TestCenter D2 1.6T Appliance

Accelerating Validation for AI-Scale Networks and Hyperscale Fabrics

Engineered to address the unique challenges of AI networking, this comprehensive 1.6T solution, available in IHS and RHS variants, spans Layer 2 through Layer 3 and beyond, identifying interconnect bottlenecks, latency hotspots, and fault-tolerance gaps that limit GPU utilization and hinder AI cluster scalability.
TestCenter D2 1.6T Appliance

Revolutionary AI Data Center Testing Platform

The VIAVI D2 1.6T validates AI infrastructure without the cost of GPU server farms. It generates up to 6.4Tbps of realistic AI traffic across 1.6T to 100G on a single 4-port platform.

Purpose-built for modern AI data centers, the D2 emulates up to 128 GPUs and NICs per port, reproduces real GPU collective communication patterns, validates congestion control under load, and scales to hyperscale AI fabrics, cutting validation cost and deployment time.

The AI Infrastructure Gap

Traditional testing tools have reached their limit. The D2 1.6T closes the three critical gaps costing AI data centers time and money:

AI Data Center
  • Traffic Authenticity Gap
    Unlike standard tools that miss MPI-based collective patterns (All-to-All, Ring-AllReduce), the D2 emulates real GPU workloads to ensure your fabric is truly AI-ready.
  • Congestion Control Gap
    While most tools lack granular control, the D2 provides hardware-level per-flow and port-level validation for DCQCN and PFC under realistic AI loads.
  • The Scale/Cost Gap
    Move beyond CapEx-heavy GPU server farms. Emulating 128 GPUs per port, the D2 delivers repeatable, production-grade certification scenarios at a fraction of the cost of real GPU clusters.

 

  • AI/ML Workload Generation
    Advanced AI/ML testing and workload traffic generation
  • Flexible interop and conformance testing
    Evolve with industry standards, stay in line with IEEE 802.3dj, IEEE 802.3df, IEEE 802.3ck, IEEE 802.3df 800GBASE-R, UEC 1.0 specs​
  • Robust 800G performance validation
    Confidently validate forwarding performance, latency, MAC capacity, functional capabilities, including Autonomous Flow Tracking (AFT) for massive-scale flow monitoring on 8-port modules.
  • Comprehensive, end-to-end solution
    Test Layer 2, and Layer 3 for routing/switching and data center applications
  • Integrated approach
    Reduce cost of ownership compared to​ solutions that specialize on specific layers​
  • Vendor neutral
    Resolve challenges at every stage, across the ecosystem, no matter the interconnect strategy​

Award-Winning Solution

Recognized in the 2026 Lightwave Innovation Reviews and a Runner-Up in the 2026 Interop Tokyo Best of Show Awards, the VIAVI TestCenter D2 1.6T Appliance sets a new benchmark for AI data center validation, combining high-capacity traffic generation, realistic AI workload emulation, and multivendor interoperability testing to help hyperscalers, cloud providers, and network equipment manufacturers confidently deploy AI-ready networks at scale.

2026 Lightwave Innovation Reviews and a Runner-Up in the 2026 Interop Tokyo Best of Show Awards

Use Cases

Integrated CCL and LLM Traffic Model Emulation

  • AI training performance depends on the network’s ability to handle complex collective communication. The D2 1.6T utilizes the industry’s largest CCL algorithm set and realistic LLM models (ChatGPT, Llama, DeepSeek, RestNet, and AlexNet) to emulate authentic AI workloads. By mixing CCL traffic types and background noise, teams can maximize test coverage and eliminate weeks of custom development and expensive GPU server costs.

Hardware-Accelerated RoCEv2 Validation at Scale

  • Validating RoCEv2 requires massive connection density to reflect production clusters. The D2 delivers unprecedented scale with 16K Queue Pairs and up to 16,000 simultaneous AI training connections per port. This enables precise validation of dual-layer congestion control, including hardware-based per-flow ECN and port-level PFC, across both lossy and lossless fabrics under real-world load.

High-Fidelity AI Fabric Benchmarking

  • Scaling AI infrastructure from 100G to 1.6T requires a unified platform capable of deep workload analysis. The D2 leverages integrated hardware and software to benchmark AI interconnects, identifying bottlenecks and latency hotspots. Through the CCL Wizard, complex AI topologies that previously took days to configure can be set up in minutes, accelerating the transition to high-capacity 1.6T architectures.

    learn more

Next-Generation Ultra Ethernet Readiness

  • As the industry shifts toward open standards, validating Ultra Ethernet Transport (UET) is critical for overcoming traditional scale and efficiency limits. The D2 provides stateful Transport and Link Layer emulation, including a comprehensive UEC conformance test suite. This allows teams to tackle congestion avoidance and verify Link Layer Retry behavior through built-in error insertion to ensure fabric resiliency.

    learn more

VIAVI D2 Advantage: Hyperscale Validation Without Compromise

While others test components, we validate the entire ecosystem at production scale.

  • Industry-Leading Scale: Support up to 1,024 ports with built-in synchronization, creating test environments 32× larger than competing solutions without external timing gear.
  • True Mixed-Speed Validation: Available in IHS and RHS variants, the D2 validates your entire 224G and 112G PAM4 roadmap, from 1.6T spines to 100G legacy ports.
  • Deterministic Lossless Metrics: Deep hardware-level PFC and congestion measurement, ensuring zero-packet-loss performance for mandatory AI workloads. 
  • AI-Driven Test Automation: TestCenter MCP Server, part of the VIAVI NITRO® AI portfolio, configures, executes, and analyzes D2 tests through natural language — no scripting or API expertise required.

Blog

Brochures

Data Sheets

Video

White Papers & Books

Support at Every Step

We provide support, services, comprehensive training and the resources you need. It’s all part of what we do to maximize the value of your VIAVI investment.

Ask an Expert

Contact us for more information or to receive a price quote. We have the experts to give you the right answer on any of your questions.