We are looking to implement merge replication on our global databases.
The decision that we must make is whether to place all of the tables in
a single publication, to create publications that are grouped by system
modules, or to create a publication for each table. What I am looking
for is best practices or recommendations based on your experience as
far as what approach to take on this. The database contains
approximately 300 tables. Each approach offers advantages and
disadvantages. The biggest disadvantage on the single publication is
flexibility. The biggest disadvantage on each table being a
publication is maintainability . Is there a scalability issue with
either of these approaches? Do all records that have foreign key
constraints have to be in the same publication?
Any guidance is very much appreciated.
Thanks,
Kent
The decision that we must make is whether to place all of the tables in
a single publication, to create publications that are grouped by system
modules, or to create a publication for each table. What I am looking
for is best practices or recommendations based on your experience as
far as what approach to take on this. The database contains
approximately 300 tables. Each approach offers advantages and
disadvantages. The biggest disadvantage on the single publication is
flexibility. The biggest disadvantage on each table being a
publication is maintainability . Is there a scalability issue with
either of these approaches? Do all records that have foreign key
constraints have to be in the same publication?
Any guidance is very much appreciated.
Thanks,
Kent