Sizing Up Voyager’s 5G Network: NVIDIA's Time and Cost Savings with Altair 5G Wireless Network Solution

• Networks & Connectivity - 5G
• Networks & Connectivity - RFID

• Electrical Grids
• Telecommunications

• Product Research & Development
• Quality Assurance

• Building Automation & Control
• Experimentation Automation

• System Integration
• Testing & Certification

NVIDIA, a pioneer in accelerated computing, built a massive 750,000 sq. ft. building named Voyager. To accompany the architectural innovation, NVIDIA wanted an equally impressive, private 5G network to support multi-access edge computing (MEC) applications and leverage the unlicensed Citizens Broadband Radio Service (CBRS) band. The first MEC application required intelligent video analytics with 5G cameras in the lobby area. A network development challenge was the 150 MHz limit within the CBRS spectrum. To handle this, NVIDIA decided to use 100 MHz minimum bandwidth to maintain the desired throughput levels and use the same frequency carrier for all radio units. This made the 5G network’s needed throughput challenging. NVIDIA also wanted to compare two different vendor radio units, one with directional transmission and one with omni-directional transmission, each with 4 downlink (DL) multiple-input multiple-output (MIMO) layers and 2 uplink (UL) MIMO layers.

NVIDIA, founded in 1993, has been a pioneer in accelerated computing. The company’s invention of the GPU in 1999 sparked the growth of the PC gaming market, redefined computer graphics, and ignited the era of modern artificial intelligence (AI). NVIDIA is now a full-stack computing company with data-center-scale offerings that are reshaping the industry. NVIDIA built a massive 750,000 sq. ft. building named Voyager, which is a reference to both Star Trek’s Voyager and the “v” in NVIDIA. The base camp reception area sits at the foot of a mountain and features numerous tiers interspersed with people, offices, and garden spaces.

NVIDIA wanted to maximize throughput for the desired in-building coverage area. However, traditional testing and physical deployment to determine the best radio-unit locations and antenna orientation were costly and time-consuming. Instead, NVIDIA deployed Altair’s simulation-driven, network planning and optimization solution. Altair developed a Voyager surrogate model including all aspects that affected the building’s wireless propagation. In an integrated system workflow, the team used Altair® HyperWorks® for geometry cleanup; Altair® Feko® with Altair® WinProp™ modules for material assignment, wireless propagation, and network analysis; and Altair® HyperStudy® for optimizing the number and positions of radio units. Altair’s initial analysis identified the best radio unit locations and orientations using WinProp’s dominant path model, a fast, accurate prediction model for wireless coverage analysis. With the radio units optimally positioned, Altair performed the final analysis with WinProp’s full-3D-ray-tracing models.

• With its integrated simulation and optimization platform, Altair was able to quickly explore numerous design iterations. With these, Altair determined the optimum number of radio units that would produce the desired throughput throughout the lobby—especially in locations for mounting cameras— and compared the performance of the two different vendor radio units. Both units operated over the same carrier frequency, which caused co-channel interference that Altair minimized by optimally positioning the units. The solution provided NVIDIA with an optimal 5G network design and eliminated time and costs associated with physical testing. Looking forward, NVIDIA will use Altair’s solution to investigate network deployment topologies and system parameters that address requirements for other MEC use cases.

• Altair was able to quickly explore numerous design iterations.
• Altair determined the optimum number of radio units that would produce the desired throughput throughout the lobby.
• The solution provided NVIDIA with an optimal 5G network design.