Affordable AI Infrastructure for the Masses
Mike McNamara
With the high complexity and integration costs of off-the-shelf artificial intelligence (AI) and machine learning (ML) solutions, finding a predictable and scalable solution that also minimizes operational expenses for your small to medium-sized business is challenging.
NetApp and Lenovo can help you overcome that challenge. We have partnered to deliver an AI and ML solution that’s efficient and cost-effective. You get an affordable solution that’s focused on an entry-level, clustered architecture with industry leading NetApp® AFF A-Series all-flash arrays and Lenovo’s award-winning ThinkSystem SR670 servers. It’s optimal for small to medium-sized businesses where most compute jobs are single node (single or multi-GPU) or are distributed over a few computational nodes. Most ordinary AI training jobs are single node, making this NetApp and Lenovo solution a superior choice.
Designed specifically for flash, AFF systems help accelerate, manage, and protect your business-critical data. For your NFS workloads, a single entry-level AFF system supports throughput of 5GBps for sequential reads and 148,000 IOPS for small random reads at latencies of 1ms or less. The Lenovo ThinkSystem SR670 server optimizes performance for accelerated AI and for high-performance computing (HPC). Supporting up to either four large or eight small GPUs per 2U node, the SR670 is well suited for your computationally intensive workload requirements.
By using the processing power of GPUs alongside traditional CPUs, the system is designed to handle AI training on your large datasets. Our solution validation has shown that you get high performance and optimal data management with a scale-out architecture that uses either one, two, or four Lenovo SR670 servers with a single NetApp storage system. The configuration that was tested during validation consisted of four SR670 servers, each with four NVIDIA 32GB V100 GPU cards, and an entry-level NetApp AFF all-flash storage system.
Benchmark testing of the NetApp and Lenovo solution demonstrated that NetApp storage provides performance that’s the same as or better than performance from local SSD storage. You also get the following extra benefits:
- Effortless sharing of data between AI systems, analytics, and other critical business systems, reducing your infrastructure overhead and streamlining your data management
- Independent scaling of compute and storage to minimize your costs and to improve your system resource utilization
- Streamlined workflows that use integrated NetApp Snapshot copies and clones for instantaneous and space-efficient user workspaces, integrated version control, and automated deployment
- Enterprise-grade data protection for disaster recovery and business continuity
Are you ready to cost-effectively accelerate and optimize your system performance and prepare for AI workloads? To learn more about the joint NetApp and Lenovo solution, read this technical report and attend this webcast on March 19.
Mike McNamara
次のステップ
.png?width=5760&format=avif&disable=upscale)