}
 	}).start();
 	
 	Thread.sleep(5000);
 }