System Design

Question 1

Single Server Design

Answer 1

HTTP Server, database, single point of failure

Question 2

Single server design with separated db

Answer 2

Two servers, Better resilency, single point of failure

Question 3

Resilience

Answer 3

The capacity to withstand or recover from difficulties

Question 4

Vertical Scaling

Answer 4

Make web server bigger, or db bigger, pros: not a lot of things to maintain

Question 5

Horizontal Scaling

Answer 5

Load Balancer, More servers if traffic grows, web servers are stateless

Question 6

Where do servers come from?

Answer 6

Own Data centers, Cloud Services, Serverless

Question 7

Failover Servers: Cold Standby

Answer 7

Do a periodic backup of DB, have a semi-ready instance of the db, redirect traffic to that db in case of failure. CONS: take long, data after backup lost. PROS: Cheap and simple

Question 8

Failover Servers: Warm Standby

Answer 8

DB is getting replicated live, switch to that db in case of failure

Question 9

Failover Servers: Hot Standby

Answer 9

Server writes to both dbs, read from second if main is down

Question 10

Failover DB: Horizontal Scaling

Answer 10

Sharding, a shard is an horizontal partition of your db, every shard has a backup.

Question 11

What is a hotspot shard?

Answer 11

The shard that is receiving most traffic

Question 12

Normalized DB

Answer 12

Data distributed across tables, relational Db. Less storage space, more lookups, updates in one place.

Question 13

Denormalized DB

Answer 13

Have everything in one query, data in one table, more storage, updates are hard.