From Leonard Erickson · 1996-04-12 09:19 · #73180 in FREE25
In reply to How a new nodelist would work..... by Jason Steck
-=> Quoting Jason Steck to Leonard Erickson <=- JS> Tuesday April 09 1996 04:59, Leonard Erickson wrote to Paul Edwards: LE> IP addresses don't have "implied" routing. They are network addreses. JS> They have an implied routing when a subnet mask is applied and one or JS> more gateways/routers are defined. That's all I was talking about. JS> On a simple TCP/IP network, the subnet mask is applied and if the JS> destination address is not in the local subnet, then the packet is JS> sent to the defined router for the destination address of the packet. JS> That gives IP a sort of "implied routing" in the informal sense. But that "routing" is not an artifact of the *address*. It's an artifact of the local routing strategy. Just as the routing a phone call takes depends on the routing options your phone company has set up. The *address* doesn't imply the routing. LE> Sounds like "hub" routing. A quick check shows that mail to you will LE> "default" route to Hub 800 (Dave Hatch) in your net. JS> One of the things about existing technology is that no OUTBOUND JS> default routing is assumed. The /0 utilization is intended only as a JS> default INBOUND route to any node. That really *is* a "political" decision. It's based on the decision that if there are long distance calls made, then the *sender* should make them. Also note that "/0" nodes are not required to provide inbound routing. Only *Host* nodes are. Zone and Region nodes are /0, but they are *not* required to provide inbound routing. But Hub nodes are. Thus, it's besty to keep /0 out of it. JS> A new system would best let each node entry independantly define JS> default inbound and outbound routing points. This would allow for a JS> very flexible, MASSIVELY powerful amatuer routing system. It would be JS> almost impossible to "break" the system since interconnectivity would JS> be so incredibly high. And it'd only work as long as the node kept that info up to date. Which, given my experience as both an NC and an RC in a couple of othernets, is not as likely as you think. Heck, when I took over as RC in one net, I found that one third of the region's *phone numbers* were wrong! I was checking because I needed to replace a NC in one net and couldn't get thru to the other nodes. After turning on the speaker, I found that I was getting "disconnected" recordings. So I looked in the Fido nodelist and found a couple of the sysops. They were able to dig up the rest for me. And then I checked the other nets in the region. Also, what happens when someone deliberately lists a wrong node as their inbound, and then does something guaranteed to generate tons of mail (like echobombing an echo). You can shut down a node that way. ... My other computer is an abacus. --- Blue Wave/DOS v2.30 * Origin: Shadowshack (1:105/51)