Nyriad has discovered a new way to store, process and transport information

Nyriad has discovered a new way to store, process and transport information

Nyriad is a global software technology company headquartered in Cambridge New Zealand. Nyriad is officially the first commercial spin-out from the Square Kilometre Array, a multi-billion dollar radio telescope and the world's largest computing project, being constructed by a consortium of 10 nations across Australia and South Africa.

SKA1 will capture approximately 160TB per second of radio data. This needs to be continuously processed and stored by two of the world's largest supercomputers. The sheer scale of the data and the computational requirements to process it led Nyriad to a breakthrough in computing density where parallel computing and distributed storage-processing become the same thing.

SKA Organisation

In 2016, Nyriad partnered with Australia's International Centre for Radio Astronomy Research to adopt Nyriad's technology for Radio Astronomy and developed the Radio Astronomy Block Device (RABiD), which is now being deployed at the Murchison Widefield Array in Western Australia.

Nyriad is working closely with ICRAR on RABiD 3.0 which will integrate with the SKA-precursor data processing framework, DaLiuGE, and enable the resilient management, storage and processing of Petabytes of astronomical data across an international network of over a dozen cloud partners and participating supercomputers.

The technology is now available commercially as NSULATE to our OEM and cloud service partners. Our NZ engineering and international business development teams will be built up rapidly as we grow our business.

Accelerating Science and Discovery
The role of GPUs

10 years ago, GPU manufacturers started to generalise their chip architectures to support general purpose processing, turning video cards into miniature supercomputers.

Today, GPUs are ubiquitous in modern supercomputing and leading the transition to faster, more efficient exascale computing solutions. As computing power increases, network and storage IO bottlenecks have come to dominate HPC costs.

Nyriad specialises in the use of GPUs for converging computing and IO to minimize data movement during processing which can reduce HPC power consumption by 70%-90%.

When you don't need to move the data, bandwidth is infinite.