Unlocking MongoDB: Key Features, Deployment Architectures, and Sharding Strategies

MongoDB Overview

MongoDB is a document‑oriented distributed database designed for high‑performance, scalable storage. Its ease of deployment and flexible schema have made it popular among developers.

Key Features

High Performance : Single‑node benchmark exceeds 100k TPS.

High Availability : Automatic election of a new primary when the primary fails.

Consistency : Supports fsync, journaling, and write concerns (w, j, wtimeout) to guarantee data consistency; rollback files preserve divergent writes.

Compression : Snappy and zlib compression achieve 2‑7× size reduction.

Distributed : Horizontal scaling of capacity and read/write throughput.

Document Model : JSON‑like documents with flexible fields; default document‑level lock since version 3.0.

Deployment Architectures

Standalone

A single mongod process provides the service. It consumes minimal resources and is easy to manage, but recovery from failures is costly.

Replica Set

Multiple mongod instances store identical data, providing redundancy and automatic failover. Primary and secondary nodes hold data; an arbiter participates only in elections.

Sharded Cluster

mongos routers distribute queries to shards, each shard typically being a replica set. Config servers store metadata. This architecture enables horizontal scaling for large data volumes and high throughput.

Sharding Strategies

Hash Sharding

Uses a hash of a single field as the shard key, distributing data evenly; suitable for equality queries.

Range Sharding

Divides data into contiguous ranges (chunks) based on the shard key, ideal for range queries but can create hotspots if the key is monotonic.

Zone Sharding

Assigns specific key ranges to designated shards, allowing data placement based on hardware differences or geographic proximity.

Data Balancing

The balancer monitors chunk distribution and moves chunks between shards to maintain balance. Its operation is transparent to applications; a time window can be configured to run only during low‑traffic periods.

db.getSiblingDB('config').settings.update(
  { _id: "balancer" },
  { $set: { activeWindow: { start: "<start-time>", stop: "<stop-time>" } } },
  { upsert: true }
)

QConf Overview

QConf is Qihoo 360’s ZooKeeper‑based configuration service that replaces traditional config files. Applications retrieve configuration from QConf to connect directly to clusters.

QConf vs. LVS

Advantages : Application transparency, configuration caching, direct connections.

Disadvantages : Requires deployment of an agent on business machines.

Recommendation: Use QConf for high‑concurrency, long‑connection services; for environments where agents cannot be installed, consider VIP with LVS.