In the ATC
Liqid Composable Disaggregated Infrastructure Lab
Labs in development
GPU lab
The GPU lab is designed with different models of GPUs to allow engineers and data scientists to evaluate GPUs from multiple ODMs. Engineers and data scientists can also develop and test AI models from an inferencing and machine learning standpoint.
SR-IOV lab
The SR-IOV lab allows users to install a type-1 hypervisor on a host or cluster to validate the ability to add components to the hosts and map the devices directly to virtual machines. This lab can also be used to show performance gains of SR-IOV vs. paravirtualization. Users can also test adding, removing and moving components from host to host, all via software.
DPU, IPU and SmartNIC lab
The DPU, IPU and SmartNIC lab is designed with solutions from the major ODMs in this market segment. The primary use case for this lab is to evaluate device/host/OS integration and the ability of the cards to support several different VNF solutions, such as content load balancer, SD-WAN and next-generation firewall solutions.
Application/ISV Validation lab
The Application/ISV Validation lab allows users to test different applications that can benefit from running in a highly scalable environment. The environment will enable users to add GPUs, NVMe storage, high-speed NICs and FPGAs as needed all via software. Impossible configurations can be built where a server with just two or three PCIe slots can support 10 or more PCI devices, providing the most efficient design possible.
Big Memory lab
The Big Memory lab showcases the integration of Dell PowerEdge servers, Intel Optane NVMe SSDs, Liqid Composable Disaggregated Infrastructure and MemVerge Big Memory Machine, allowing users to build configurations that standard servers cannot support due to physical DIMM slot capacity. Users can create an environment that uses local DRAM for cache memory and Intel Optane NVMe SSD as capacity memory to scale a Big Memory solution past 8PB of capacity. Users can then test in-memory databases, run in-memory analytics tests and even evaluate in-memory computing solutions.
Component comparisons and upgrade testing
The Component Comparison lab allows users to test and compare performance gains between different ODM solutions, such as GPU, while keeping the OS, application and server to stay the same. Users can install and configure the OS and application once and test different product releases from an ODM.