IBM Unveils ESS 3500 Software Defined Storage Server For AI

ESS 3500 GBL

Pointers at Glance

  • IBM launches Enterprise Storage Server with a faster controller CPU and maximum throughput.
  • The ESS has been designed to work with Nvidia’s DGX-dense compute servers for training AI.

IBM has included a new member to its Spectrum Scale Enterprise Storage Server (ESS) portfolio that features a faster controller CPU and more throughput. The ESS is designed to work with Nvidia’s DGX-dense compute servers for training AI.

ESS 3500 Details

The new ESS 3500 is a 2U design with 24 drive bays and a maximum raw capacity of 368 TB. But it can achieve up to 1 PB through LZ4 compression, a first for the series that earlier ESS versions do not have. The ESS 3500 can achieve up to 91GB/s of throughput performance, comparatively better than the 80GB/s of the older models.

IT runs Spectrum Scale, IBM’s scale-out parallel file system that spans on-premises, cloud, and edge networks. It uses dual active controllers with either 100 Gbit Ethernet or 200 Gbit HDR InfiniBand ports and a 48-core AMD Epyc processor on each controller.

The ESS 3500 targets Nvidia’s DGX dense compute systems: all GPUs and memory but no storage. It uses Nvidia’s GPUDirect Storage technology, which creates a direct data path between GPUs and storage through NVMe or NVMe over Fabrics (NVMe-oF).

Generally, data needs to be loaded into the CPU and main memory before being moved to the GPU for processing. GPUDirect allows the system to bypass the CPU and main memory entirely and provides a direct connection between storage and GPU memory.

IBM says that with this system, auto parts maker Continental was able to improve AI training time for self-driving vehicles by as much as 70% using IBM Spectrum Scale and IBM ESS 3500 with a DGX system. The ESS 3500 is available now.

Read here: The 10 Hottest Data Storage Startups of 2022

Skip to content