Data Center
Participate in insightful discussions regarding Data Center topics
57 Discussions

Intel Intelligent Fabric Optimizes Connectivity for the Data Center of the Future

Ed_Doe
Employee
0 0 2,660

At Intel Innovation, Intel announced a groundbreaking step – Intel® Tofino™ 3 Intelligent Fabric Processor to power the Data Centers of the Future.


Conventional data centers are complex. Workloads are distributed across multiple networks (on-premises, cloud, hybrid environment to name a few). Real-time data, complex algorithms and infrastructure traffic can lead to network congestion that results in latency and bandwidth challenges, thereby reducing application efficiency and performance. This can lead to service outages and unwanted increases in OPEX. As a result, data center operators are facing an unprecedented demand to remediate issues rapidly (or better yet prevent them) while maintaining high-availability and performance in distributed workload environments.


Furthermore, new services and workloads that include, for example, 5G, IoT, AI/ML, AR/VR, are growing rapidly and will require unquenchable demand for bandwidth. This demand will cause today’s data centers to quickly outgrow their ability to analyze, process and secure the data in real time.


Conventional methods of increasing throughput from existing silicon will be insufficient and will demand sophisticated features that provide actionable network intelligence to quickly troubleshoot and resolve issues.


Chart showing how AI workloads, general compute, storage, and optimization functions all come together under one fabric.



Intel® Intelligent Fabric Vision


Intel’s vision to meet these needs is the Data Center of the Future (DCoF) and the Intel® Intelligent Fabric is Intel’s architectural approach to address the DCoF’s fabric requirements. It is a novel Ethernet connectivity framework aimed at optimizing data center compute fabrics from cloud-to-edge, optimizing connectivity for cloud-native and remote applications. It brings together a full range of Intel connectivity products and technologies that include Intel® Tofino™ Intelligent Fabric Processors (Intel® Tofino™ IFPs), Intel® Xeon® Scalable processors, Intel® Infrastructure Processing Units (Intel® IPUs), Intel® Ethernet Network Adapters, Intel® FPGAs, and Intel® Silicon Photonics to provide intelligence, performance, scalability, and manageability.


List of various Intel hardware products.


Intel has taken a leadership role developing industry standards designed to enable cloud and edge data centers, high performance computing (HPC) and communication service providers (CoSPs) to reduce overhead and uplift networking performance. By building on these standards, customers will better utilize resources with a secure, programmable, and versatile solution that enables them to balance data processing, movement, and storage in a modern data center, while reducing network complexity, OPEX, and CAPEX.



Silicon at the Heart of Intel® Intelligent Fabric


The Intel® Tofino™ 3 Intelligent Fabric Processor (IFP) is purpose built for data centers of the future. It supports up to 25.6 Tbps of non-blocking switching throughput and up to 10 billion packets per second of forwarding capacity with highly consistent packet rates and forwarding latencies. This is double the capacity from Intel® Tofino™ 2.


The mission for the Intel Tofino 3 IFP is to drive and develop a new architecture that encompasses programmability, power optimization, high-performance and deep insights into the network and avoid any tradeoffs on packet buffering or other table scale parameters that are typical to many switch ASICs.


The Intel Tofino 3 IFP provides P4 programmability and accelerates packet processing via a Match-Action Pipeline architecture, for edge-to-cloud workload environments. It also optimizes power consumption for hyperscale use cases, accelerates machine learning (ML) for distributed training workloads and supports real-time in-band network telemetry (INT) spanning the Edge-to-Cloud spectrum.


The Intel Tofino 3 IFP’s set of robust hardware and software capabilities will allow network engineers to spend more time advancing rather troubleshooting and maintaining their network. What’s more, Intel’s comprehensive suite of workload-monitoring tools, Intel® P4 Studio Software Development Environment (Intel® P4 SDE), Intel® Deep Insight Network Analytics Software, and simplified integration of container-based microservices, will provide long-term investment protection with agility and reduced complexity. This further leads to more secure and self-healing networks and cloud infrastructure.



Switch to Intelligence


Intel IFPs deliver a dynamic suite of intelligent, software-defined solutions to transform your network switching infrastructure into an intelligent fabric. With the Intel Tofino 3 Intelligent Fabric Processor (IFP), IT and network administrators will see increased performance with power optimization in distributed work-load environments.


For more information, visit: intel.com/IFP which provides all available information about these exciting new products. On the page you can find our new Intel Tofino Intelligent Fabric Processors Brochure which covers the full family of Tofino products as well as the new Intel Tofino 3 Intelligent Fabric Processor Product Brief.


To learn more about Intel Intelligent Fabric, attend Intel’s OCP Summit Expo Hall Talk by Barry Davis on Nov 9, 1:10 PM PST.







Notices & Disclaimers
Intel technologies may require enabled hardware, software or service activation.
No product or component can be absolutely secure.
Your costs and results may vary.
© Intel Corporation. Intel, the Intel logo, and other Intel marks are trademarks of Intel Corporation or its subsidiaries. Other names and brands may be claimed as the property of others.
About the Author
Edward (Ed) V. Doe is vice president of the Connectivity Group and general manager of the Barefoot Division (BXD) at Intel Corporation. Doe holds a bachelor’s degree in electrical engineering from the University of Western Ontario, London, Ontario, Canada and two years of graduate studies towards a master’s degree in electrical engineering at Stanford University.