You can use the Salesforce standard or Bulk API to process Salesforce data in a replication task.
Use the standard API to process a normal amount of Salesforce data with standard reporting on the results of the standard API load.
Use the Bulk API to process large amounts of Salesforce data while generating a minimal number of API calls.
With the Salesforce Bulk API, each batch of data can contain up to 10,000 rows or one million characters of data in CSV format. When the replication task creates a batch, it adds any required characters to properly format the data, such as adding quotation marks around text.
When you use the Bulk API, enable primary key chunking to increase performance when you extract data from large tables. When you enable primary key chunking, the Bulk API splits the data set into chunks based on the record ID and creates extract queries for each chunk. The Bulk API combines the data when all the extract queries are complete. Salesforce supports primary key chunking for custom objects and certain standard objects.