ECMWF’s Extreme Data Challenges towards a Exascale Weather Forecasting System

Tiago Quintino (ECWMF) gave this presentation at PASC19, Zürich, Switzerland (12-14 June 2019).

Abstract: ECMWF's operational weather forecast generates massive I/O in short bursts, currently approaching 100 TiB per day, in two hour-long windows. From this output, millions of user-defined daily products are generated and disseminated to member states and commercial clients all over the world. As ECMWF aims to achieve Exascale NWP by 2025, we expect to handle around 1 PiB of model data per day and generate 100's of millions daily products. This poses a strong challenge to a complex workflow that is already facing I/O bottlenecks. To help tackle this challenge, ECMWF is developing multiple solutions and changes to its workflows, and incrementally bringing them into operations. For example, it has developed a high-performance distributed object-store that manages the model output, for the needs of our NWP and Climate simulations, making data available via scientific meaningful requests, which integrate seamlessly with the rest of the operational workflow. We will present how ECMWF is leveraging this and other technologies to address current performance issues in our operations, while at the same time preparing for technology changes in the hardware and system landscape and the convergence between HPC and Cloud provisioning.

(Authors: Tiago Quintino, Simon Smart, James Hawkes, Baudouin Raoult, all ECMWF)