The RealityServer is a scalable platform, which is component-based. The distributed computing environment is called DICE and is developed for scalability. The multicast based-clustering methodology is unique with self-forming clusters, so no configuration is required. It is a master-less architecture which means that all nodes can be head nodes. RealityServer includes also a configurable fault tolerance.
On top of DICE, a framework has been installed with integrated new rendering modes. There are multiple rendering modes supported, including OpenGL, rasterizer, iray, and sketch renderer.
The native web service provides access to high end renderers through Web Service commands. The code can be in the language of the customer's choice and support of standard protocols is being provided.
The RealityServer offers code in the language of the customer's choice as well. It provides thin-client support for tablets and smart phones, with HTTP connection. There is a standard browser support via Flash or Silverlight and application logic on client side is provided.
The RealityServer has web-based on-line product configurations, such as the car configurator and the furniture configurator. MyDeco.com is the UK's first 3D interior Design Portal which can be customized completely.
The RealityServer is also used as a huge data visualization tool in oil and gas and for cosmology modes. Here, solutions for petabytes are necessary.
For the local industrial design review, the RealityServer is used in Japan by an automotive manufacturer, where it has been integrated in the review pipeline.
Customers are spread all over the world but connected to the RealityServer which is interactive and collaborative. Customers have realtime access anywhere while the RealityServer is running on a cluster in Berlin.
Tom-Michael Tamm introduced 3ds max and RealityServer as the company's latest project. It is a 3D Cloud render service integrated in 3ds max for Autodesk/Amazon/NVIDIA. The single host is equipped with 2 Xenos (2 quad cores) with 22 GB RAM and 2 NVIDIA GPU's. A sufficient number of hosts exist and it is a highly elastic model.