A new concept introduced in SQL 2016. The concept is indeed good but does it really fly with DBA’s likewise business users. The concept aims to introduce local server performance for hot data and cloud storage for old data without changing any of your current application. That is all good on paper but what are the caveats and limitations. The main three limitations that would probably stop me from using the feature are:
- Uniqueness is not enforced for a UNIQUE and primary key constraints
- You will not be able to update likewise delete records on stretch-enabled table
- You will not be able to run create index or alter index operations on a stretch-enabled tables
- You will not be able to run ALTER Table operations
- No support for some data types (timestamp, sql_variant, xml, geometry, geography, hierarchyid, UDT
- In addition to the above limitations , stretch tables don’t support some features such as computed columns, check constraints, foreign key constraints, spatial indexes, CLI (clustered columnstore indexes, xml indexes and etc. To view the full list of limitation please visit https://msdn.microsoft.com/en-us/library/dn935016.aspx
I am hoping most of the above will be supported before the actual product release. Otherwise don’t see any benefit of using it as it stands.
No comments:
Post a Comment