System requirements

Functional:

1. Search and compare Hotels based on filters like budget,location etc
2. Book Hotels.
3. Real time availability updates for the users.
4. Secure Payment.
5. Personalized recommendations.

Non-Functional:

1. Low latency for hotel search service and should be highly available.
2. The booking service should be consistent. No two people should be allowed to book the same room.
3. System should be partition tolerant.

Capacity estimation

Estimate the scale of the system you are going to design...

API design

Host url - www.booking.com

API design -

1. GET /api/v1/search?name=?&location=?&price=? -> returns hotels based on filters
2. POST /api/v1/book -> Books a hotel

{

hotelId:""

roomNo:""

roomType:""

NumberofDays:""

}

3. POST /api/payment/initiate -> Creates a payemnt

{

bookingId:""

paymentAmount:""

}

Database design

1. User -> Id(PK),Name,Phone,Address,Location
2. Hotel -> Id(PK),Name,Location,Pin,Phone
3. Booking -> Id(Pk),UserID(FK),HotelID(FK)
4. Payment- Id(PK),BookingId(FK),UserId(FK),HotelID(FK)

1. We will be using MySQL as the primary DB for all the services.

1. For real time room availability, the client will setup a websocket connection with the server. The server will poll the 3rd party hotel API every X minutes for the real time availability. Whenever the room data changes for the hotel a notification is sent to the client.

High-level design

1. Client( Browser)
2. API gateway (Auth,Rate Limiting,etc)
3. Search Service -> For searching Hotels
4. Booking Service -> For user Bookings

Request flows

1. The client will search for hotels based on the filters from the browser.
2. The request will go from the API gateway to the search service. The search service will fetch the data from the DB and return it.
3. Once the user selects a hotel room and clicks on book, a new booking will be created, and the user will be redirected to the payments page.
4. Once the user initiates a payment invoice is generated.

Detailed component design

1. Database can be sharded based on the location for high availability and throughput.
2. Caching commonly searched queries to reduce latency time.
3. CDN for caching static assets.

Trade offs/Tech choices

1. MySQL is chosen as the DB because of structured data.
2. AWS lambda can be used for servers. (search service and booking service)
3. AWS API gateway for Auth,Rate limiting etc.

Failure scenarios/bottlenecks

Try to discuss as many failure scenarios/bottlenecks as possible.

Future improvements

What are some future improvements you would make? How would you mitigate the failure scenario(s) you described above?