****🗓️ 01112024 2102
📎 #data_processing
online_transactional_processing
TLDR
Type of data processing for large number of concurrent transactions
Typically user facing
Access Pattern
Usually only touch a small number of records in one query
- Application requests records using some kind of key, and the storage engine uses an index to find the data for the requested key
- Disk seek time is often the bottleneck
Requirements
| Requirement | Description |
|---|---|
| ACID compliant | Refer to internal link |
| Concurrent | Able to handle large amounts of concurrent operations |
| Scale | Must be able to scale up / down to meet demand irrespective of volume |
| Availability | Must always be available / always ready to accept transactions |
| High throughput and short response time | Nanoseconds / even shorter response times |
| Reliability | Typically read / manipulate selective & small amounts of data |
| Security | Since OLTP can store sensitive customer information |
| Recoverability | Must have the ability to |
Relational DBs were built specifically for transaction applications