Learn how Intel® Intelligent Fabric optimizes data center connectivity through enhanced intelligence, performance, visibility, and control. Explore resources to understand how cloud service providers (CSPs), communications service providers (CoSPs), and high performance computing (HPC) / AI clusters can benefit from the full range of Intel technologies that deliver innovation to connectivity fabric for data centers moving to cloud native technologies.
Optimized, Scalable, Intelligent Data Center Fabric
To meet continuously changing needs, data center advancement requires an optimized, flexible network architecture to manage escalating and dynamic accommodations for virtualization/containerization, workload optimization, and bandwidth expansion. Intel® Intelligent Fabric is an integrated architectural approach delivering uncompromising performance and business agility to address data center acceleration within the network.
Intel® Intelligent Fabric Vision for the Data Center of the Future
Software-driven and developer-led, Intel® Intelligent Fabric prioritizes decentralized computing and smart infrastructure in support of data center innovation and current architectures. The Ethernet-based leaf/spine network topology is based on three categories of compute (CPUs, XPUs, and IPUs) and a suite of Intel hardware and software—delivering optimization across all network elements.
Efficiently connect all data movement between processing, acceleration, and storage in a modern data center, while increasing scalability and flexibility through the increased intelligence, high performance, visibility, and control of Intel® Intelligent Fabric.
Open-source P4 programming technology for developer customization, packet forwarding control, and enhanced end-to-end security.
Scale and deploy for large-scale workloads with high-speed Ethernet and high-throughput switching capacity. High-speed silicon photonics for optical data transport with high reliability.
Visibility and Control
Real-time edge-to-cloud telemetry and optimized congestion control. Self-healing capabilities for configuration and issue resolution.
The Intel® Tofino™ Expandable Architecture is a massive solution meant for massive use cases and represents a powerful new way to capitalize on the Intel® Intelligent Fabric’s incredible potential.
The solution provides network owners and operators with the flexibility that they need to customize their networks for a variety of workloads, while also enabling the addition of supporting network functions. The Intel Tofino Expandable Architecture delivers capabilities that greatly exceed those of traditional network switching.
Intel Tofino Expandable Architecture offers four key benefits:
Performance and Scale – Increasing switch and router table and buffer capacity by two orders of magnitude when compared to a standard data center-focused ASIC.
Lowers Total Cost of Ownership – Support across hardware, software, and IP make the Intel Tofino Expandable Architecture an excellent choice for use cases that need larger tables/deeper buffers and more advanced shaping/scheduling functions than are available with standard switches.
Open Standards Support – Provides customers choice on when and how to optimize their network by offering the freedom to choose from a wide variety of products.
Flexibility – The P4 programmability within Intel® Tofino™ Intelligent Fabric Processors (IFPs) allows users to dial in performance, while the power of Intel® CPUs and acceleration capabilities of Intel® FPGAs and Intel® IPUs both extend and augment Tofino IFP functionality. Efforts are underway to extend the existing P4 compiler and P4 Studio to all networking components.