My C++ Redis Server

---

Overview

my_redis_server is a lightweight, Redis-compatible, in-memory data store written in C++.
It supports strings, lists, and hashes, full RESP protocol parsing, multi-client concurrency, and periodic disk persistence.

---

Features

• Common Commands: PING, ECHO, FLUSHALL
• Key/Value: SET, GET, KEYS, TYPE, DEL, EXPIRE, RENAME
• Lists: LGET, LLEN, LPUSH, RPUSH, LPOP, RPOP, LREM, LINDEX, LSET
• Hashes: HSET, HGET, HEXISTS, HDEL, HKEYS, HVALS, HLEN, HGETALL, HMSET
• Persistence: Data is dumped to dump.my_rdb every 5 minutes and on shutdown; loaded at startup if present.
• Concurrency: Handles multiple clients using threads.
• RESP Protocol: Fully parses both inline and array formats.

---

Directory Structure

├── include/                  # Header files
│   ├── RedisCommandHandler.h
│   ├── RedisDatabase.h
│   └── RedisServer.h
├── src/                      # Source files
│   ├── RedisCommandHandler.cpp
│   ├── RedisDatabase.cpp
│   ├── RedisServer.cpp
│   └── main.cpp
├── Concepts,UseCases&Tests.md # Design and test documentation
├── Makefile                  # Build script
├── README.md                 # This file
└── test_all.sh               # Automated test script

---

Installation

You need a C++17 (or later) compiler.

Build with Make:

make

Clean build artifacts:

make clean

Manual compilation:

g++ -std=c++17 -pthread -Iinclude src/*.cpp -o my_redis_server

---

Usage

Start the Server

./my_redis_server          # Default port 6379
./my_redis_server 6380     # Custom port

On startup, you’ll see:

Database Loaded From dump.my_rdb
# or
No dump found or load failed; starting with an empty database.

The server saves the database every 5 minutes and on shutdown (Ctrl+C).

---

Connect with a Client

You can use the standard redis-cli or your own RESP client.

redis-cli -p 6379

Example session:

127.0.0.1:6379> PING
PONG

127.0.0.1:6379> SET bob "pawan kalyan"
OK

127.0.0.1:6379> GET bob
"pawan kalyan"

---

Supported Commands

Common

• PING → PONG
• ECHO <msg> → <msg>
• FLUSHALL → clear all data

Key/Value

• SET <key> <value>
• GET <key>
• KEYS *
• TYPE <key>
• DEL <key>
• EXPIRE <key> <seconds>
• RENAME <old> <new>

Lists

• LGET <key> — get all elements
• LLEN <key>
• LPUSH <key> <v1> [v2 ...]
• RPUSH <key> <v1> [v2 ...]
• LPOP <key>
• RPOP <key>
• LREM <key> <count> <value>
• LINDEX <key> <index>
• LSET <key> <index> <value>

Hashes

• HSET <key> <field> <value>
• HGET <key> <field>
• HEXISTS <key> <field>
• HDEL <key> <field>
• HLEN <key>
• HKEYS <key>
• HVALS <key>
• HGETALL <key>
• HMSET <key> <f1> <v1> [f2 v2 ...]

---

Design & Architecture

• Concurrency: Each client handled in a separate thread.
• Synchronization: Single mutex guards all data structures.
• Data Stores:
  • kv_store for strings
  • list_store for lists
  • hash_store for hashes
• Expiration: Lazy eviction on access, with TTL map.
• Persistence: Simple text-based RDB dump/load (dump.my_rdb).
• Singleton Pattern: One shared database instance.
• RESP Parsing: Custom parser for both inline and array formats.

---

Testing

• Use test_all.sh to run a suite of commands for validation:

  ./test_all.sh

• See Concepts,UseCases&Tests.md for more examples and test cases.

---