380
Performance modeling of the GOES R-series ground segment product generation function

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Wednesday, 20 January 2010
Julie McNeil, NOAA/NESDIS, Greenbelt, MD; and A. J. Ryberg Jr. and J. Gurka

The GOES-R Series provides advanced imaging capability for the development of atmospheric, oceanic and land weather products. Compared to the current GOES imager instruments, the Advanced Baseline Imager (ABI) provides imagery at finer spatial resolutions (0.5 km to 2 km), radiances in 16 spectral channels and faster scan rates (5 minutes for Full Disk). The Ground Segment (GS) will take advantage of the improved imagery data in the production of an enhanced product suite distributed to the user community. A new GS system is being designed, developed and implemented to effectively handle the processing of over 65 unique products, while meeting latency, refresh, and precedence specifications.

This poster provides an overview of the design and application of the Ground Segment Performance Model. The model is a discrete event simulation using COTS software and custom libraries and logical blocks, integrated with Microsoft Excel for data analysis. Based on user-specified input parameters, the model simulates product generation from satellite data downlink through Level 0, Level 1b and L2 product generation across multiple sites. The results include average and peak CPU utilization, data storage levels, and latency timing per product.

The Performance Model provides the capability to independently assess performance of the Ground Segment, characterized in terms of product latency, refresh rates, data throughput, storage size, and resource use (computational, storage, bandwidth). The model also supports studies in the architecture trade-space, including processing capacity, bandwidth sizing, and failover scenarios.