Cisco equips network service providers with tools to reinvent their networks for Artificial Intelligence integration.
In a significant move, networking and security company Cisco has announced a new advanced network architecture - Cisco Agile Services Networking - designed to meet the demands of AI workloads and connectivity. The architecture, an evolution of Cisco's Converged SDN Transport (CSDN-T) architecture, is purpose-built to handle the evolving traffic characteristics of AI, such as high data intensity, low latency requirements, and distribution across clouds, data centres, and edge locations [1][2].
According to Jeetu Patel, the EVP and chief product officer at Cisco, the advent of AI agents will lead to an influx of new digital workers requiring high-bandwidth, secure, and energy-efficient connectivity [2]. Recognising this, Cisco Agile Services Networking provides an AI-native, automated, and high-capacity network designed to simplify operations and increase resilience [2].
The architecture treats the network edge as flexible functions deployable anywhere, enabling greater agility and efficiency in network deployment [1]. It is optimized for intelligent service delivery, enabling service providers to monetize services deployed closer to their end-users [2].
Cisco Agile Services Networking empowers service providers to transform their network infrastructure into AI-ready platforms. It allows providers to support emerging AI service models such as GPU-as-a-Service and Inference-as-a-Service, creating new revenue streams by enabling them to offer specialized AI compute and intelligent services to industries that rely on AI workloads [2][4].
The architecture incorporates new converged access and edge routers powered by Cisco Silicon One, expanding efficiency and functionality for service providers. This leads to operational cost savings and enhanced service offerings [4]. Cisco also introduces a multi-agent AI framework integrated with its Crosswork Network Automation platform, enabling service providers to automate complex network operations and decision-making, reducing manual effort and accelerating service delivery, which improves profitability [4].
By seamlessly integrating terrestrial and satellite networks, service providers can extend AI connectivity to remote and underserved areas such as maritime, aviation, IoT, and disaster response sectors. This broadens market reach and unlocks new monetization opportunities in previously hard-to-serve regions [4].
Cisco's architecture simplifies network management by unifying software-defined networking fabrics and operational controls, making it easier and more cost-effective for providers to deploy and assure AI connectivity services [3]. The network automation and assurance features are available now, including new predictive AI technology for network capacity planning and resource allocation, and integration with Splunk for automated decisions [3].
The first Silicon One-powered 8000 series is available today, with additional models targeted to ship in spring and summer of 2025. Cisco is also introducing new Silicon One systems and platforms, including the A100 and K100 devices for fixed, centralised, and modular systems [1].
Cisco's new solutions are designed to handle the increase in data volume and variety from AI applications. The new Cisco Silicon One systems and platforms will offer consistent control, sustainability, security, and manageability for network building from the silicon and systems level [1]. Cisco is also extending Routed Optical Networking to simplify network, improve management, and use less energy by further converging IP and optical layers [1].
Moreover, Cisco's new 400g ULH coherent pluggable optics enable ultra-long-haul applications and connect locations up to 3000km apart without traditional transponders [1]. The architecture is more energy-efficient and less resource-intensive, with one edge architecture use case projected to reduce greenhouse gas emissions by 324 metric tonnes CO2e across an assumed product lifetime and all product life phases when combined with Cisco's Takeback and Reuse program, and lower energy consumption by 80% [2].
In conclusion, Cisco Agile Services Networking offers service providers a path forward for AI connectivity, enabling them to monetize the delivery of assured services and networking. The architecture is a significant step towards meeting the demands of AI workloads and connectivity, and it presents a significant opportunity for service providers in the AI revolution.
Technology played a crucial role in the development of Cisco Agile Services Networking, a data-and-cloud-computing solution designed for AI workloads. Recognizing the influx of new digital workers requiring high-bandwidth, secure, and energy-efficient connectivity, Cisco Agile Services Networking is an AI-native and automated network, optimized to simplify operations and increase resilience.
