As part of its cloud transformation initiative, Haier Innovation Center planned to migrate multiple core workloads to AWS, including 2D design, 3D design, graphics rendering, and office automation (OA). These data-intensive workloads rely heavily on efficient data storage and management, making the storage architecture a critical consideration. However, mainstream cloud file services and generic SMB solutions were unable to meet Haier's requirements.
Haier uses Windows-based desktops for large-scale rendering and simulation, with files shared via SMB. Many traditional SMB implementations depend on single gateways or active-standby designs, introducing single points of failure and limited load balancing. In AWS environments, these limitations are further amplified by throughput constraints, making it challenging to support high-concurrency, low-latency access for large numbers of rendering nodes.
Design and rendering workloads generate massive volumes of data, with many compute nodes accessing files concurrently. The storage system must deliver high bandwidth, stability, and continuous availability while remaining cost-efficient. However, generic SMB solutions typically rely on a single storage tier, making it difficult to balance high-performance access with low-cost data retention and backup.
In AWS, standard cloud file services offer limited flexibility in performance tuning, particularly when large compute clusters mount storage concurrently. At the same time, mixed Windows and Linux environments place higher demands on availability, operational continuity, and cross-platform compatibility. A single off-the-shelf cloud service is often insufficient to support the complex, mission-critical workloads of large manufacturing enterprises.
After a comprehensive evaluation, Haier selected TuringData’s cloud-native storage solution tailored for its rendering and simulation workloads on AWS. The TuringData file system delivers highly available, high-concurrency SMB services through a virtual IP address pool and a distributed metadata architecture, enabling multi-node load balancing and rapid failover without disrupting application I/O.
Deeply integrated with AWS infrastructure, the solution leverages cloud elasticity and multi–Availability Zone deployment to provide a stable and scalable storage foundation for mixed Windows and Linux environments. In addition, the solution introduces smart tiering, combining high-performance Amazon EBS as the hot tier with Amazon S3 as the cold tier under a unified namespace that remains fully transparent to end users. This architecture significantly reduces total cost of ownership (TCO) while maintaining high performance for active workloads.
With TuringData, the Haier Innovation Design Center team works efficiently in parallel without worrying about managing data infrastructure, enabling rapid creativity, innovation, and high productivity.
TuringData’s built-in flexibility makes the move to cloud simple and easy from a data perspective. Haier Innovation Design Center was able to deploy a complete TuringData environment in a very short time.
With multi-node load balancing and automatic virtual IP failover, TuringData removes single-point bottlenecks and delivers stable, high-throughput SMB access for Haier’s cloud-based rendering and design workloads, even under heavy concurrency.
TuringData automatically keeps active data on high-performance storage while transparently archiving inactive data to low-cost Amazon S3, reducing long-term cloud storage costs without impacting user access or workflows.
Through distributed parallel access, TuringData fully utilizes underlying cloud block storage bandwidth. Combined with multi–Availability Zone deployment, it provides a flexible and resilient SMB storage platform that supports Haier’s cloud migration and future growth.
Find out how much GPU utilization you're leaving on the table. Schedule a free 30-minute architecture review with our experts.