Dell PowerSwitch S5448F-ON 100GbE Switch

The Dell PowerSwitch S5448F-ON is a high-density, open-networking top-of-rack switch engineered for modern data-center leaf and spine deployments. Built on a merchant silicon architecture, it delivers 48 ports of 25GbE SFP28 alongside 8 ports of 100GbE QSFP28, yielding aggregate throughput designed for demanding east-west traffic patterns typical of hyperconverged infrastructure, high-performance computing, and cloud-scale workloads. The platform ships with the Open Network Install Environment (ONIE) bootloader, enabling customers to run Dell's own Enterprise SONiC Distribution or third-party network operating systems without vendor lock-in.

The S5448F-ON is purpose-built for leaf-layer roles in spine-and-leaf Clos fabric architectures, providing the low-latency, line-rate forwarding required by latency-sensitive applications such as NVMe-over-Fabrics storage, AI/ML training clusters, and financial services workloads. Its 1U form factor integrates into standard data-center rack footprints while supporting flexible cabling topologies including breakout configurations on its QSFP28 uplink ports. The switch includes front-to-back and back-to-front airflow options to accommodate hot-aisle/cold-aisle containment strategies.

As a member of Dell's open-networking PowerSwitch portfolio, the S5448F-ON aligns with disaggregated network models where hardware and software lifecycles are managed independently. Integration with automation frameworks such as Ansible, Puppet, and REST APIs enables consistent, programmable network operations within DevOps and NetOps workflows. Redundant hot-swappable power supplies and fans reinforce the availability requirements of always-on enterprise and service-provider environments.

Ideal for

• Top-of-rack leaf switch in high-density 25GbE server access deployments within spine-and-leaf data-center fabrics
• Storage networking fabric for NVMe-over-Fabrics and high-throughput SAN environments requiring low-latency, lossless forwarding
• AI and ML compute cluster interconnect where high bandwidth and minimal latency between GPU nodes are critical
• Hyperconverged infrastructure aggregation supporting platforms such as VMware vSAN and Nutanix at scale
• Open-networking deployments running Enterprise SONiC or third-party NOS where hardware and software lifecycle independence is required
• Network segmentation and micro-segmentation enforcement at the access layer within zero-trust data-center security architectures

Technical specifications

Available from Omnixon Global. Submit an RFQ and our team will confirm configuration and availability for your order.