NLA1200 Heterogeneous Knowledge-Based Processors with NetRoute™ Technology
*Certain versions of this device are not available for sale in Germany or shipment to or from Germany, and these versions should not be used in any product destined for the German market.
These 28 nm knowledge-based processors are based on heterogeneous architecture.
The architecture enables cutting-edge performance (2.4 billion searches per second), while delivering unprecedented power savings. This is achieved at deterministic latency, which is essential for the majority of telecommunications applications, including routers (core, edge and multiservice), enterprise, datacenter and carrier Ethernet switches, security appliances and wireless infrastructure equipment. The heterogeneous knowledge-based processor (KPB) provides network awareness and enables real-time modifications and updates to the routing configuration, making it ideal for packet classification, policy enforcement and Layer 2 / Layer 3 forwarding. The integration of an on-chip user data array facilitates a lower overall system cost to the end user by reducing bill-of-materials costs as well as ASIC complexity.
The 28 nm heterogeneous KBPs deliver flexibility and scalability to address the needs of a large landscape of networking applications. Furthermore, the heterogeneous architecture is complemented with a feature-rich and easy-to-use software toolset encompassing Broadcom's proprietary NetRoute™ technology, which allows support for up to 2.4x the knowledge database capacity per device (9.6x larger database capacity per system) without compromising performance.
- Concurrently delivers up to 24x higher performance, 2.4x knowledge database capacity per device
- 2.4 billion decisions per second to address growing line rates for IPv4 and IPv6 packets
- High-speed serial links for enhanced communication from KBP to system packet processors. 12.5 Gb/s SerDes for 300 Gb/s aggregate bandwidth
- Dual-port mode for off-the-shelf network processors, FPGAs and customer ASICs
- Improved search flexibility with context buffers and Key Processing Unit (KPU)
- Range Encoding Engine (REE) delivers efficient database compaction for ACLs
- Support for on-chip associated data for search tables to enable lowest system latency
- Cable Modem Termination Systems
- Carrier Access Products
- Data Center Servers (switching)
- Deep-Packet Inspection
- EDGE and Terabit Routers
- Enterprise WLAN
- Line Cards
- Metropolitan Packet Aggregation
- Storage Area Networks
- Wireless Base Stations