Edge AI deployments are pushing network bandwidth to its limits, and Innodisk’s latest 10 GbE LAN Series aims to meet that demand with a lineup built for low-latency connectivity and compact integration.
The series includes both M.2 and PCIe form factors, leveraging Intel E610/X710 controllers to deliver high-throughput performance while supporting advanced features like DPDK 25.07, PTP, and SR-IOV. These capabilities are designed to accelerate packet processing, time synchronization, and virtualized resource management—critical for applications such as inference, autonomous vehicle coordination, smart factory vision, and surveillance.
Key specs
- Form factors: M.2 2242, M.2 2280, PCIe low-profile
- Controller: Intel E610/X710 Ethernet
- Support: DPDK 25.07/16.11, PTP, SR-IOV
- Connectivity options: Copper and SFP+ (optical)
- Temperature range (selected models): -40°C to 85°C
The EGPL-T2F1 stands out as the first M.2-based SFP+ module, offering both optical and direct-attached copper connectivity. It was recognized with the Embedded World 2026 Best in Show Award for its flexibility and performance. Other models, such as the dual-port EGPL-T203, provide wide-temperature operation, making them suitable for harsh environments like outdoor deployments or factory automation.
Performance and flexibility
The daughterboard architecture, combined with high-speed shielding cables, simplifies integration in space-constrained systems while reducing complexity. This makes it easier to upgrade existing designs without major redesigns—a key consideration for embedded developers working under tight deadlines.
Innodisk’s focus on edge AI networking addresses a growing gap where traditional 1 GbE connections struggle to handle real-time data processing, high-resolution video streams, and multi-sensor integration. The series is positioned as a solution for developers building systems that require both speed and adaptability.
The lineup includes additional models like the EGPL-T103, ELPL-T101, and ELPL-T201, providing options for different deployment needs. Availability and pricing have not been confirmed but are expected to align with the demands of edge AI infrastructure.
