Question

I am on the process of setting up transactional replication for our SQL server. I plan to use a remote distributor to take off as much load as possible from our main server because performance is key for our business. I know that performance impact depends a lot on the transaction amount, our main database is currently 70 GB and it grows steadily around 3.5 GB per month.

Are there known performance issues when setting up SQL replication using a remote distributor? Would adding RAM or processor power avoid performance impact? How much more (RAM/Processor power) would you recommend?

---- This would be our topology:

Publisher: Main application database server (SQL Server 2008).

Distributor and Subscriber: Remote database server (SQL Server 2008).

Thanks in advance for your help.

Was it helpful?

Solution

Are there known performance issues when setting up SQL replication using a remote distributor?

Yes, because there is an extra network hop when using a remote Distributor. The network can be a potential bottleneck and can affect Publication database log size. Microsoft recommends a fast network of 100 Mbps or faster.

Would adding RAM or processor power avoid performance impact?

Yes, Replication agents can take advantage of processing power and adding RAM to the Distributor will improve performance by allowing it to cache more.

How much more (RAM/Processor power) would you recommend?

This really depends on your workload. I would recommend setting up a test/development environment which mimics your production environment as closely as possible and develop a performance baseline to determine resource consumption under a typical workload. Some other things to look at would be latency, throughput, concurrency, and duration of synchronization.

Here are some resources that cover replication best practices and improving performance:

Best Practices for Replication Administration

Enhance General Replication Performance

I hope this helps.

Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top