On Thu, 2013-07-25 at 10:45 -0400, Bryan Fields wrote:
I have no interest in limiting myself to what some old ignorant hams can understand. This is one of the reasons I got out of radio for a number of years, as it was not new or challenging due to these guys. I want to sit down an study if I don't understand it, not gripe about it (and my angina) on FM.
Ignoring a problem doesn't solve it either. Just yesterday on the PBBS scene someone made a comment on how the amprnet doesn't appear to be of much use - not in those exact words but that's the impression he left in his message. I highly disagree with him on that as the way to keep packet in general fresh is to provide more and unique L7 to the ham community. The most cost-effective way would be through IP based apps as IP is common amongst almost every cpu-based platform. Not every elderly ham is or wishes to be ignorant to our technologies either... it takes specialized people who can relate to them in simpler terms to help them understand. We should view such scenarios as a challenge not as a deterrant to our progress.
We really have two discussions here. One is global BGP peering and a 44-net back bone, the other is the end user connecting over tunnels or dedicated links to the backbone.
End users won't want to connect if there's nothing unique we can provide them at the L7 level... so then all this work would achieve what goal?
The backbone should be based on standard routing protocols and be vendor agnostic (ie ALU/CSCO/JNPR), as I could see having a router and peering link donated, but if we need to rack a linux box, we'll have an issue.
I don't discount any of the network discussions that have been taking place, and many of the suggestions are viable solutions if the goal is for us to be some sort of glorified ISP backhaul, however we don't *need* the amprnet to do that. If we are to include something to encourage end-user ham usage at the lowest cost possible to them, than that's what we should be looking at. I totally agree with you that the protocols need to be vendor agnostic.
The individual backbone POP's can interconnect with the end users over any method they want. (I'd say we need to support IPIP or GRE tunnels as standard at any POP)
I see the backbone POP's serving the gateway hosts in the areas they are in... more along the lines of a border-router. The gateway hosts should be the edge router (so to speak) to their area.
For an end user wanting a /29 or /31 they would pick a "local" AMPR POP, and link in with them.
Agreed, thus making the gw host more of an amprnet edge router... but again, of what benefit is there to the end user who may be using 2m, 70cm, or even 2.4Ghz to link into the amprnet that they can't already do on commercial internet?
Here in Connecticut, the State Police amateur club have constructed a TheNet based network State Wide all linked in via microwave on the backhaul for the single purpose of providing a state-wide service just for WL2K. They've been very successful _without_ the need of amprnet connectivity to achieve this goal (sponsored in part by Homeland Security).
I've been lucky enough to have the admin of this system test drive a box I'll be configuring for them with RMS, axMail-Fax, URONode, FBB, and a webmail interface which will be available via the amprnet... I hope to have this project complete by autum. This would be something unique offered at the L7 level.
We need to "sell" (for lack of a better word) the amprnet to end users. A lack of patience with 'ignorant hams' allows us to shoot ourselves in the foot.
Thoughts?
I'm all for the idea on creating the backhaul for a way where we can eliminate a single-point-of-failure whether it be BGP, ISIS, OSPF, the protocol to me is irrelivant... I would just hate to see all this work done when we can't gain any end users. Remember, without end users, ISPs go under!
</$0.02>