Speed the development of BNG solutions with the Intel® BNG Acceleration Kit.
With the rapid growth of internet subscribers and connected devices, cloud service providers and communication service providers find it difficult to meet today’s capacity needs for the following:
- 4K video streaming
- work-from-home teleconferencing
- Internet of Things (IoT) propagation, such as for smart home appliances
- Artificial intelligence/machine learning (AI/ML)
The Intel® Broadband Network Gateway Acceleration Kit (Intel® BNG Acceleration Kit) makes it easier for network equipment manufacturers to quickly develop BNG solutions such as OpenBNG. The Intel® BNG Acceleration Kit is available for the Intel® Tofino™ Expandable Architecture, which provides massive performance, programmability, intelligence and visibility/control.
The Intel® BNG Acceleration Kit takes advantage of the P4 programmability of Intel® Tofino™ Expandable Architecture, enabling vendors to customize and optimize their BNG solutions for customers’ unique deployment needs.
This acceleration kit includes the necessary reference implementations to build a fully featured OpenBNG solution using the Intel® BNG Acceleration Kit’s foundational libraries. OpenBNG solutions that are created using the Intel® BNG Acceleration Kit benefit from the power of Intel® Xeon® Scalable processors and the acceleration capabilities of Intel® FPGAs and/or Intel® Infrastructure Processing Units (Intel® IPUs) to extend and augment Intel® Tofino™ Intelligent Fabric Processor (Intel® Tofino™ IFP) functionality.
This hardware acceleration increases the efficiency of the CPU, eliminating heavy networking constraints by partially or completely delegating resource-consuming network functions to the Intel® IFP and Intel® FPGA, resulting in very low latency. This enables manufacturers to bring new solutions to market quickly and meet their customers’ rapidly increasing capacity and flexibility requirements, while reducing development time as well as OpEx and CapEx.
- Hardened while customizable L2/L3 switching and routing
- Increased capacity
- Multiple-pipe programmable packet processing
- BNG-specific functionality
- Software programmability
- Fast time to market
- Open source components
- Hardware and software disaggregation
Overall, a BNG’s tasks include packet classification, session termination and collecting statistics for accounting. Once traffic classification occurs, the Intel® Tofino™ IFP marks the packets with additional metadata that tells the Intel® FPGA how to treat that traffic. The FPGA extracts that metadata and enqueues the packets accordingly, then the packets are stored in high bandwidth memory (HBM) while they await dequeuing. The scheduling and shaping algorithm propagates the packets through the scheduling hierarchy and, when appropriate, releases the packets back to the Intel® Tofino™ IFP for routing to the rest of the network towards their destinations. The Intel® BNG Acceleration Kit includes four reference implementations that perform all the necessary packet management functions in the BNG:
- CPU control plane.
- FPGA HQoS
The need for a flexible architecture that offers more bandwidth and lower latency is continuing to grow. Intel has worked with the ecosystem to develop a broad, end-to-end portfolio of fully programmable, interoperable and modular components—including the Intel® BNG Acceleration Kit —that can enable a wide range of OpenBNG scenarios.