Keysight has introduced the APS-ONE-400, a modular network cybersecurity test platform that fits 4x100GE test capacity into a single rack unit. The appliance generates hyperscale Layer 4-7 traffic, encrypted TLS (Transport Layer Security) throughput, security strikes, and Zero Trust Network Access (ZTNA) testing for the network equipment manufacturers (NEMs), service providers, and data center operators that validate their networks before deployment.

In that single rack unit, the APS-ONE-400 delivers 400 Gbps of Layer 4-7 throughput, 380 Gbps of hardware-accelerated encrypted TLS throughput, and 95 Gbps of Elephant Flow throughput—the large, sustained data transfers common in AI and large language model (LLM) training. It supports 100GE, 25GE, and 10GE fanout, and runs either as a standalone appliance or alongside Keysight’s existing APS-100/400GE systems, including the APS-M8400, APS-M1010, and APS-ONE-100.

Paired with the APS-M8400 appliance or APS-M1010 management controller, the platform scales into hyperscale testbeds that generate up to 16 Tbps of Layer 4-7 traffic, 20 billion concurrent connections, 15 Tbps of TLS traffic, and 25 million TLS connections per second. Capacity can be added later, so a lab can start with one appliance and scale up as its test needs change.

The design is built for the traffic that AI and machine learning workloads push through data center and service provider networks, including post-quantum cryptography (PQC)-encrypted flows and the Elephant Flow datasets generated during LLM training. Hardware acceleration handles the TLS encryption, so encrypted throughput stays close to the platform’s raw line rate rather than bottlenecking on the cryptography. Consolidating that load generation into one rack unit reduces the rack space, power, and cooling a multi-tool test setup would otherwise require. The APS-ONE-400 joins Keysight’s test tools for AI data center networks, which include its 1.6T Ethernet AI fabric emulation for 224G lanes.

“Keysight’s modular APS-ONE-400 compute node delivers new heights in realism, emulating the traffic flows associated with generative AI models and agentic applications at hyperscale, including the huge Elephant Flow datasets common to LLM training use cases and PQC-encrypted traffic flows, all in a compact form factor that conserves critical lab resources,” said Ram Periakaruppan, Vice President and General Manager of Keysight’s Network Test and Security Solutions.

Source: Keysight