There is no reason why you can't combine two or more of the available architectures in a single application. In fact, some combinations can be extremely powerful. 

For example, you can combine the disk-based architecture described in Using a dedicated file on disk with another approach such as those described in Connecting to another dataset or Using a multi-tiered architecture. These combinations are easy because all models use a client dataset to represent the data that appears in the user interface. The result is called the briefcase model (or sometimes the disconnected model, or mobile computing). 

The briefcase model is useful in a situation such as the following: An onsite company database contains customer contact data that sales representatives can use and update in the field. While onsite, sales representatives download information from the database. Later, they work with it on their laptops as they fly across the country, and even update records at existing or new customer sites. When the sales representatives return onsite, they upload their data changes to the company database for everyone to use. 

When operating on site, the client dataset in a briefcase model application fetches its data from a provider. The client dataset is therefore connected to the database server and can, through the provider, fetch server data and send updates back to the server. Before disconnecting from the provider, the client dataset saves its snapshot of the information to a file on disk. While offsite, the client dataset loads its data from the file, and saves any changes back to that file. Finally, back onsite, the client dataset reconnects to the provider so that it can apply its updates to the database server or refresh its snapshot of the data.