Search

Mian Zubair
Mian Zubair @zubairm_ ·
The difference between a junior and senior engineer isn't the code they write. It's how they set up the system before writing any code. Circuit breakers on day one. Not after the outage. #SystemDesign
2
Sumit Kandpal
Sumit Kandpal @sumitkandpal_ ·
🚀 System Design Day 1 System Design = building apps that handle many users. Flow: User → Server → DB → Server → User Goals: Fast • Scalable • Reliable Learning in public! #SystemDesign #100DaysOfCode
2
16
Rohan Mehta
Rohan Mehta @RohanM0190 ·
Ancient architects built for 1,000 years. Modern devs build for the next sprint. We need to find the middle ground for the future of finance. 🏛️⛓️ #SystemDesign #Architecture
1
Supaboard.dev
Supaboard.dev @supaboardapp ·
How would you design a fraud detection system for 1M transactions per day? Think about: scoring in under 50ms without blocking payments, adapting to new fraud patterns your model hasn't seen, and keeping false positive rates below 0.1%. What's your approach? #systemdesign
3
Rupam Das
Rupam Das @Rupam_irp ·
#Day6 of Decoding System Design 🚀 Going deeper into how data is actually transferred and secured. Today’s focus: 1️⃣ Transport layer (TCP vs UDP) 2️⃣ TLS/SSL (how HTTPS secures communication) 3️⃣ Evolution of HTTP (1.0 ➡️ 1 ➡️ 2 ➡️ 3) #SystemDesign #Networkingg
Understanding TCP vs UDP 👇 Both operate at the Transport Layer (L4), but with very different goals: TCP (Transmission Control Protocol) • Connection-oriented (3-way handshake) • Reliable delivery (ACKs + retransmissions) • Ordered data (sequence numbers) • Flow & congestion control 👉 Guarantees correctness, but adds latency Use cases: Web (HTTP/HTTPS), APIs, file transfer, databases UDP (User Datagram Protocol) • Connectionless (no handshake) • No guarantee of delivery • No ordering • No retransmission 👉 Faster, but best-effort delivery Use cases: Streaming, gaming, VoIP, DNS Key intuition: TCP → “Make sure everything arrives correctly” UDP → “Send fast, don’t wait” Extra insight: • TCP packets = segments • UDP packets = datagrams • MTU (~1500 bytes) limits packet size Applications decide how to handle loss: • Retry (like HTTP) • Ignore (like video streaming)
Understanding TLS/SSL (how HTTPS works) 👇 1. SSL vs TLS • SSL → old (deprecated) • TLS → modern secure protocol 2. Where TLS fits Flow: DNS → TCP handshake → TLS handshake → HTTP request TLS happens before any actual data is sent 3. Why TLS is needed Without encryption: Attackers can intercept communication (Man-in-the-Middle attack) 4. How TLS works (simplified) Step 1: Client Hello • Sends supported encryption methods Step 2: Server Hello • Chooses method + sends certificate Step 3: Verification • Client verifies certificate (trusted authority) Step 4: Key exchange • Both derive a shared session key 👉 Now communication is encrypted 5. TLS 1.3 improvements • Faster handshake (1-RTT) • 0-RTT for resumed connections • Only strong cipher suites
HTTP/1.0 • One request per connection • No reuse → slow HTTP/1.1 • Persistent connections (keep-alive) • Pipelining (limited) • Better caching Problem: 👉 Head-of-line blocking (one request blocks others) HTTP/2 • Single TCP connection • Multiple parallel streams (multiplexing) • Header compression (HPACK) • Server push Improvement: 👉 Better utilization of one connection Limitation: 👉 Still suffers from TCP-level head-of-line blocking HTTP/3 (QUIC over UDP) • Uses UDP instead of TCP • Independent streams (no blocking across streams) • Built-in TLS (encryption mandatory) • Faster connection setup (0-RTT possible) Key intuition: HTTP evolution = solving bottlenecks of previous versions • HTTP/1 → too many connections • HTTP/2 → TCP bottleneck • HTTP/3 → removes TCP limitation
10
Nishant dev
Nishant dev @nishantdevx ·
🚀 Day 26/30: Database Sharding When one DB becomes a bottleneck → shard the data. Split records across multiple databases using a shard key. But beware: Sharding increases massive operational complexity. Use only when scaling demands it. #systemdesign #databases #Day26
5
Adarsh Melath
Adarsh Melath @AdarshMelath__ ·
When one lone server starts to strain, A new design must ease the pain. Break the load, spread the key, A world of shards, for you and me. Dive into how sharding transforms scalability! medium.com/@adarshmelath1… #DatabaseSharding #SystemDesign #Backend
The Digital Expansion: From Single Server to a World of Shards

Imagine you and your friend start a tech company in a small garage. As your product gains users, you begin hiring more employees. At first…

From medium.com
15
PRAJAKTA
PRAJAKTA @prajakta_mane_ ·
Replying to @prajakta_mane_
(no locks, no consensus protocol needed) CRDTs rely on one key property: all operations must be commutative, associative, and idempotent — meaning the order you apply them doesn't matter, and applying the same op twice is safe. This guarantees convergence. #systemdesign
31
StackPilot
StackPilot @TheStackPilot ·
Last seat. Two users. One conflict. ✈️ If your system says “confirmed” twice… you’ve already lost. Here’s how real systems prevent double-booking 👇 #SystemDesign #BackendC
19