Introduction

This is the first article on a series about OpenDDS. We'll cover many aspects regarding its use for allow different processes to communicate each other.

What's DDS

DDS is an acronym for Data Distribution Service. It's essentially a protocol for exchanging data over a network. It's a standard defined by OMG and it's widely used in distributed application, simulators and so on. It's based on a DCPS (Data-Centric Publish Subscribe) model.

The basic elements of a DDS communication are the following ones:

• Topic: This is a message containing data that can be exchanged between processes. Data is represented like a structure that can contain different data types, like integers, strings and so on;
• Data Writer: This is the component that the process uses in order to send data. The process write data that must be sent into the data writer;
• Data Reader: This is the component that receives the data and make it available to the process;
• Publisher: This is the component that control the network flow of the message, applying i.e. QoS policies (we'll cover it later);
• Subscriber: This is the component that control the input flow.

The main advantage of this architecture are:

• Scalability: A single process can send and receive many topics, and in a simulation network there can be many processes (hundreds, even thousands). There's no need for existing processes to be aware of the existence of new processes. It simply receives and sends data, but who receives its data and who sends data that are received by it, is not defined;
• Flexibility: It's possible to add and remove processes from the simulation without changing anything to existing processes. A process simply sends and receive data over the network. This allows to upgrade the simulation without affecting existing nodes;
• Performance: OpenDDS (and DDS in general) is designed to be very performant. It's widely used for real-time simulations (for example many HLA implementations use DDS as network layer;
• Interoperability: OpenDDS is an implementation of the standard, and is compatible with other commercial DDS implementation. In addition to this, with OpenDDS has libraries for C++ and Java, allowing to exchange data with processes written in different languages.

Conclusions

We have introduces the OpenDDS architecture and some of its features. In next articles we'll see how to use OpenDDS in more details and we'll see some interesting application.

Add comment