VIAVI TestCenter B3-800-OSFP-8-1200A 8-Port OSFP 800G/400G Appliance Bundle
VIAVI B3 8-port OSFP 800G test appliance bundle with 800G and 400G speeds
- Default time source
- Stratum-3 rated oscillator
- Transmit line clock
- ± < 1 PPM
- Long-term clock accuracy
- ± 4.6 PPM
- Frame time-stamp resolution
- 2.5ns
Overview
The VIAVI TestCenter B3 800G appliances are QSFP-DD800 and OSFP800 test platforms. They validate hyperscale data centers, AI clusters and high-performance Ethernet fabrics from Layer 1 through Layer 3. The family comes in 4-port and 8-port configurations and generates up to 6.4 Tbps of test traffic.
In 112 Gbps PAM4 mode, the appliances support 1x800G, 2x400G, 4x200G and 8x100G, in line with Ethernet Technology Consortium 800GBASE-R and IEEE 802.3ck and 802.3df 800GBASE-R. In 56 Gbps PAM4 mode they support 1x400G, 2x200G and 8x50G per IEEE 802.3cd and IEEE 802.3bs. Auto Negotiation and Link Training (AN/LT) follow IEEE 802.3df for all supported speed modes. Smart Port Technology is a licensed feature that allows single-port and speed upgrades.
AI infrastructure testing
For AI data center work, the B3 emulates GPUs and NICs using full-stack RoCEv2 transport and Ultra Ethernet Transport (UET). It supports Collective Communication Library (CCL) traffic patterns and LLM training traffic emulation. UEC 1.0 support covers Transport Layer and Link Layer emulation, plus a conformance test suite with over 50 built-in test cases. Autonomous Flow Tracking (AFT) on 8-port models monitors very large numbers of flows, for example in BGP route import scenarios.
Software and automation
The appliances run under VIAVI TestCenter. Automation is available through Command Sequencer visual programming, GUI to Script, and a REST API. The TestCenter Model Context Protocol (MCP) Server is part of the VIAVI NITRO AI portfolio. It lets AI agents configure, execute, monitor and analyze tests through natural language. XStream mode raises traffic generation and protocol scale.
Key Features
- Flexible port speeds supporting 1×800G, 2×400G, 4×200G, 1×400G, 2×200G, 8×100G and 8×50G PAM4
- RS-FEC, Auto Negotiation (AN) and Link Training (LT) per IEEE 802.3df
- AI infrastructure testing for RoCEv2, CCL and Ultra Ethernet Transport (UEC 1.0), including transport and link-layer LLR testing
- L2/L3 routing and switching and data center protocol testing
- Autonomous Flow Tracking (AFT) on 8-port modules
- TestCenter MCP Server for natural-language test configuration, execution and analysis
- QSFP-DD and OSFP test platform for high-scale testing
- Layer 1 debug including RS-FEC performance analysis and interconnect monitoring
- Support for optical transceivers, AOC, passive copper (DAC) and active electrical copper cables (AEC)
- Up to 6.4 Tbps of test traffic
- Smart Port Technology licensed single port and speed upgrades
- XStream mode for increased traffic generation and protocol scale
- Self-calibrating inter-chassis timing chain with daisy chaining of up to 255 appliances
- Command Sequencer visual programming and GUI to Script automation
- UEC 1.0 conformance test suite with over 50 built-in test cases
Specifications
| Default time source | Stratum-3 rated oscillator |
|---|---|
| Transmit line clock |
± < 1 PPM
At the precise nominal Ethernet rate on initial shipment
|
| Long-term clock accuracy |
± 4.6 PPM
15 years of operation
|
| Frame time-stamp resolution | 2.5ns |
| External time sources | GPS and CDMA-based; IEEE 1588v2 and NTP packet-based; TIA/EIA-95B-based |
| Appliance time synchronization |
± 20ns
VIAVI-patented self-calibrating inter-chassis timing chain using dedicated port on appliance
|
| Daisy chain | Up to 255 appliances |
| Operating temperature range |
41° F to 86° F (5° to 30° C)
Ambient temperature
|
| Relative humidity | 20% to 80% |
| Max power draw | Maximum 2100W at 100-240 VAC |
| Product dimensions | 29.5 in D x 17.1 in W x 3.5 in H (74.9 cm x 43.4 cm x 8.9 cm) |
| MSA interface | OSFP800 |
| Bundle | B3 8-Port OSFP 800G/400G Bundle |
| Unit installed weight |
57 lb. (25.85 kg)
8-port appliance
|
| Speed modes |
1x800G, 2x400G, 4x200G, and 8x100G in 112Gbps PAM4 modes; 1x400G, 2x200G and 8x50G in 56Gbps PAM4 mode
Speeds available depend on the purchased bundle and speed options
|
|---|---|
| QSFP/OSFP interconnects | CR, SR, LR, FR, DR, ZR at multi-rate (800/400/200/100/50G) |
| FEC | RS (544,514) FEC supported for all PAM4 speed modes |
| Optical transceivers, 112 Gbps PAM4 mode | 1x800G: 800GBASE-SR8, 800GBASE-DR8, and 800GBASE-FR8; 2x400G: 800GBASE-SR8, 800GBASE-2FR4, 400GBASE-DR4, 400GBASE-FR4; 4x200G: 800GBASE-SR8; 8x100G: 800GBASE-SR8 |
| Copper cable, 112 Gbps PAM4 mode |
1x800G/2x400G/4x200G/8x100G: 800GBASE-CR8
Copper cable types include Direct Attached Copper Cable (DAC), Active Electrical Cable (AEC), and breakout cable
|
| Optical transceivers, 56 Gbps PAM4 mode | 1x400G: 400GBASE-SR8, 400GBASE-DR4, 400GBASE-FR4, 400GBASE-LR4, 400GBASE-LR8, 400GBASE-ZR, and 400GBASE-ZR+; 2x200G: 400GBASE-SR8 and 200GBASE-FR4; 8x50G: 400GBASE-SR8 |
| Copper cable, 56 Gbps PAM4 mode |
1x400G/2x200G/8x50G: 400BASE-CR8, 200GBASE-CR4
Copper cable types include Direct Attached Copper Cable (DAC), Active Electrical Cable (AEC), and breakout cable
|
| Auto-Negotiation/Link Training (AN/LT) |
AN/LT supported for 1x800G, 2x400G, 4x200G, 8x100G, 1x400G, 2x200G, and 8x50G
IEEE 802.3 compliant
|
| Layer-1 debug tools and features | Pre/Post FEC Codeword statistics, Tx Emphasis settings, Rx Eye view, FEC Counters, PRBS Gen/Check, Front-end L1 Summary Status, Xcvr MDIO access, PCS monitoring |
| PRBS patterns | PRBS-9Q, PRBS-10Q, PRBS-11Q, PRBS-13Q, PRBS-15Q, PRBS-20Q, PRBS-23Q, PRBS-31Q, PRBS-49Q, PRBS-58Q |
| L1 FEC alarms | High Symbol Error Rate, Loss of Alignment Marker Payload Sequence, Loss of Alignment |
| Number of streams, Stats/Streams (Tx/Rx), default mode | 800G (4K/32K), 400G (16K/32K), 200G (8K/16K), 100G (4K/8K), 50G (2K/4K) |
|---|---|
| Number of streams, Stats/Streams (Tx/Rx), XStream mode |
32K/32K for 800G, 400G, 200G, and 100G
Enabled by the XStream mode
|
| Number of paths/raw streamblocks | 127 (800/50G), 511 (400/200G), 255 (100G); 511 (800/400/200/100G) enabled by XStream mode |
| Frame transmit modes | Port based (rate per port), stream based (rate per stream), burst, timed, random frame size with unique seed |
| Min/max frame size (w/CRC) | 64-16383 |
| Min/max Tx rates | 1 packet per 1.37 seconds to 101% of line rate |
| Real-time Tx stream adjustments | Change rate and frame length settings without stopping the generator or analyzer |
| Per-stream statistics analyzed in real time | Tx and Rx frame counts and rates; Tx and Rx Layer 1 byte counts and rates; FCS errors and rates; Min, Max, and Average Latency (32K streams); Real Time Dropped Frame count; Advance Sequence Tracking: Duplicate, reordered, late, and inordered |
| Per-port statistics analyzed in real time | Tx and Rx frame counts and rates; Tx and Rx Layer 1 byte counts and rates; PRBS errors; FCS errors and rates |
| Flow control | Support Priority Flow Control |
| Transmit timestamp resolution |
800G: 2.5 ns Tx timestamp resolution
With intra-chassis and inter-chassis synchronization
|
| Supported encapsulations | Layer 2: Ethernet II, 802.1Q, 802.1ad; Layer 3/4: IPv4, IPv6, TCP, UDP |
| Supported Tx signature capability | Fully compatible with VIAVI hardware; contains sequence number and highly accurate timestamp |
| Capture buffer size | 1.28 MB per port (MAX) |
| Capture filter patterns |
User defined pattern definitions can logically combine 8 filters of up to 32 total bytes
Patterns can be applied to start, filter (quality), or stop capture
|
| Latency modes |
LIFO, LILO, FIFO or FILO
Latency calculation methods supported by benchmark tests
|
| Route Insertion Table (RIT) entries per port |
128K (800/400/200G), 64K (100G), 32K (50G) 4-byte entries
For dynamic label or random IP/MAC address assignments
|
| RIT entries per port, XStream |
1M (800G), 256K (200/100G) 4-byte entries as primary table, and a secondary 64K (400/200/100G) entries table
Enabled by XStream
|
| RIT or list VFD entries per stream | 8 RIT insertions per stream (800/400/200/50G); 6 RIT insertions per stream (100G); 4 VFD insertions per stream for all supported speeds |
| GPU/NIC emulation | 8 GPUs / NICs emulated per port |
|---|---|
| CCL traffic patterns | AlltoAll with PXN, All to ALL without PXN, RingAllReduce, Halving and Doubling, Dual Tree, RingReduceScatter, RingAllGather, Alternating Ring, Point to Point, Mixed Traffic Patterns and Background Traffic |
| LLM model emulation |
Chat GPT, llama, DeepSeek, RestNet, and AlexNet
To emulate training traffic models
|
| RoCEv2 configuration | RoCEv2 profile applied per port; Up to 4K Queue Pairs per port; RC and UD transport mode; DCQCN congestion control; Traffic mode – Continuous and Bursty; Full Mesh, One-to-One, One-to-Many; Bi-direction and Uni- direction traffic |
| UET configuration | State full Link Layer; Link Layer Retry (LLR); Link Layer Error insertion; State full Transport layer; Reliable ordered delivery (ROD); Reliable un-ordered delivery (RUD); High number of Fabric End Point (FEP); RCC Congestion Control; Packet trimming; Dynamic entropy change; Traffic mode – Continuous and Bursty; Full Mesh, One-to-One, One-to-Many; Bi-direction and Uni- direction traffic |
| RoCEv2 throughput | 800G, 400G, 200G, 100G |
| UET throughput | 800G, 400G, 200G, 100G |
| Congestion avoidance | Packet Trimming support; Multi pathing support |
| Congestion control | ECN, PFC combination for lossy and lossless fabric; ECN detection, CNP notification; CNP interval; QoS priority for CNP notification; DCQCN; NSSC and RCC |
| Key results and statistics |
JCT, Ideal JCT; Tail latency (P95/P99/P99.9); Algorithm bandwidth (algbw) and Bus bandwidth (busbw) based on JCT; CCL Traffic Chart algbw and busbw based on transmission time; Total Transmission Time, Total Training Time; Average, Max and Min Iteration Time; Throughput, packet loss, out of sequence packets, packet latency; ECN, CNP TX/RX; Receiver Window Size; Packet Rate; RTT; Retransmission Rate
Available on Web and Windows GUI
|
| UEC conformance test suite |
over 50 built-in test cases
Validates implementations against the UEC 1.0 specification
|
|
Trackable flow count (AFT)
B3-800-OSFP-8-1550A, B3-800-OSFP-8-1300A, B3-800-OSFP-8-1200A, B3-800-OSFP-8-800A
|
Tx: 1 Million, Rx: 4K
AFT only supported on B3 8-port 800G and 400G; sub-speeds will be available in future software releases
|
|---|---|
|
HW stream blocks
B3-800-OSFP-8-1550A, B3-800-OSFP-8-1300A, B3-800-OSFP-8-1200A, B3-800-OSFP-8-800A
|
128 |
|
Max RIT table entries
B3-800-OSFP-8-1550A, B3-800-OSFP-8-1300A, B3-800-OSFP-8-1200A, B3-800-OSFP-8-800A
|
128K |
|
Rx out of sequence
B3-800-OSFP-8-1550A, B3-800-OSFP-8-1300A, B3-800-OSFP-8-1200A, B3-800-OSFP-8-800A
|
Support Out-of-Seq at per flow level |
|
Removed generator / analyzer features
B3-800-OSFP-8-1550A, B3-800-OSFP-8-1300A, B3-800-OSFP-8-1200A, B3-800-OSFP-8-800A
|
Pattern Fill / Pattern Match |
|
Egress flow tracking
B3-800-OSFP-8-1550A, B3-800-OSFP-8-1300A, B3-800-OSFP-8-1200A, B3-800-OSFP-8-800A
|
Supported |
| GUI operation | Windows-based workstation with 10/100/1000 Mbps Ethernet NIC; mouse and color monitor |
|---|---|
| Automation scripting | Linux or Windows-based workstation |
| REST API support | Mac, Linux, or Windows-based workstation |
Specifications are taken from the manufacturer's documentation. See the datasheet for complete and current values.
Applications
- AI data center and GPU cluster validation
- AI back-end fabric, GPU, AI accelerator and AI NIC testing
- 800G switch and router validation
- Multi-vendor interoperability validation
- Cloud computing and streaming services QoS validation
- Data center ToR and EoR switches and fabrics
- Chipset and silicon verification
- Terabit router testing
- Full Layer 2–3 testing including RFC 2544, 2889, 3918, 8239 and 6349
- AI-driven test automation
Options & Accessories
| Option | Description |
|---|---|
| HWO-B3-ROCE-ECN-800-4 | VIAVI RoCE ECN-800G for B3-800-OSFP or B3-800-QD — 4 X 800G PAM4 ports total per appliance |
| HWO-B3-ROCE-ECN-400-8 | VIAVI RoCE ECN-400G for B3-800-OSFP or B3-800-QD — 8 X 400G PAM4 ports total per appliance |
| HWO-B3-ROCE-ECN-200-16 | VIAVI RoCE ECN-200G for B3-800-OSFP or B3-800-QD — 4-port B3: 8 X 200G PAM4 ports. 8-ports B3: 16 x 200 PAM4 ports total per appliance. |
| HWO-B3-ROCE-ECN-100-16 | VIAVI RoCE ECN-100G for B3-800-OSFP or B3-800-QD — 4-port B3: 8 X 100G PAM4 ports. 8-ports B3: 16 x 100 PAM4 ports total per appliance. |
| HWO-B3-ROCE-ECN-1500A | VIAVI RoCE ECN-800G/400/200/100 Suite for B3-800-OSFP or B3-800-QD |
| HWO-B3-ROCE-ECN-1200A | VIAVI RoCE ECN-800G/400G Suite for B3-800-OSFP or B3-800-QD — 4-port B3 and 8-port B3: 4 X800G PAM4 ports or 8 X 400G PAM4 ports total per appliance. |
| HWO-B3-ROCE-ECN-1100A | VIAVI RoCE ECN-800G/200/100 Suite for B3-800-OSFP or B3-800-QD |
| HWO-B3-ROCE-ECN-1000A | VIAVI RoCE ECN-800G/200 Suite for B3-800-OSFP or B3-800-QD |
| HWO-B3-ROCE-ECN-700A | VIAVI RoCE ECN-400G/200/100 Suite for B3-800-OSFP or B3-800-QD |
| HWO-B3-ROCE-ECN-600A | VIAVI RoCE ECN-400G/200 Suite for B3-800-OSFP or B3-800-QD |
| HWO-B3-ROCE-ECN-500A | VIAVI RoCE ECN-400G/100 Suite for B3-800-OSFP or B3-800-QD |
| HWO-B3-ROCE-ECN-300A | VIAVI RoCE ECN-200G/100 Suite for B3-800-OSFP or B3-800-QD |
| HWO-B3-CCL-4 | VIAVI CCL for B3-800-OSFP or B3-800-QD — Speeds are included: 4 X 400G PAM4 or 4 X 800G PAM4 ports per appliance |
| HWO-B3-UET-800-8 | UET 800G for B3-800-QD or B3-800-OSFP — 4X 800G PAM4 ports total per test module |
| HWO-B3-UET-400-8 | UET 400G for B3-800-QD or B3-800-OSFP — 8X 400G PAM4 ports total per test module |
| HWO-B3-UET-1500A | UET 800G/400G/200G/100G Suite for B3-800-QD or B3-800-OSFP |
| HWO-B3-UET-1200A | UET 800G/400G Suite for B3-800-QD or B3-800-OSFP |
| HWO-B3-UET-600A | UET 400G/200G Suite for B3-800-QD or B3-800-OSFP |
| HWO-B3-UET-500A | UET 400G/100G Suite for B3-800-QD or B3-800-OSFP |
| HWO-B3-UET-300A | UET 200G/100G Suite for B3-800-QD or B3-800-OSFP |
| HWO-B3-UET-200-8 | UET Transport Layer-400G/100G Suite for B3-800-QD or B3-800-OSFP — 4-port B3: 8 X200G PAM4 ports. 8-port B3: 16 X 200G PAM4 ports total per test module |
| HWO-B3-UET-100-8 | UET 400G/100G Suite for B3-800-QD or B3-800-OSFP — 4-port B3: 8 X100G PAM4 ports. 8-port B3: 16 X 100G PAM4 ports total per test module |
| AON-B3-AI-4-1500A | B3 RoCE ECN and CCL 800G/400G/200G/100G Bundle (AION) |
| TPK-1106 | CCL Automation Benchmark Test Package — Requires RoCEv2 and CCL license. Works with A1, B2, and B3 only. |
| HWO-B3-800-OSFP-8-50G | 50G 112G/56G PAM4 HW Speed Option for B3-800-OSFP-8-T1S |
| HWO-B3-800-OSFP-8-100G | 100G 112G/56G PAM4 HW Speed Option for B3-800-OSFP-8-T1S |
| HWO-B3-800-OSFP-8-200G | 200G 112G/56G PAM4 HW Speed Option for B3-800-OSFP-8-T1S |
| HWO-B3-800-OSFP-8-400G | 400G 112G/56G PAM4 HW Speed Option for B3-800-OSFP-8-T1S |
| HWO-B3-800-OSFP-8-800G | 800G 112G/56G PAM4 HW Speed Option for B3-800-OSFP-8-T1S |
| HWO-B3-800-OSFP-8-PORT | B3-800-OSFP-8 Single Port Enablement |
| SWO-B3-800-OSFP-8-MACSEC | MACsec Software on B3-800-OSFP-8-T1S Appliance |
| BPK-1407 | Autonomous Flow Tracking (AFT) Package — Only supported on B3 8-port 800G and 400G, sub-speeds will be available in future software releases. |