I’ve run Linux on servers for years, so the promise of Whitebox Switching genuinely excited me. The idea of disaggregation between hardware and software—buying a powerhouse and choosing my own OS—felt like the ultimate freedom.
I picked up an Edgecore 5610 (48x10G sfp+ / 4x40G qsfp) to prove it. But that excitement quickly hit a proprietary brick wall.
The “Not So Open” Paradox Despite the labels, “Open Networking” often feels like a pay-to-play club. Once hardware hits EOL, the software ecosystem vanishes: • Cumulus Linux: No legacy support post-NVIDIA; no open release. • Pica8: No community/trial versions available. • Open Network Linux (ONL): A great framework, but the ASIC integration remains a “black box.”
Why, when these systems go End of Support, can’t the software be released? Without it, this high-performance gear is just e-waste.
The Engineering Grind 🛠️ I’ve spent more time then I’d like to admit, pushing OpenNSL and cross-compilation to the limit. I’ve successfully: ✅ compiled Open Network Linux for the PowerPC (P2020). ✅ Got fans and thermal management behaving. ✅ Got SFP+ ports and different sfps recognized and powered.
But loading software only gets you to the door. The Final Boss is the Broadcom Trident ASIC. I have the peripherals humming, but the packet-switching logic is a locked vault.
Next Stop: Reverse Engineering. 🔍 Since official paths are closed, I’m moving this into my reverse engineering projects. I’m going to see if I can manually bridge the gap to the silicon and give this hardware a second life. It shouldn’t take a five-figure contract to move a packet in a lab.
Can you help? 🆘
Does anyone have a lead on a NOS, an old SDK build, or legacy files for a P2020/Trident combo? Let’s keep this gear off the scrap heap.