overview

Weka maximizes GPU investments for WeRide by delivering high I/O bandwidth at scale to its GPU Farm.

WeRide focuses on strategic alliances among Artificial Intelligence (AI) technology companies, car makers, and mobility service platform providers. WeRide RoboTaxi, a joint venture company of WeRide, Baiyun Taxi, and SCI (Guangzhou) Group, has a robotaxi program in Guangzhou, China that covers 144.65 square kilometers in the Huangpu and Guangzhou Development Districts, providing the locals with robotaxi services on their daily commute. This is the first robotaxi service that is open to the public in China.

quotes

“We built a GPU farm, and we needed a high-speed data pipe to feed it. We evaluated open source solutions, HDFS, and the public cloud. We chose Weka for its ability to provide cost-effective, high-bandwidth I/O to our GPUs, product maturity, customer references, and stellar on-demand support.”

Paul Liu, Engineering Operations Lead

Challenge

Established in 2017, WeRide is a multi-faceted AI start-up that works on L4 autonomous driving vehicles and partnerships with transportation platform providers that support robotaxi services. This requires IT infrastructure that supports the full AI data pipeline from the edge, to the core, to the cloud. The company processes data at the petabyte (PB) level, ingesting a daily mixture of large video and image files generated from information collected from over 2 million kilometers of driving distance. The data is annotated at the core, trained by the AI model on the cloud-based cluster, and then fed back to the on-premises AI engine. The WeRide team faced the challenge of managing hundreds of terabytes of data with files of mixed sizes through the entire AI pipeline and keeping its GPU farm fully utilized in a cost-efficient manner.

THE NEW SOLUTION HAD TO MEET SEVERAL KEY CRITERIA:
icon3-2

Maximize utilization of the GPU farm

icon-5

Offer a hybrid implementation model to reduce datacenter footprint

Breakthrough Economics

Deliver the best economic value for capacity planning and future performance

icon2

Provide flexibility through hardware-agnostic compatibility with commodity servers

SOLUTION

THE WEKA FILE SYSTEM SOFTWARE ON COMMODITY SERVERS AND AWS

Extensive cost-analysis of various scenarios utilizing open source solutions and HDFS to feed data to the GPU farm proved that all alternative approaches would not to be cost-effective. Ultimately, WeRide chose a hybrid implementation using WekaFS. WekaFS has a two-tier architecture that takes commodity servers and the public cloud and presents them as a single hybrid storage solution. On premises, WeRide manages hundreds of TBs of NVMe flash on Weka, using commodity Intel x86-based servers from AMAX and Mellanox Ethernet network switches. WeRide was also the first customer to use Weka to deliver high-bandwidth I/O to its GPU resources in AWS China. The WeRide team lauds the collaboration with the Weka Technical Support team that made for a seamless public cloud implementation.

Benefits and ROI

WeRide was able to realize several benefits and tremendous return on investment by choosing WekaFS:

Acceleration of product development

icon2
Hardware-agnostic software

Breakthrough Economics
Cost-efficiency

icon3-2
Improved Resource Utilization

icon01
Stellar technical support

Cerence Case Study
Case Studies

Cerence Case Study

Download
Genomics England Case Study
Case Studies

Genomics England Case Study

Download
IDC MarketScape Report
Analyst Report

IDC MarketScape Report

Download

Start Solving the Big Problems

Get Started