04-16-2020, 04:06 AM
(This post was last modified: 04-16-2020, 04:08 AM by danielt.
Edit Reason: Grammar...
)
I'm afraid to say I'm a little sceptical of the router proposal, or at least I am when it is "just" focused on being a router.
Whilst Pine64 direct sales/pricing strategy is undeniably important, I think Pine64 has gained a lot of attention recently by combining its sales model alongside taking on products that target underserved areas that are of interest to hackers: powerful SBC (with PCIe), Arm64 laptops, non-Android phones, watches without the back glued on.
So I'm not skeptical because the Rock64Pro can't be an adequate router. It can. It is simply that I don't the unmet need. There are are already a ton of great platforms to run the various open source router operating systems on and those platforms have been available to us for nearly 20 years now.
So the question about the router proposal is what new things it brings to the table?
Personally I think the answer could be "lots of PCIe slots". In other words a "router" implemented using a daughter board containing a 4-to-1 PCIe bridge can lead to some very interesting options. The most flexible form is a raw PCIe bridge with a chassis for low profile PCIe cards (network cards, M.2 converters for SSD). However that has some problems since connecting something like an M.2 4G modems would still be difficult (because there is no USB on the M.2 slot of a simply PCIe to M.2 converter). That might mean trading some flexibility for being a better router (maybe provide mostly M.2 and support one or two full height PCIe slots with a right angle riser to keep things compact?). Either way I think there has to be focus on how to plug in things to a PineRouter that can not already be trivially connected to an off-the-shelf router.
Whilst Pine64 direct sales/pricing strategy is undeniably important, I think Pine64 has gained a lot of attention recently by combining its sales model alongside taking on products that target underserved areas that are of interest to hackers: powerful SBC (with PCIe), Arm64 laptops, non-Android phones, watches without the back glued on.
So I'm not skeptical because the Rock64Pro can't be an adequate router. It can. It is simply that I don't the unmet need. There are are already a ton of great platforms to run the various open source router operating systems on and those platforms have been available to us for nearly 20 years now.
So the question about the router proposal is what new things it brings to the table?
Personally I think the answer could be "lots of PCIe slots". In other words a "router" implemented using a daughter board containing a 4-to-1 PCIe bridge can lead to some very interesting options. The most flexible form is a raw PCIe bridge with a chassis for low profile PCIe cards (network cards, M.2 converters for SSD). However that has some problems since connecting something like an M.2 4G modems would still be difficult (because there is no USB on the M.2 slot of a simply PCIe to M.2 converter). That might mean trading some flexibility for being a better router (maybe provide mostly M.2 and support one or two full height PCIe slots with a right angle riser to keep things compact?). Either way I think there has to be focus on how to plug in things to a PineRouter that can not already be trivially connected to an off-the-shelf router.
