use case : synchronization/Replication across resources

There is a large class of distributed applications which are intermittently
connected, for a variety of reasons. Such systems often need to deploy some
form of a synchronization or replication mechanism.  

Since every such system may have different requirements from the point of
view of hardware, deployment platform and communication protocols, it's not
possible to create a framework fully encompassing all such needs. However,
it would help to have a basic protocol for synchronization/replication for
an 'abstract' or general distributed computing framework.  

One might argue that such protocols could be implemented using the existing
use cases in XP. Well, that's what people are doing now anyway.  Even with
the advent of 3G (and 4G in future), it's still hard to imagine a world with
fully connected nodes.  In contrast, the information distribution is going
to increase with time, resulting in a need for a distributed
synchronization.
 
Such an abstract framework might incorporate various patterns like periodic
sync, event based or on-demand sync, partial/complete sync of the resources.

Regards,
Shailesh Karandikar,
Dendrite International (R&D)

Received on Tuesday, 29 January 2002 15:20:28 UTC