Reverse or inverse ARP from windows/linux - no way (!?!?)
in [Design]
|
From: markp on 29 Mar 2010 12:59 "Didi" <dp(a)tgi-sci.com> wrote in message news:851c373c-9078-4880-aca1-18fb3e83389e(a)g19g2000yqe.googlegroups.com... On Mar 29, 6:32 pm, "Boudewijn Dijkstra" <sp4mtr4p.boudew...(a)indes.com> wrote: > Op Mon, 29 Mar 2010 17:10:34 +0200 schreef glen herrmannsfeldt > <g...(a)ugcs.caltech.edu>: >> >> In comp.protocols.tcp-ip markp <map.nos...(a)f2s.com> wrote: >> (snip, someone wrote) >> >> Some will respond to a ping to the broadcast address, others not. >> >> Some will even respond differently to the subnet broadcast address (e.g. >> 192.168.0.255) than the global broadcast address (255.255.255.255). >> > Hmm, do you still think it is "simple" (just kidding :-)). :-) OK, the ARP way sounds like you've got to send a ping first to establish the route. However, since the only things you are worried about are your web-servers you can guarantee they respond to that as you would have programmed them too. As an aside, wouldn't the DHCP server when allocating addresses in the first place do that automatically? The original idea of sending a UDP broadcast enquiry packet and getting replies is quite simple though. Only your servers would reply, and you could even send the desired server's MAC address in the enquiry packet so only one device that matches it would respond. Mark.
From: Didi on 29 Mar 2010 13:16 On Mar 29, 7:59 pm, "markp" <map.nos...(a)f2s.com> wrote: > "Didi" <d...(a)tgi-sci.com> wrote in message > > news:851c373c-9078-4880-aca1-18fb3e83389e(a)g19g2000yqe.googlegroups.com... > On Mar 29, 6:32 pm, "Boudewijn Dijkstra" > > <sp4mtr4p.boudew...(a)indes.com> wrote: > > Op Mon, 29 Mar 2010 17:10:34 +0200 schreef glen herrmannsfeldt > > <g...(a)ugcs.caltech.edu>: > > >> In comp.protocols.tcp-ip markp <map.nos...(a)f2s.com> wrote: > >> (snip, someone wrote) > > >> Some will respond to a ping to the broadcast address, others not. > > >> Some will even respond differently to the subnet broadcast address (e.g. > >> 192.168.0.255) than the global broadcast address (255.255.255.255). > > > Hmm, do you still think it is "simple" (just kidding :-)). > > :-) > > OK, the ARP way sounds like you've got to send a ping first to establish the > route. Nope. Just a single ARP packet type 3 (if RARP misused, not sure which it has to be for inarp - but same thing) to the known MAC destination, which will reply in a single packet to the originating MAC address. No loop, no broadcast involved. > However, since the only things you are worried about are your > web-servers you can guarantee they respond to that as you would have > programmed them too. As an aside, wouldn't the DHCP server when allocating > addresses in the first place do that automatically? If you mean would not everyone on the subnet be aware of everyone elses IP & mAC address data by snooping the dhcp, then no. Only the dhcp request is broadcast, the rest goes peer to peer (and may well be invisible to much of the subnet even in promiscuous mode). > The original idea of sending a UDP broadcast enquiry packet and getting > replies is quite simple though. Only your servers would reply, and you could > even send the desired server's MAC address in the enquiry packet so only one > device that matches it would respond. Yes, of course it is simple. I just want to use the wintel PC's as TV-sets (i.e. I am not delivering any software for them, my software runs under DPS on the device). In fact this works so far, I am just preparing for the day I have this issue at a customers site. Dimiter (the thread got lengthy, here is a third link to the device we are talking about: http://tgi-sci.com/tgi/nmctb.htm )
From: robertwessel2 on 29 Mar 2010 15:32 On Mar 29, 10:16 am, glen herrmannsfeldt <g...(a)ugcs.caltech.edu> wrote: > In comp.protocols.tcp-ip Didi <d...(a)tgi-sci.com> wrote:> On Mar 29, 7:37?am, "JosephKK"<quiettechb...(a)yahoo.com> wrote: > > (snip) > > >> APR lower level than UDP? ?I don't think so. ?Same > >> level at best, below UDP is MAC and PHY only. > > Uhm, not so sure. Same level as UDP in that they both have the > > Ethernet protocol type set to "internet", perhaps; but the > > data inside a UDP packet are encapsulated into that level, > > whereas the ARP data are purely Ethernet encapsulated. > > This should put it one level lower - at least unofficially? > > MAC at layer 2 (ethernet switch level), IP at layer 3, > TCP and UDP at layer 4. > > Sometimes I think that ICMP should be layer 4 (as in ping), > other times in layer 3. It does go inside an IP packet. An ARP packet looks absolutely nothing like an IP packet. > ARP does not have the IP (X'0800') ethernet type, so it > seems that it should also be layer 3. ARP, in most senses, is not technically part of the TCP/IP stack at all, except that it's needed by some types of links. No more than Ethernet is really a part of TCP/IP (although it is used by many implementations). While ARP is defined in an RFC, there's plenty of stuff in the RFCs that not really part of TCP/IP proper. In any event, ARP, as defined in RFC826, supports non-Ethernet format MAC addresses, and non-IP format network addresses. In fact, the format of the hardware and network addresses, and their lengths, are specified in the ARP packet for each query and response. And to the OP: does the OS on this box have any support for the Windows networking protocols (Samba, if this is *nix)? If you enable that, and ID the node as a workgroup node with some easily recognizable name, then the users should be able to find it in Explorer (somewhere under "My Network Places"). If you published a share on your node you could put a single html page with a link (or perhaps a redirect, or just a display Im at IP 1.2.3.4) to the correct IP address (that redirect page would be recreated each time your device was assigned a new IP address).
From: Didi on 29 Mar 2010 15:54 On Mar 29, 10:32 pm, "robertwess...(a)yahoo.com" <robertwess...(a)yahoo.com> wrote: > .... > And to the OP: does the OS on this box have any support for the > Windows networking protocols (Samba, if this is *nix)? If you enable > that, and ID the node as a workgroup node with some easily > recognizable name, then the users should be able to find it in > Explorer (somewhere under "My Network Places"). No, DPS is not unix. But thanks for the hint, I'll look into that "Samba" thing. If not too complex I certainly could add some (probably partial) support so the name becomes visible in "my network places" or sort of, this would be perfect as a result. Dimiter
From: Hans-Bernhard Bröker on 29 Mar 2010 18:44
Didi wrote: > I tried today to figure out a simple way to give users of our > new netmca ( http://tgi-sci.com/tgi/nmcatb.htm ) to locate its > IP address once it gets one via dhcp when there is no internet > at the moment I believe that, strictly speaking, that can't happen. If you have no internet at the moment, you don't have DHCP either. Remember that DHCP itself is a UDP service. UDP in turn works on top of IP, and that, for better or for worse, is "internet". > Turned out there is nothing like an easy way to do that! Well, the problem is nowhere near as easy as it appears at first sight. It's called a "network" because it's _work_ to set up a properly functioning net. > How on Earth is that possible?! You'll want to look up "zero config networking". That's what the big guys came up with to address this very same issue. You'll see Apple mentioned rather a lot, for their "Rendezvouz"/"Bonjour" project. And let me point out I'm completely flabbergasted that nobody mentioned this before me --- not over here in c.a.embedded, anyway. I mean, come on guys: not a single owner of an Apple Airport base station speaking up, wondering what all these people keep talking about for days, when a "normal" WLAN box just does the job??? |