Hardware-based point-to-point encryption encompassing Ethernet speeds up to 400G is being deployed in cloud, carrier and enterprise networks to address the market demand for enhanced data security. In addition, the stringent timing requirements of 5G radios are driving greater timing accuracy.
The 400GbE PHY device incorporates Marvell’s 56G PAM4 SerDes technology, IEEE 802.1AE 256-bit MACsec encryption and highly accurate PTP timestamping and is intended to deliver security and timing precision for hyperscale data centre, edge, enterprise and 5G infrastructure applications.
In combination with the company's Prestera CX 8500 400GbE switch, the transceiver allows data to be securely moved to the smart edge to meet the bandwidth and latency demands for critical applications enabled by 5G and artificial intelligence.
The 88X7121P, the latest addition to Marvell’s Alaska® C family of Ethernet transceivers, supports both retiming and gearboxing applications and is footprint- and software-compatible to Marvell’s 88X7120 PHY, providing design flexibility for customers.
The device is fully compliant to IEEE standards for 400GbE, 100GbE and 50GbE, and exceeds the electrical specifications to interface with QSFP-DD and OSFP optical modules.
The 88X7121P’s 256-bit MACsec-based encryption for point to point links provides enhanced security and allows for flexible deployment of MACsec encryption without incurring the cost and power burden of including this functionality in the switch ASIC.
According to the company, highly accurate Class C PTP timestamping incorporated in the device enables enhanced timing precision required for carrier, wireless backhaul and 5G infrastructure applications.
Commenting Faraj Aalaei, executive vice president of the networking business group at Marvell said that the PHY has been, “optimised for differentiated, innovative solution deployments to address the distinctive security and performance requirements of hyperscale data centres, enterprise networks and 5G infrastructure.”
“We see an uptrend in data centres transitioning to 100GbE, 200GbE and 400GbE to meet ever-increasing processing and I/O bandwidth demands, particularly as artificial intelligence and machine learning workloads ramp. Enhanced protection and security of data is also becoming a key requirement for cloud operators and network-equipment vendors,” said Bob Wheeler, principal analyst for networking at The Linley Group.