I red this article https://technet.microsoft.com/en-us/library/dd263442(v=sql.100).aspx and http://sqlblog.com/blogs/linchi_shea/archive/2010/12/01/linked-server-and-performance-impact-direction-matters.aspx .
1)Is blog valid explanation for push/pull performances? Why is row by row implemented in push?
2)Are same rules of push and pull buffer/pipeline applied all over tsql or those rules are only for replications? What about linked servers?
3)Why is there buffer on push, why is limited on 40k and is it possible to change it?
4)If linked server uses some other rule, can you please tell me which one and why is difference in performance? What is the reason for that rule (if any)?
- Edited by Marko_O 17 hours 23 minutes ago