Rx 1.0 Solves the Problem of Asynchronous Data Access
Rx tries to solve the problem of asynchronously accessing data from an event-based UI. The standard Iterator pattern with its basic interfaces IEnumerable and IEnumerator is not adequate for asynchronous operations, so Rx solved the problem by introducing the Observer pattern, having two main interfaces, IObservable and IObserver. Instead of a client iterating step by step over a data collection, the collection pushes data to the client as a result to an asynchronous call, closing the call’s loop.
This scenario is especially useful in cloud computing, according to Erik Meijer, the leader of the Rx project. Cloud computing involves asynchronous calls due to inherent network latencies, so for a client is much better to ask for data, then continue with other tasks and receive the data when it becomes available. While this problem seems simple to be solved using more standard ways, the power of Rx comes from its embedded support for LINQ and concurrency. So, the user not just gets a piece of data in an asynchronous way, be he can generate asynchronous LINQ queries over collections of data and run them concurrently. Rx deals with concurrency by being built on top of Parallel Extensions (PFX), a managed concurrency library including Parallel LINQ and Task Parallel Library.
Tom Gilb & Kai Gilb Jan 26, 2015