Bigtable -Google's NoSQL database
Bigtable — Easy Notes
1. What is Bigtable?
Bigtable is Google's NoSQL database designed for very large amounts of data.
Think of it like:
Bigtable = NoSQL database + huge data + very fast read/write
It is built for high throughput and consistently low latency.
2. When should you choose Bigtable?
Bigtable is a good choice when:
| Situation | Example |
|---|---|
| Very large data > 1 TB | Billions of customer records |
| Data changes very quickly | Stock prices changing every second |
| High read/write speed required | Millions of requests per second |
| NoSQL data | Data that doesn't need complex table relationships |
| Time-series data | Sensor data over time |
| Real-time processing | Live analytics |
| Big Data | Large-scale data processing |
| Machine Learning | Training/processing large datasets |
Easy example
Suppose you have 10 million IoT sensors.
Every sensor sends:
Sensor ID → Temperature → Time S001 → 28°C → 10:01 S002 → 31°C → 10:01 S003 → 27°C → 10:01 ...
And this happens every second.
That creates an enormous amount of continuously changing data.
👉 Bigtable is a good fit for this type of workload.
3. Bigtable is NoSQL
Unlike a traditional relational database such as MySQL/PostgreSQL, Bigtable doesn't focus on relationships like:
Customer ↓ Orders ↓ Products
Instead, it is optimized for very fast access to huge datasets.
So remember:
SQL database → relationships and transactions
Bigtable → huge scale + fast reads/writes
4. Time-series data
This is an important Bigtable use case.
Time-series data = data collected over time.
For example, stock prices:
10:00 → ₹100 10:01 → ₹102 10:02 → ₹101 10:03 → ₹105
Or server monitoring:
10:00 → CPU 40% 10:01 → CPU 55% 10:02 → CPU 72%
Bigtable works well because the data naturally has time/order.
5. How does data get into Bigtable?
The transcript mentions two main approaches.
Streaming
Data continuously comes into Bigtable:
IoT devices ↓ Dataflow / Spark / Storm ↓ Bigtable
For example, millions of sensors continuously send temperature data.
Batch
Data can also be processed in batches:
Large dataset ↓ Hadoop / Dataflow / Spark ↓ Bigtable
This is useful when you don't need every record processed immediately.
6. How does an application access Bigtable?
An application can read/write Bigtable using APIs and clients.
Simple picture:
Application ↓ Java / HBase Client / REST ↓ Bigtable
For example:
User opens dashboard ↓ Application asks Bigtable ↓ Bigtable returns millions/billions-scale dataset results ↓ Dashboard displays data
7. Very important: Bigtable vs Firestore
Since you just studied Firestore, this distinction is useful:
Firestore:
Flexible NoSQL database for mobile, web and application development
Example:
Instagram-like app ↓ User profile Posts Comments Messages ↓ Firestore
Bigtable:
NoSQL database for massive-scale, high-throughput data
Example:
Millions of sensors ↓ Millions of readings ↓ Bigtable
Remember this shortcut 🧠
Firestore → App data
Bigtable → Big data
Cloud Storage → Files/objects
Cloud SQL → Relational SQL data
BigQuery → Analytics/data warehouse
One-line exam definition
Bigtable is Google's fully managed NoSQL big-data database designed for massive, rapidly changing datasets requiring high throughput and consistently low latency.
Comments
Post a Comment