Advancing ADAS Development with Massive Data and Machine Learning
The development of advanced driver assistance systems (ADAS) and autonomous vehicle technology generates a massive amount of data. Auto makers, OEMs, startups, and technology partners need to capture and analyze enormous amounts of unstructured data from test fleets equipped with LiDAR, radar, video, and other sensors.
The industry is at a crossroads. As the amount of data being captured is increasing, the need for capacity and speed is skyrocketing. Data complexity and the need to support hundreds of concurrent users is driving an urgent need for intelligent data management. The ADAS community is challenged to achieve capacity, speed, and intelligence in a manner that’s cost effective.
The following example illustrates the amount of data generated by test vehicles. A typical test fleet vehicle equipped with 3-4 cameras, 3-4 LiDAR sensors, and other sensors will create upwards of 10-20 TB per hour. With even a small test fleet of 10 vehicles, driving 8 hours per day can generate over an exabyte of data in 3-5 years.
Data generated by test vehicles
45GB
of LiDAR data per hour
1.3TB
of Camera data per hour
360MB
of Radar data per hour
Quantum Solutions for ADAS and Mobility
Quantum helps developers of ADAS and autonomous technology capture and cost-effectively manage huge amounts of video and unstructured data with speed and efficiency across complex organizations and workflows.
Higher Performance, Greater Capacity at a Fraction of the Cost
Many developers of ADAS look to traditional networked-attached storage products or public cloud storage for their data needs. These solutions work when dealing with terabytes and petabytes of content, but when dealing with 10+ PBs, 100’s of PBs, and even exabytes of data, these solutions are not practical, technically sufficient, or economical. From a technical perspective, storing and protecting all this content on scale-out NAS can often cost well over $1,000 per TB to store both a primary and secondary copy.
Performance and Security at a 50 – 75% Savings Over Alternatives
With Quantum’s policy-based data management capabilities, content is stored on fast storage during the active period when data is being processed and analyzed, but a backup copy is stored, then archived, on lower-cost storage that can be less than $50 per TB at scale. Because a majority of the data being generated needs to be stored and protected indefinitely, and is seldomly accessed for additional simulations, most of the data can be stored at a lower cost. Many Quantum customers store 80 – 90% of their total data on this low-cost storage tier, enabling them to develop ADAS for a fraction of the cost of competitive solutions.
特色的好处
High-Performance Ingest and Analysis
Complete simulations faster and accelerate ADAS technology development with the ultra-fast parallel processing of StorNext File System.
Cost-Effective Data Storage
Preserve and protect PB-scale unstructured data at a fraction of the cost of competitive storage alternatives.
Integrated Data Protection
Achieve the performance and security of an on-premise infrastructure with capabilities to support hybrid cloud as needed.
Featured Products
Quantum R-Series
Capture data quickly, inside a moving vehicle, without loss or corruption—and automatically ingest that data so you can innovate and develop safer vehicles.
StorNext File System
StorNext®, our award-winning file system, is a combination of a high-speed file system and advanced data management software.
ActiveScale™ Object Storage
See how ActiveScale is helping data-driven enterprises manage, scale, and protect their most valuable data.
Scalar® Tape Storage
The lowest-cost storage for long-term retention and archiving, and to provide an ‘offline’ copy of data to protect against hackers.