19 September 2014
ESA/ESTEC
CET timezone

NGAPP final presentation

19 Sept 2014, 10:50
1h 30m
Newton 2 (ESA/ESTEC)

Newton 2

ESA/ESTEC

Speakers

Mr Armin Luntzer (Department of Astrophysics, University of Vienna (UVIE)Mr Bernhard Kausl (RUAG Space Austria (RSA))Mr Roland Ottensamer (Department of Astrophysics, University of Vienna (UVIE)

Description

Astronomical science missions in space usually carry a payload of instruments that generate data at high rates. These high data rates often conflict with the available down-link telemetry budget, so that a number of data processing steps will have to be carried out on board by dedicated payload processing units. One of the potential candidate platforms addressing future demands for on-board data processing capabilities was subject to the Next Generation Processing Platform (NGAPP) study. This platform is the Massively Parallel Processing Breadboard (MPPB) architectural concept conceived by Recore Systems, the Netherlands, featuring two of their novel Xentium digital signal processing (DSP) cores in a Network-on-Chip (NoC) infrastructure, which is currently also being investigated by ESA in an effort to integrate it into a space-qualifiable ASIC, dubbed “Scalable Sensor Data Processor” or “SSDP”. The objectives of NGAPP were to evaluate the MPPB with respect to processing performance and capabilities, aiming to evaluate applicability for possible space missions, by subjecting it to software benchmarking and architectural analysis. The presentation will illustrate the flow of the NGAPP study, beginning with an overview of the project's organization by highlighting general background aspects, followed by a selection of potential use cases and necessary science data processing capabilities, along with consequences relating to system requirements. The MPPB and its key features will be presented. In addition, a concept of a lean operating system, conceived as an unexpected output of the study, will be introduced. Finally, the overall system performance will be presented and a set of key points for improvements will be suggested as input to the SSDP with respect to its use as a payload processor for future missions.

Presentation materials