There are some good open source software router projects around: VyOS, BSDRP, FREESCO and others.
All of them have quite comprehensive soft capabilities but in other side, all of them have one common problem: lack of standard hardware they’re running on.
Every admin who decided to use open source in networking should build own unique PC-based system. It’s not good. It’s bad, I think. Mainly because of lack of community around such complex solutions, lack of docs and lack of any end-user support (imagine that one nice day admin said “bye-bye” to the company and what? Who will support his systems? Next admin? No. He will remove them and build his own “network puzzles”
But anyway, some companies uses Vyatta as a software for their networking hardware: Brocade, Ubiquiti, may be some others. But all of them offer hardware. Some standartizied hardware that has common ideas, common structure, common use cases. And therefore it can be supported not only by company itself but by wide community of end-users.
And now, imagine that we have abilities to create a hardware for open-source modular router. Imagine that this hardware will support not only standard 1000Base, but 10G, 40G, SDH channels and so on. Imagine that it will support special hardware to speed-up routing tasks, to support multiservice (i.e. PBX stacking, etc), and many-many other things. Oh. I know. It’s hardware will support OpenBTS project integration and be able to run corporate GSM base station atop of it (hey, Cisco, are you scared???)
So, in this case we’ll have a standard open hardware that runs open software (VyOS Multiservice Community Edition, of course :). With end-users community around it and all other advantages.
So, that’s my opinion. I’m interested in your mention of this topic. How do you think, will the open-source hardware project of such [imaginable] router be demanded?