Does DRS Use Concurrency?

Concurrency is key to performance improvement. DRS uses concurrency in multiple phases. There are two types of concurrency: read/write concurrency and thread concurrency.

Read and Write Concurrency

To migrate data quickly during full migration and ensure a stable connection, DRS extracts data at row-level with concurrency. As a result, the migration speed is maximized and repeated migration of large amounts of data will not occur.
To speed up data writes, DRS supports table-level concurrent writes, greatly improving data transmission speed and achieving millisecond synchronization delay in the same city.
Figure 1 Read and write concurrency

Thread Concurrency

When you migrate a large number of database objects (for example, more than 10,000 tables), the structure migration will become a bottleneck. DRS optimizes concurrency for such scenarios. It uses multiple threads to concurrently query and replay structures, greatly improving structure migration performance.

Figure 2 Thread concurrency