System Design Template / Approaches

Question 1

What are the 6 sections for system I need to include?

Answer 1

(1) Feature Expectations (5 min)
(2) Performance Estimations (5 min)
(3) Design Goals (5 min)
(4) High level design - core building blocks (5-10 min)
(5) Design Deep Dive (important sections) (15-20 min)
(6) Justify Design (5 min)

Question 2

What should you cover in feature expectations?

Answer 2

(1) Use cases
(2) Scenarios that will not be covered
(3) Who will use
(4) How many will use
(5) Usage patterns

Question 3

What should you cover in performance estimations?

Answer 3

(1) Throughput (QPS for read and write queries)
(2) Latency expected from the system (for read and write queries)
(3) Read/Write ratio
(4) Traffic estimates
- Write (QPS, Volume of data)
- Read (QPS, Volume of data)
(5) Storage estimates
(6) Memory estimates
- If we are using a cache, what is the kind of data we want to store in cache
- How much RAM and how many machines do we need for us to achieve this ?
- Amount of data you want to store in disk/ssd

Question 4

What should you cover in design goals?

Answer 4

(1) Latency and Throughput requirements
(2) Consistency vs Availability [Weak/strong/eventual => consistency | Failover/replication => availability]

Question 5

What should you cover in High Level Design?Q

Answer 5

You should define the core components of the system here from a high level:

    (1) APIs for Read/Write scenarios for crucial components
    (2) Database schema
    (3) Basic algorithm
    (4) High level design for Read/Write scenario

Question 6

What should you cover in the design deep dive?

Answer 6

You should dive deep on a core feature of the system (ask the interviewer to determine what)

Cover the following:

1. How to scale the algorithm
2. Scaling the individual components - availability, scalability, consistency of each component (consistency and availability patterns)
3. Cover the following components and how they would fit in:
   - DNS
   - CDN (push vs pull)
   - Load balancers (where to use them, active/passive, algorithm, layer 4 or 7)
   - Reverse proxy
   - Application layer scaling (microservices, service directory, etc.)

• Database approach
  -> Relational vs NoSQL
  -> High level schemas for data and tables
  -> RDBMS -> master-slave, master-master, federation, sharding, Denormalization, SQL tuning
  -> or NoSQL -> Key-value, wide column, Graph, Document, Fast lookups (

RAM -> Redis, Memcached
AP -> Cassandra, RIAK, Voldemort
CP -> Hbase, Mongo, Couchbase, DynamoDB

Caches -> Client caching, CDN caching, Webserver caching, Database caching, Application caching, Cache @Query level, Cache @Object level

Eviction policies of cache:
&raquo_space; Cache aside
&raquo_space; Write through
&raquo_space; Write behind
&raquo_space; Refresh ahead

Asynchronism -> message queues, Task queues, Back pressure

Communication protocol:
- TCP
- UDP
- REST
- RPC

Question 7

What should you justify at the end of the system design?

Answer 7

(1) Throughput of each layer
(2) Latency caused between each layer
(3) Overall latency justification

Question 8

What are the high level components of the system design that should be included in your interview?

Answer 8

• Storage - Database, database schema, and caching strategy
• Network traffic (load balancer, routing strategy, distributing requests)
• Vertical vs Horizontal Scaling strategy
• Data replication and scaling strategy - sharding
• Content Delivery Network (CDN)
• Stateless web tier / stateless architecture
• Data centers
• Message queue / producer/consumer service if needed

Question 9

Latency for L1 cache reference

Answer 9

0.5 ms

Question 10

Latency for branch mispredict

Answer 10

5 ns

Question 11

Latency for L2 cache reference

Answer 11

7 ns

Question 12

Latency for Mutex unlock/lock

Answer 12

100 ns

Question 13

Latency for main memory reference

Answer 13

100 ns

Question 14

Latency for Compress 1K bytes with Zippy

Answer 14

10,000 ns

Question 15

Latency for Sending packet CA to Netherlands to CA

Answer 15

150 ms

Question 16

Latency for send 2k bytes over 1 Gpbs network

Answer 16

20,000 ns

Question 17

Latency for Read 1 MB sequentially from memory

Answer 17

250,000 ns

Question 18

Latency for Round trip with the same datacenter

Answer 18

500,000 ns

Question 19

Latency for disk seek

Answer 19

10 ms

Question 20

Latency for read 1MB sequentially from a network

Answer 20

10 ms

Question 21

Latency for read 1MB sequentially from disk

Answer 21

30 ms

Question 22

What is the 4 step process for effective system design interview performance? What is the time allocation for each?

Answer 22

Step 1 - Understand the problem and establish the design scope - 5-10 minutes

Step 2 - Propose high level design and get buy in - 10-15 minutes

Step 3 - Design deep dive - 10-25 minutes

Step 4 - Wrap up - discuss other additional points - 3-5 minutes

Question 23

What are the do’s and don’ts of system design interviews?

Answer 23

Do:
- Always ask for clarification and study the problem
- Understand the problem requirements
- Understand there’s no right or wrong answer
- Let the interviewer know what you are thinking
- Suggest multiple approaches if possible
- Go into detail about each component once you have the high level design down
- Bounce ideas of the interview and collaborate with them - treat them as a teammate

Don’ts:
- don’t be unprepared for typical interview questions
- dont’ jump to a solution without clarifying the requirements and assumptions
- don’t be afraid to ask for hints
- don’t think in silence
- don’t think your interview is done once you give the design
-