Design PayPal

1. Cases:

Case 1: A user open PayPal from some websites (e.g. shopping)
Case 2: A user login PayPal
Case 3: A user pay money using PayPal
Case 4: A user search payment history

Case 5: A seller login PayPal
Case 6: A seller configure payment when user access PayPal
Case 7: A seller search payment history

2. Constraints

Users: 100 Million users
Daily Active user: (assuming 1%) 1 million users
Average Transaction Per Second: 1 million users * 1.5 / (3600 * 24) = 17.36
Peak Transaction Per Second: 17.36 * 10 = 173.10 Tran/Sec

Assuming one machine's capacity is 10 trans/sec, requiring 18 machines. 

Daily new user (assuming 0.1%): 100,000 users
Average new user Per Second: 100,000 / (3600 * 24) = 1.57
Peak new user Per Second: 1.57 * 10 = 15.7 User/Sec

Assuming one machine's capacity is 10 users/sec, requiring 2 machines. 

3. Service:

User Service
(Register, Update, Remove, Login, Logout)

Transaction Service
(TranferFromOtherSystemInPaypal, TransferFromPaypalOut, TransferBetweenPaypals)

Record Service
(Insert, Delete, Search)

4. Data

User Account Data
(Database: MySQL)

Transaction Records
(Database: MySQL (double check functions))

Transaction Log
(Files)

5. Core Algorithms

1. Information Hashing
     
2. Transaction ACID
     Atomicity: Log first (Commit or Rollback)
     Consistency: Consistency checker
     Isolation: Independent Transactions
     Durability: Redundancy

6. All in One

Privacy Manager's Input: user name and password
Privacy Manager's Output: Hashed strings