Remote Visualisation describes the process of rendering images in real-time based on data, which is not stored locally on the system where you are observing the rendering with the possibility of interaction.
To achieve this two different approaches are common, remote computation where the data is kept remotely but the rendering is performed locally and the actual remote visualisation where data and rendering are performed remotely and the images are transmitted afterwards to the local system.
Use cases and requirements
Remote visualisation provides improved security, since the servers are maintained by security experts and the actual data does not have to leave the secure areas. Administration cost for the end-user is reduced since the remote system and the software is maintained by professional system administrators. Hardware, like storage, CPU and graphics boards are provided as well as a suite of visualisation tools on the remote system. Thus the user has immediate access to the visualisation of the computational simulation results without having to transfer the datasets.
No special requirements are needed on the local system, but accounts to the remote systems either Linux Cluster or SuperMUC are required.
Due to packet loss during the image transfer minor visual artefacts could occur. If the geographical distance is huge or a poor internet connection is used additional latencies might be observed during interaction. The remote visualisation service of LRZ is shared by many users, and the resources are limited in some cases all graphics boards of the remote systems could be booked.
The RVS cluster of SuperMUC Phase 2 contains NVidia Tesla K40m cards.
The RVS clusters of SuperMUC Phase 1 and the Linux Cluster contain NVidia TeslaK20Xm cards.