Before an autonomous vehicle can be road ready, manufacturers are required to train their Advanced Driver Assistance System (ADAS) with an enormous amount of real environment data. All of which needs to be collected beforehand via large fleets of sensor-equipped test vehicles over many journeys. To make these masses of data accessible to all relevant stakeholders, it must be uploaded to ever-present storage and cloud infrastructure. As high data quality is essential for efficiently training the ADAS, the data is ideally pre-processed by intelligent software algorithms running on high-performance servers.
But how and where can those tasks be handled? The only place where these huge amounts of data can be ingested, processed and uploaded to the cloud with reasonable speed is a colocation data center, because it provides the necessary highly secure on-premise environments with direct connections to the Points of Presence (PoP) of all major public clouds.
The process of collecting, transferring, processing and uploading data is the essential use case for every developer of an ADAS. Taking an end to end approach to building up the complex chain of technologies and solutions, the NTT Technology Experience Lab has brought together five highly experienced technology partners into the Data Garage project.