PALO ALTO, Calif., STAMFORD, Conn., and PHILADELPHIA—Broadcom Inc., Charter Communications, and Comcast have announced a landmark effort to jointly develop Unified DOCSIS chipsets for network nodes, smart amps, and cable modems, enabling both FDX and ESD versions of the DOCSIS 4.0 specification.
The effort is important because it sets the stage for operators to deliver upwards of 25 Gbps speeds over existing networks.
Comcast is showcasing early versions of technology supporting 25 Gbps speed and support for up to 3 Ghz during the SCTE TechExpo in Atlanta.
“By pushing the benefits of AI and ML to the edge of our network, we are maximizing network performance, improving the internet experience, and introducing new and exciting services to support our customers' incredible appetite for more, higher quality data,” said Charlie Herrin, president, Technology, Product, Experience, Comcast. “Supporting a single technology that can be used by all operators efficiently, over their existing networks, allows the industry to leverage its collective investment, and to increase the benefits of AI at the edge.”
“Charter, Comcast, and Broadcom are collaboratively driving industry innovation through the Unified DOCSIS 4.0 specification,” added Rich DiGeronimo, president of Product and Technology, Charter Communications, which operates the Spectrum brand. “By integrating Broadcom's cutting-edge technology into our network infrastructure and cable modems, our network will leverage artificial intelligence to enhance power, reliability, and cost-effectiveness, ultimately delivering superior services to our customers.”
This specification brings together all the benefits of both DOCSIS ESD and DOCSIS FDX and includes support for both mid-split and high-split configurations.
Designed for compliance with the existing DOCSIS 4.0 standard and creating a foundation for the next generation, the technology solutions will enable all operators through a single and open standard, delivering on the following advantages:
- Economies of scale for the industry
- Robust vendor ecosystem
- Streamlined CableLabs certification and operator qualification
The companies also reported that this collaboration will dramatically enhance DOCSIS networks with advanced Artificial Intelligence and Machine Learning, utilizing Broadcom’s embedded Neural Processing Unit (NPU) in the network nodes, smart amps, and modems. Operators will be able to increase operational efficiency and improve network security and reliability through specific advancements, including:
- Channel anomaly detection and network self-healing
- Predictive network power management and reduction
- Real-time network pattern detection, classification, and optimization
- Advanced cybersecurity, intrusion detection, and phishing protection to fight AI-based attacks
- Monitoring and optimizing home IoT and WiFi device performance
This three-way collaboration also augments each company’s investment in the future of DOCSIS through an agreement to start developing a path to 25 Gbps. Among the enhancements of this evolution will be the extension of DOCSIS to higher spectrum. By expanding the supported bandwidth on hybrid fiber coax (HFC) networks up to 3 GHz, this pioneering technology delivers many benefits to cable operators’ existing networks without requiring them to build new infrastructure. While this initiative is in the early planning stages, it would ultimately allow for increased network speeds for continued competitive service offerings and the ability to apply operator software into next generation advanced products.
“Broadcom is pleased to work with our operator partners in bringing the industry together around Unified DOCSIS,” added Charlie Kawwas, Ph. D., President of Broadcom’s Semiconductor Solutions Group. “With the 3 GHz capabilities, we’re providing the tools for cable operators to continue to enhance the leading capabilities of their networks while competing with higher and higher speed offerings. We’ve seen the huge benefits to operators and consumers of embedded AI/ML in access networks. From advanced diagnostics and network self-healing to state-of-the-art AI-based cybersecurity, the sky's the limit for innovative machine learning applications.”
