Storing unstructured automotive data in the cloud could cost millions

Today’s vehicles are required to capture large amounts of unstructured data, ingested through cameras, lidar, and long-range and short-range radar. Advanced driver-assistance systems (ADAS) need data for features like adaptive cruise control, blind spot detection, emergency braking, and collision avoidance. Level 4 and Level 5 autonomous cars literally depend on data to drive—that is, to drive intelligently and react appropriately.

Auto manufacturing plants are generating more data, too, as they introduce electric vehicles and revamp plant floors to use smart factory technology. Manufacturers use data for visualization, understanding yields, and identifying defects in real time. Production lines are capturing images of welds to ensure quality and images of printed circuit boards to make sure there is no delamination.

Manufacturing engineers and continuous improvement engineers use data to optimize their production and process lines and reduce bottlenecks. Design engineers and developers rely on data to create new cars and applications. Data scientists produce models and test them in simulation environments by using an algorithmic pattern that can later be tested in a real environment.

Automakers also need to hold on to data for compliance reasons—for example, so that insurance companies and transportation boards can review the raw acceleration or stopping data. This data needs to be stored in its original format with flexible levels of retention periods. This storage governance helps authorities determine that the data hasn’t been tampered with.

Now multiply all this data by the number of automotive plants and fleets around the globe. These massive amounts of unstructured data need to be stored somewhere that’s easy to access. And the quickest way to deal with this data is to push it into the public cloud and worry about restructuring later.

“Later” will arrive in 1 or 2 years, when many automakers take cars with ADAS and autonomous systems into the field. They will increase production from 100 to 100,000 cars and start to feel the pressure—particularly the financial pressure—of keeping that data in the public cloud.

The answer to these ballooning costs and high egress charges is to remember that cloud is an operational model. It doesn’t have to involve a hyperscaler like Google, AWS, or Microsoft Azure. You can reproduce the agility and speed of the cloud in your own data center.

For short-term requirements of 3 to 6 months, the public cloud is typically less expensive because there’s no initial investment in hardware and infrastructure. But for longer time frames, an on-premises private cloud can cut storage costs by up to 90%.

A private cloud also delivers more security and control over sensitive data. Particularly in the autonomous vehicle landscape in Silicon Valley, intellectual property is key, and a private cloud keeps your most important assets protected in an environment where there’s significant employee turnover.

Some data might also need to be kept on premises for compliance reasons as ADAS and autonomous functions come into play. Those regulations will be developing over the next few years and need to be considered when data storage is restructured.

NetApp offers several data storage models that address the needs of automakers at different stages of growth and development.

To keep your data on premises and under your control while managing costs, create a private cloud with the NetApp^® StorageGRID^® object-based storage solution. You can choose a capital expenditure (capex) or operational expenditure (opex) model of any size. StorageGRID lets you take advantage of integrated data-management capabilities for machine learning and deep learning development processes, and also for capacity and application requirements. And it helps you optimize your data durability and availability across multiple geographies.

For automakers that do want public cloud storage, NetApp can deliver it with more flexibility. You can switch cloud providers if availability, performance, or cost changes and you need to move to a different service.

Whether you build or buy, on premises or in the public cloud, you can get flexible cloud consumption models with NetApp Keystone.

To learn more, find out how NetApp solutions help speed up data transformation for automakers. Read up on how NetApp Keystone provides a portfolio of payment solutions and storage-as-a-service offerings. And request a test drive of NetApp StorageGRID to see the benefits of centralized storage management.