You do not need to know how the OData protocol is defined and implemented to access data though an OData web service. However, the following brief description may help you understand conceptually what’s going on. The OData protocol follows the principles of Representational State Transfer (REST), meaning that client requests are independent of each other; no session state or data is stored on the server between client requests. Instead, all necessary state is transferred from client to server along with each client request. In the case of OData, client requests include create, read, update, and delete (CRUD) operations. Individual requests from clients are HTTP methods, which include GET, PUT, POST, MERGE, and DELETE, the contents of which specify the details of what is being requested. The server responds to those requests over HTTP.
OData data consists of feeds that are collections of entries. For a database, a feed might represent a table and the entries would be the individual records in that table. An OData web service may also contain a service document that lists the available feeds, service operations that are functions available on that service, and a metadata document that describes all relevant information about the service. All of the resources made available by an OData service are described in the metadata document.