A user or an application is constantly connected to a data source till the time it is operation on data.

i.e. client directly deals with the data present on server without keeping any local copy of data.

• Since client is getting the data from server directly, hence data is revatively current than in disconnected architecture.
• Since everyone is working on same data in multiuser environment, it easier to manage concurrency conflicts and enforce security constraints.

• Since user needs to be connected all the time, hence this architecture must have a constant network connection.
• It is tough to scale up such applications, demanding huge bandwidths and faster algorithms.

A user or an application is not constantly connected to a data source till the time it is operating on data.

i.e. Client takes stores a copy of data taken from server in local cache. All the operations are done on this local copy and finally changes are made on server in one go with final copy.

• Client works on the local copy after it downloads data from server. Hence it need not to be connected with server all the time. i.e. Network connection is required only for downloading & uploading data.
• Server resources  are free for other users meanwhile.
• Scalability is much easier.

Disadvantages:

• Since user is disconnected after taking data once, hence all updations on server data are not reflected in client's local cache.
• Data locked by one user can restrict access for other users for much longer periods.