Wednesday, February 6, 2013

MoSys unveils bandwidth engine IC with on-board macro functions for 400G network equipment

USA: MoSys, a leader in semiconductor solutions that enable fast, intelligent data access for network and communications systems, announced the newest member of the Bandwidth Engine family of products that accelerates intelligent networking functions.

The Bandwidth Engine 2 – Macro delivers the highest access rate and throughput of any single device today coupled with offload accelerators for single rate and two rate three color marker (srTCM, trTCM) metering, statistics, and accounting applications.

As network performance and feature requirements continue to scale, architectural improvements are required. Networking equipment has transitioned to highly parallel, multi-threaded processing System-on-Chip complexes which require an insatiable amount of memory bandwidth. The Bandwidth Engine 2 family has three purpose-built variants, Burst, Access and Macro, to meet these growing needs and is intended for high-reliability, carrier-grade applications.

Using sixteen 15 Gigabits per second (Gbps) SerDes lanes, the Bandwidth Engine 2 interface operates at 480 Gbps, providing the host with up to 384 Gbps CRC protected, effective data throughput. This represents an unprecedented 80 percent overall efficiency, well beyond the capability of standard memory subsystems and alternative serial interface solutions, while using less than half of the board area, interface pins, and power resulting in substantial system-level cost savings.

The new device, MSR820, with its on-board accelerators, is capable of fire-forward operations which can update records entirely internal to the device, reducing the number of memory bus transactions from six down to one, as well as relieving the host of the computations required for the update.

The MSR820’s macros can be saturated using only 8 SerDes lanes, further reducing the power, pincount and host resources. The macro functions can retire entire operations in under 30 nanoseconds (ns), far quicker and at substantially lower power than alternative solutions, making Bandwidth Engine 2 – Macro a device unique to the industry in its capabilities.

“The industry is challenged to provide high-performance line cards that can aggregate hundreds of Gigabytes of bandwidth and deliver ever increasing intelligence,” stated John Monson, VP of Marketing at MoSys.

“The MSR820 Bandwidth Engine - Macro, delivers up to twelve billion operations per second for onboard or host-based processing, eliminating as many as 6 to 8 transactions with a single command.  This industry-leading performance capability, combining memory bandwidth, intelligence features and efficiency improvements, enables networking and compute architects to achieve both increased speed and intelligence for packet or data processing applications.”

MoSys’ Bandwidth Engine family of ICs utilizes the GigaChip Interface, an open, 90 percent efficient, reliable transport protocol optimized for chip-to-chip communications. The devices are compatible with CEI-11G and XFI SerDes, which allows a seamless interface with high performance FPGAs as well as standard libraries available from ASIC providers.

A complete package of RTL and tools is available to support the Bandwidth Engine interface.

MoSys' first generation Bandwidth Engine ICs have been fully qualified for carrier-grade applications and is available for volume production now.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.