Boosting AI efficiency: FS rolls out advanced 200G networking switch
The solutions provider FS has launched the N8550-24CD8D, a next-generation 200G switch for use in AI-focused storage systems and data centres. The high performance switch has been designed to support growing AI workloads and the complex demands of hybrid infrastructure upgrades.
It provides fast data transfer speeds and versatile port configuration that can adapt to various networking needs. The switch is purpose-built for scalability, flexibility, growth, and overall performance, for the next wave of advanced networking technologies.
With AI workloads on the increase and global enterprises needing faster data without packet loss, upgrading from 100G / 200G to 400G architectures is becoming increasingly important. FS’s high-density N8550-24CD8D switch has been engineered to meet these needs, helping data centres and network infrastructures upgrade to handle the greater demands of modern infrastructures.
The switch connects to the core (spine layer) of data networks via 400G uplinks and has 24 200G ports, each capable of being split into either two 100G ports or four 50G ports. FS hopes the model will provide the necessary flexibility, adapting to the connected equipment’s capabilities.
For data centres upgrading to a 400G spine, the switch has eight 400G uplinks and can easily integrate with FS’s other high-end switches, like the N9550-32D or the N9550-64D, both of which are designed to operate at the core. The N8550-24CD8D can help organisations upgrade to a full 400G infrastructure gradually, preventing disruption and the need for high CAPEX up front.
Two advanced protocols are also featured – EVPN-VXLAN and MLAG. EVPN-VXLAN is designed to extend Layer 2 networks over Layer 3 infrastructure, typically beneficial in cloud environments. The MLAG protocol allows multiple switches to act as a single logical unit, improving overall performance and making management simpler.
The switch can act as an aggregation leaf in AI storage network architectures, letting users link servers and storage devices for low-latency and reliable data transfer through the use of RoCEv2, PFC, and ECN.
Built on the Broadcom Trident 4 chip, and with large data buffers, the N8550-24CD8D is built to deliver fast, high-performance networking, the company says.
(Image source: “Spine” by jurvetson is licensed under CC BY 2.0.)
Want to learn more about AI and big data from industry leaders? Check outIntelligent> Automation Conference, BlockX, href=”
Explore other upcoming enterprise technology events and webinars powered by TechForge here.
