According to a new press release, Alluxio has unveiled a new high-performance data platform called Alluxio Enterprise AI, aimed at meeting the increasing demands of artificial intelligence (AI) and machine learning (ML) workloads within enterprise data infrastructure. This platform addresses challenges associated with AI infrastructure, such as low performance, data accessibility, GPU scarcity, and underutilized resources. Alluxio Enterprise AI offers significant speed improvements, achieving up to 20 times faster model training and up to 10 times accelerated model serving while utilizing GPUs efficiently.
“Alluxio Enterprise AI provides customers with streamlined solutions for AI and more by enabling enterprises to accelerate AI workloads and maximize value from their data,” said Haoyuan Li, founder and CEO of Alluxio. “The leaders of tomorrow will know how to harness transformative AI and become increasingly data-driven with the newest technology for building and maintaining AI infrastructure for performance, seamless access, and ease of management.”
The key features of Alluxio Enterprise AI include enhanced performance for model training and serving, intelligent distributed caching tailored to AI workload I/O patterns, seamless data access across various environments, and a new distributed system architecture called DORA. This architecture allows for infinite scalability and can handle up to 100 billion objects with commodity storage. Alluxio aims to remove bottlenecks and streamline AI infrastructure, helping organizations optimize their next-generation workloads.
Alluxio Enterprise AI can be integrated into existing AI infrastructure, serving as a bridge between AI compute engines and data lake storage. It seamlessly integrates with popular ML frameworks like PyTorch, Apache Spark, TensorFlow, and Ray, as well as various storage systems, both on-premises and in the cloud. Alluxio’s solutions aim to help organizations achieve higher performance, seamless data access, and easier management of AI model training and serving in machine learning pipelines.
Image used under license from Shutterstock.com