¶ Current Models
¶ Power Supply
USB-C 5V/3A UL Power Adapter with ON/OFF Switch
Features:
- UL and FCC Approved and Listed
- Convenient ON/OFF Switch
- DOE Level VI Efficiency
- Over Circuit and Short Circuit Protection
- Compatible with all Defend-O-Tron V100 models and other SBCs with USB-C Connector
- 5V/3A Output and 47 inches Cable Length
Specifications / Contents:
Model: RAS-PWR13-PI
Input: AC 100-240V/50-60Hz
Output: DC 5V/3A
Cable length: 47 inches
Weight: 0.25 lbs
¶ Future Roadmap — V200 (In Development)
The V200 is in active design as of 2026. Specifications below are targets, not commitments — expect refinement before release. We'll publish a firm spec sheet and release window once the platform enters validation.
The next-generation Defend-O-Tron is being designed around the Radxa E54C network-computer platform, aimed at giving the appliance significantly more headroom for advanced threat analysis, on-device AI inference, and larger network deployments.
¶ Targeted upgrades over V100
| Spec | V100 (current) | V200 (target) | Improvement |
|---|---|---|---|
| SoC | Rockchip RK3568B2 | Rockchip RK3582 | New generation |
| CPU | 4× Cortex-A55 @ 2.0 GHz | 2× Cortex-A76 + 4× Cortex-A55 (6 cores) | Higher single-thread + efficiency cores |
| RAM | 2 / 4 GB LPDDR4X | 16 GB LPDDR4X (32 GB option) | 4–8× |
| NPU | 0.8 TOPS @ INT8 | 5 TOPS @ INT8 | ~6× |
| Network ports | 1× GbE management + 2× 2.5 GbE protect | 4× Gigabit Ethernet | More flexibility for branch / multi-segment networks |
| Storage | M.2 PCIe 2.1 x1 NVMe | M.2 NVMe + optional eMMC (8/16/32 GB) | More built-in storage options |
| Display | HDMI 2.0 | HDMI 2.1 (up to 8K) | Higher console fidelity |
| USB | 2× USB 3.0 Type-A | 1× USB 3.0 OTG Type-C + 1× USB 3.0 Type-A + 2× USB 2.0 | More peripheral options |
¶ What the extra horsepower unlocks
- Larger networks, more segments. Four Gigabit ports give branch deployments and multi-VLAN environments more wiring flexibility without external switching.
- On-device AI threat analysis. The 5 TOPS NPU is sized to run small Large Language Models — Radxa documents 1.5B-parameter LLMs (e.g. DeepSeek) running locally on the same hardware. That opens the door to running anomaly-detection and threat-classification models directly on the appliance, without forwarding traffic or logs to cloud analysis.
- Future-proofing. 16 GB RAM is a generous baseline for the existing workload (CrowdSec engine, Suricata, audit subsystem, dashboards, container stack) and leaves room for the next several years of defense components.
¶ Open Hardware
If you'd like to be notified when the V200 enters availability, watch this page or the project home.