NoSQL Overview

NoSQL Use Cases

Traditional relational databases face limitations with modern web-scale applications:

• Single-machine MySQL systems struggle with large datasets exceeding storage capacity
• Index sizes can surpass available memory resources
• Read/write workloads overwhelm single-server capabilities

Caching solutions like Memcached emerged to reduce database load by storing frequently accessed data in memory. Modern applications require NoSQL databases to handle diverse data types including user locations, social graphs, and unstructured logs.

NoSQL Characteristics

• Horizontally scalable architecture
• High performance for read/write operations
• Schema-less data models
• Eventual consistency guarantees

Redis (Remote Dictionary Server) is an in-memory key-value store supporting persistence and multpile data structures.

Installation

# Linux installation
wget https://download.redis.io/releases/redis-6.2.6.tar.gz
tar xzf redis-6.2.6.tar.gz
cd redis-6.2.6
make
make install

Database Operations

SELECT 3       # Switch to database 3
DBSIZE         # Check key count
FLUSHDB        # Clear current database
SET user:101 "data" EX 3600  # Set key with expiration

Core Data Structures

Strings

APPEND buffer "data"  # Append to value
INCR counter          # Atomic increment
GETRANGE text 0 4     # Substring extraction

Lists

LPUSH tasks "itemA"   # Prepend to list
RPOP tasks            # Remove last element
LTRIM logs 0 99       # Keep last 100 items

Sets

SADD admins "user1"    # Add member
SINTER group1 group2  # Set intersection
SPOP notifications 5  # Get random members

Hashes

HSET profile name "Alex"   # Set field
HINCRBY profile age 1     # Increment field
HGETALL profile           # Get all fields

Sorted Sets

ZADD rankings 100 "playerA"  # Add with score
ZREVRANGE rankings 0 2       # Top 3 players
ZCOUNT scores 80 100          # Count in range

Specialized Data Types

Geospatial

GEOADD cities 13.4050 52.5200 "Berlin"
GEODIST cities "Berlin" "Paris" km
GEORADIUS cities 15 50 500 km WITHDIST

HyperLogLog

PFADD visitors "192.168.1.1"
PFCOUNT visitors          # Unique count
PFMERGE total v1 v2       # Combine counters

Bitmaps

SETBIT attendance 101 1   # Mark present
GETBIT attendance 101      # Check status
BITCOUNT active_users      # Count active

Persistence Mechenisms

RDB (Snapshotting)

Periodic full dataset snapshots. Configure in redis.conf:

save 900 1    # After 900sec if ≥1 change
save 300 10   # After 300sec if ≥10 changes

AOF (Append-Only File)

Log-based persistence recording write operations:

appendonly yes
appendfsync everysec   # Balance safety/performance

High Availability

Replication

# Slave configuration
replicaof 192.168.1.10 6379
replica-read-only yes

Sentinel Configuration

sentinel monitor primary-cluster 192.168.1.10 6379 2
sentinel down-after-milliseconds primary-cluster 5000
sentinel failover-timeout primary-cluster 180000

Cache Management Strategies

Cache Penetration Mitigation

• Bloom filters for existence checks
• Cache null results with short TTL

Cache Avalanche Prevention

• Staggered expiration times
• Hotspot data never expires
• Pre-warming cache before events

Cache Breakdown Protection

• Distributed mutex locks
• Request coalescing
• Failover mechanisms