By Grant Gustafson | April 28, 2020 |
The development of advanced driver assistance systems (ADAS) - also known as autonomous driving (AD) - relies heavily on massive amounts of real-world training data that consists of data gathered over the course of millions of miles of test driving and thousands of concurrent simulations. To be successful, car manufacturers must be able to simultaneously ingest thousands of concurrent streams of data, apply artificial intelligence (AI) techniques to Petabyte size training data sets, and archive these datasets for decades to come.
It is an enormous undertaking, and with the spotlight on car manufacturers to get this right - if they don’t, lives will be put at risk - the pressure is on to be first to market with a highly reliable solution.
Dell EMC automotive data solutions form the backbone of ADAS development workflows, offering the scalability, performance, and simplified management required to support continuous ADAS data ingest, high bandwidth parallel simulation, and in-place sensor data fusion for AI training sets.
Dell EMC Isilon scale-out NAS delivers the performance and scalability to accelerate time to market and remain competitive in the ADAS space. Already in use today by many of the leading ADAS developers, Isilon’s all-flash performance enables concurrent ingest of vehicle sensor coupled with concurrent, high bandwidth access to data for embedded system development and data-intensive AI solutions.
Dell is a long time partner of Bright Computing to automate end-to-end provisioning, management, and monitoring of compute infrastructure and machine learning workloads – on-premises, at the edge, and in the cloud.
Underpinning Dell’s automotive data solutions is Bright Cluster Manager Auto Scaling Hybrid Cloud, which provides ease-of-use orchestration through Kubernetes and the popular HPC schedulers, and Bright Cluster Manager for Data Science to enable data scientists to become immediately productive by bringing everything together in a simple, easy to deploy and manage solution.
Bright Auto Scaler works on on-premise nodes as well as on cloud nodes. It can auto scale HPC scheduler queues from a pool or a prioritized list of pools. This allows you to run competing workloads on a shared compute infrastructure without statically assigning nodes to queues. Instead, Bright Auto Scaler dynamically assigns nodes to queues based on workload demand and policies that you create and control. And you can do the same thing using Kubernetes namespaces.
Last week, Robert Stober, Bright Computing’s Director Product Management, joined a Dell webinar on this very topic, to explain to the automotive industry how they might adopt a Bright / Dell solution into their ASAD / AD strategy.
You can view the webinar here: https://p.allego.com/vtW101v9uWk4.