Admittedly been quite lazy and busy at same time.. Busy with client projects but enough time to blog but lazy not to blog. So, breaking shackles of laziness here is first post of 2012. Started working on an exciting new SQL Server Project. Can share only very skimpy details for obvious reasons.
Problem statement:
There are 350+ databases geographically dispersed in a country and management had decided to get data to a central server. There is not need to merge all 350 databases into a single database but they need it on single server.
Additional Objectives:
- Central Server is just for reporting and there are not updates on Central Server
- Bandwidth will be constrained considering they are geographically dispersed and over WAN Connections.
- Failback Mechanism needed in case primary data replication fails
- BLOB data present
- Process of automatic failback if primary replication fails
Due to so large number of databases challenges are
- What type of data replication should be used that has least impact on
- Network resources
- Server Resources (Central Server should handle Data Replication load and Reporting load)
- What data replication technology to use
- SQL Log Shipping
- SQL Transactional Replication
- SQL Merge Replication
- SAN replicator Software
- Any combination of above technologies
- What is server size needed at Central Location
- To handle any of data replication technologies
- To handle Report Loads
- Defining what is failure and how to recover from a failure or provide an alternative method
Based on these we needed to create Solution Architecture and design of data Replication..
What do you think is appropriate data replication technology, do you think we need to get more requirements in terms of latency before trying different options.
Will keep you all updated how we are progressing..
--Guru