20 System Design Concepts Explained in 10 Minutes

Comments · 563

• DR

  @draft_day · 2 years ago

  “Software engineering is about finding new and complicated ways to store and move data around”

  175

• RE

  @realdaly · 3 years ago

  this is like the slow version of Fireship

  104

• AN

  @anantprakashsingh8777 · 3 years ago

  <a href="https://www.youtube.com/watch?v=i53Gi_K3o7I&amp;t=330">5:30</a> years and years of watching YouTube videos and this has to be the most compelling way of asking for a sub yet, I would&apos;ve subscribed had I not been subscribed already.

  375

• AL

  @allthingscoding0 · 1 year ago

  I’m gonna need to vertically scale my brain, horizontally scale my attention span, and spin up a mental CDN to cache all this knowledge 🤣🤣

  123

• YO

  @yogeshputtaswamy8464 · 3 years ago

  I love how the topics are connected to one another in a simple and neat way.

  1K

• MO

  @mokoboko2482 · 2 hours ago

  the dropping acid pic gets me every time

• AN

  @ankittayal8291 · 3 years ago

  Scaling- (because getting more users) <br> <br>Vertical scaling (add resources) - Increasing CPU,RAM <br>Horizontal scaling (add replica) - Increasing number of machines &nbsp; <br> &nbsp; - scale infinitely, Redundancy, Fault tolerance, don&apos;t need good machines <br> <br>load balancer (server known as reverse proxy) {directs incoming requests to the apporiate server using algorithms like Round Robin(balance traffic by cycling through pool of servers) Hashing( hashing incoming request ID) } <br> <br>Static content- Content Delivery Networks CDN (network of servers located all over the world) (copying files from origin server to CDN servers done by push or pull basis) <br> <br>Cashing - storing frequently accessed data to RAM <br> <br>IP - Internet Protocol Address &nbsp; &nbsp; Internet protocol suite (IP , TCP, UDP) <br> <br>DNS Domain name system - buy from DNS registrar , create DNS A record (IP address of my server) <br> <br>Application layer protocol - HTTP Hypertext Transfer Protocol (client server model){client initiates a request which includes 1.Request Header(Address of destination and client, metadata), 2.Request Body} <br> <br>API Paradigms (REST API, GraphQL, gRPC) <br>REST Representational State Transfer &nbsp;(standardization around HTTP API making them Stateless and following Consisting Guidelines){successful request should include 200 OK code in its response header, bad req from client would return a 400 Bad Request code, Issue with server 500 Internal server error level code} <br>GraphQL(single request aka query)(fetch multiple resources with single request, no over fetching) <br>gRPC (RPC framework){server to server communication} <br>gRCP web - Protocol Buffers(data is serialized in binary format {storage efficient}) - performance improvement compared to JSON (used by REST API)(human readable) <br> <br>web sockets - Bidirectional Communication <br> <br>SQL (Relational Database Management System like MySQL, PostgreSQL){Efficient Storage using data structures like B-Trees, Fast &amp; Power Retrieval of data using SQL(Structured Query Language) queries since data is organized into rows and tables}(RDMS is ACID compliance(Atomicity{Every transcation is All or Nothing}, Consistency{foreing key and other constraints will always be enforced}, Isolation{different concurrent transcations wont interfere with each other}, Durability{data is stored on disk})) <br> <br> <br>NoSQL(Non relational databases)(No consistency){consistency makes DBs harder to scale} (NO enforcing of foreign key constraints){break up DB and scale horizontally (sharding)} &nbsp;DynamoDB(key-value stores), MongoDB(Document Stores), neo4j(graph DBs) <br> <br> <br>sharding done by shard key (range based sharding)(like ID of person) <br>Replication-(read only copies) &nbsp; (Leader follower Replication){ (every write would be sent to leader who sends them to followers, every read can go to leader or follower) &nbsp;Leader--&gt;(replication)--&gt;follower} - <br>(leader leader replication)Replica for every region - &nbsp;CAP theorem (to keep replicain sync) <br> <br> <br>Message Queues (System receiving more data than it can process) durable storage and can be replicated for redundancy or sharded for scalability <br> <br>(Replication and sharding){ (read/write)Leader--&gt;(replication)--&gt;follower(read)}

  52

• DE

  @DEBIEL918 · 2 years ago

  Amazing video! I hate how slow so many videos are and how much fluff they contain. You manage to explain these concepts in SUCH a simple way, which is a truly valuable and rare skill. Also +1 for the <a href="https://www.youtube.com/watch?v=i53Gi_K3o7I&amp;t=329">5:29</a> subscribe button use case.

  18

• ED

  @EdwinManual · 3 years ago

  I had to unsubscribe neetcode but I was able to see the network request.

  304

• IT

  @itsmehajar4842 · 1 year ago

  This was literally a summary of my last 4 years in university but why does it feel like it&apos;s my first time actually understanding these concepts :&apos;) I already graduated as a SWE. I never thought it was actually this deep and DB weren&apos;t just assets we use lol

  9

• KI

  @KingstonFortune · 3 years ago

  CS590 Software Architecture summarized in 10 mins, awesome...this was definitely the most bulky course and you did way better than my professor in explaining

  31

• MA

  @mayankgupta804 · 9 days ago

  I love how ACID-compliant SQL-based databases are. Always feel enlightened when learning the theory.