GigaIO Introduces the Industry’s First PCIe 4.0 Fully Managed Accelerator Pooling Appliance for the Composable Data Center
Pool GPUs and FPGAs, and compose them at native PCIe latency for the ultimate in
data center agility, performance, and cost savings
GigaIO Networks announced their new Hydra product line, the industry’s first fully managed PCI Express (PCIe) Gen 4.0 Pooling Appliance, a high-performance expansion chassis for the disaggregating and pooling of PCIe accelerator devices. Using the off-the-shelf Ready-To-Run composition software of their choice, data center managers can transform previously static resource silos into elastic, shareable resource pools, and liberate expensive stranded resources such as accelerators from the legacy servers imprisoning them. The resulting software-defined infrastructure delivers cloud-like flexibility and agility, while drastically slashing TCO (Total Cost of Ownership) and delivering blazing performance to their users. GigaIO’s open standards approach to disaggregation and composability with Redfish APIs and off-the-shelf software delivers freedom from vendor lock-in and removes a significant barrier to widespread adoption of composable infrastructure.
Disaggregated Composable Infrastructure (DCI), made possible by-products such as Hydra, removes the bottlenecks to making changes to legacy IT infrastructure which prevents businesses from entering new market opportunities. It also increases financial efficacy by balancing CapEx and OpEx spending and facilitates the deployment of AI and advanced analytics to optimize customer value.
With its composability and utilization optimization capabilities, this accelerator appliance advances GigaIO’s vision for composable technology as the future of the software-defined data center architecture.
The GigaIO Accelerator Pooling Appliance is the industry’s highest-performing PCIe accelerator expansion chassis fully supporting the PCIe 4.0 specification with up to 1Tb/sec bandwidth into and out of the box. It delivers advanced provisioning and monitoring of computing accelerators and next-gen NVMe assets, supporting up to 8 double-width PCIe 4.0 x16 accelerator cards and 4 PCIe 4.0 x16 low-profile cards. This flexible expansion platform enables users to add any PCIe 4.0 compliant application accelerators, GPUs (including the new NVIDIA A100), FPGAs, DPUs, IPUs, thin-NVMe-servers, and specialty AI chips.
GigaIO’s advanced Redfish® API interface provides an industry-first capability with management software that enables appliance provisioning as well as discovery, inventory, port configuration, diagnostics, monitoring, fault detection, utilization auditing, and performance monitoring. Administrators can provision devices, redeploy them and configure PCIe ports in seconds without service interruptions. Hydra also enables centralized analytics such as device utilization and performance and provides a fault summary by continuously monitoring errors to quickly resolve problems and improve reliability.
“Our customers require solutions that provide greater flexibility, performance, and cost savings and once again GigaIO delivers,” says Niraj Mathur, VP Product Management at GigaIO. “Our leadership in delivering dynamically composable solutions for AI and HPC workloads in real-time, using industry-standard PCI Express technology, enable us to deliver solutions based on open industry standards.”