AI Server Clustering

Enabling higher bandwidth and lower latency multi-rack connectivity for larger AI clusters.

Requirements and Challenges​

  • Larger clusters of GPUs are required to address the increasing bandwidth and memory demands of AI workloads
  • AI infrastructure must scale GPU clusters across racks as server racks can only accommodate a limited number of GPUs due to power and thermal management constraints
  • At PCIe 5.0 speeds, passive PCIe DAC reach is limited to 3 meters
  • Passive PCIe DACs and general-purpose AECs lack advanced cable and fleet management features essential to managing data center infrastructure​

Aries Smart Cable Module (SCM) Benefits​

  • Aries SCMs extend high-bandwidth PCIe 5.0 and CXL signal reach beyond the rack at 128 GB/s for up to 7 meters to enable larger GPU clusters in a multi-rack architecture
  • Aries SCMs improve cable routing, serviceability, and air flow with thin copper cables and feature low power CMOS process and L1.0 Low-Power Mode to maintain existing rack power and thermal density
  • Aries SCMs offer system-wide visibility and management features with Astera Labs’ COSMOS software that enables enhanced security, quick debug, and flexible firmware upgrade.
  • Aries SCMs are compatible with leading cable assembly suppliers to support multi-sourced cable solutions — enabling a flexible supply chain.
Similar Applications