"crbd98@yahoo.com" wrote:

> Hello All,
>
> I know that this problem is not SQL Server specific but I decided to
> post it here because it is a problem that I am certain many of you
> expericenced (or at least thought about).
>
> In my system, I have a central SQL Server Database and number of
> remote "data providers". Some of these data providers are other SQL
> Servers and others are other databases or other data provider services
> that expose a database-like interface.
>
> We provide two levels of search capability in the system. The user can
> search either the central database, or search one specific data
> provider. One of the new requirements, was to provide a global search,
> which would search the central database and all the data providers at
> once.
>
> Performing a distributed is prohibitive in our scenario because we
> have many data providers and they remotelly located. To minimize the
> search time, my first reaction was to create a copy of the data from
> the remote providers on the central database and keep it in synch with
> the data in the remote data providers. By doing this, I would expect
> to simplify the search by having simply a search in the central
> database.
> However, a number of new question concerning data synchronization
> arise:
> 1. How can I keep the data in the central db in synch with the remote
> data providers? (If they all were SQL Server databases, I would use
> merger replication...but they aren't?)
> 2. How to handle disconnects? If the remote provider disconnets and
> reconnects, some of the data might be stale. In pronciple, I could
> take a new snapshot and keep the data in synch from that point.
> However, this is very expensive!! Do you know of any other techniques
> for data synchronization that would minimize network traffic?
> 3. Is this a good approach? What do you think?
>
> Your opinion is gratly appreciated.
>
> Kind regards
> CD
>