Re: Which replication topology?
- From: "glondi" <glondi@xxxxxxxxx>
- Date: Wed, 23 Nov 2005 10:10:46 +0100
Thank you Hillary for your answer! I still have some questions:
"Hilary Cotter" <hilary.cotter@xxxxxxxxx> wrote in message
news:Oc63ja%237FHA.3984@xxxxxxxxxxxxxxxxxxxxxxx
>A central subscriber and transactional replication will work for you. If a
>branch office needs to see data in the central subscriber or branch office
>b use a linked server to connect to the central subscriber of branch office
>b directly. Make sure you use sp_serveroption 'RemoteServerName','Data
>Access', True' to enable this.
would it be a big advatage over suppling rich client application a list of
conncetion strings to other branch offices and central server
so that user can choose server from the supplied list? I see that both
solutions differs in number of connections needed.
I mean that in my solution each client from branch A who want to connect to
branch B server will have its ows connection.
In your solution in this situation those clients will not be directly
connected to branch B server but through their local server using linked
server connection
which will produce only one connection to server of branch B. Am I right?
Has your solution other advatages over mine?
>
> If you need data to flow bi-directionally in the same tables look at
> merge. If its different tables use transactional.
lets say that I resign from making changes at central server and want to
implement one way replication in the same tables.
I mean branches will asynchronously send channges to central server (for the
purpose of data consolidation, reporting).
I will set unique range of identities to each branch. Is transactional
better then merge in this case?
> Regarding identity ranges, a better solution is to have a location
> identifier in conjunction with the identity column and make this your pk.
I want my client application which is used now as a client to embedded
database to scale it to disturbed database topology.
That is why I have chosen manuall allocation of identity ranges ( I dont
want to make changes to database schema and what is more important changes
to client application).
Best regards,
glondi
.
- References:
- Which replication topology?
- From: glondi
- Re: Which replication topology?
- From: Hilary Cotter
- Which replication topology?
- Prev by Date: Re: Which replication topology?
- Next by Date: Re: Logshipping Monitor does not work correctly
- Previous by thread: Re: Which replication topology?
- Index(es):
Relevant Pages
|