Slashdot Log In
First RFC1149 Implementation
Posted by
michael
on Mon Apr 30, 2001 05:45 AM
from the don't-try-this-at-home dept.
from the don't-try-this-at-home dept.
Crossfire writes: "IP over Avian Carrier (RFC1149) was just a joke, right? It would seem not. The Bergen Linux Users Group has made it a reality! It would also seem that Alan Cox was present for the event too, given by the photos on the page." This is just excellent. Kudos to everyone involved.
This discussion has been archived.
No new comments can be posted.
First RFC1149 Implementation
|
Log In/Create an Account
| Top
| 143 comments
(Spill at 50!) | Index Only
| Search Discussion
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
(1)
|
2
(1)
|
2
Re:wow (Score:3)
spoofing pigeons (Score:3)
:)
Improvements possible (Score:3)
1.1. Use barcodes instead of OCR, or
1.2. Use magtape instead of paper.
2. Transmit redundant packets to cut lossage.
3. Use better trained pigeons.
4. Secure packets against rain damage using cling film (saran wrap).
--
Combine with the 11 pound plane... (Score:3)
Ok, so the ping times are gonna suck and having to wait months for a new packet to be constructed if the previous failed also isn't too good.
"How come I didn't get that email from Bob?"
"Oh, the model plane lost a wing half way across the ocean, sorry."
Geoff
CURSES! (Score:3)
Re:Just imagine... (Score:3)
It's just TCP/IP transport layer. You can add whatever encryption and authentication protocols on top as you like.
There is no theoretical difference between this transport and any other (ethernet anyone?) ... to the application it is all the same ... albeit perhaps a little slower ...
Re:Just imagine... (Score:3)
A DOS attack of the Avian carrier. Clip their wings, perhaps?
I would picture a DOS as being where you send *armies* of your own pigeons to the secondary base camp to elbow out the legitimate pigeons from dropping their messages.
Perhaps a herd of cats invading the receiving base camp might deter packets from landing, and cause bad pings.
Buggy protocol (Score:3)
I just loved the ping reports:
$ ping -i 450 10.0.3.1
PING 10.0.3.1 (10.0.3.1): 56 data bytes
64 bytes from 10.0.3.1: icmp_seq=0 ttl=255 time=6165731.1 ms
64 bytes from 10.0.3.1: icmp_seq=4 ttl=255 time=3211900.8 ms
64 bytes from 10.0.3.1: icmp_seq=2 ttl=255 time=5124922.8 ms
64 bytes from 10.0.3.1: icmp_seq=1 ttl=255 time=6388671.9 ms
9 packets transmitted, 4 packets received, 55% packet loss
round-trip min/avg/max = 3211900.8/5222806.6/6388671.9 ms
What happened with the 5 packets? Was the "carrier lost" or perhaps eaten? Or did it just "drop the packet?"
Re:I guess this one is for... (Score:3)
Especially RFC's about routers. Those can be ignored and your router will still work with all future network protocols.
Or perhaps, while one or two RFC's are meant to be funny, the other RFC's are not. RFC's are often the accepted standard though not always.
That may be very hard to understand at first but I think it's true none the less.
I wanna be a CPIP firewall! (Score:3)
No, I can see this working in the real world.. (Score:3)
So we translate all our communication data into TCP datagrams.. print them out.. stick them in envelopes... write the zip code and house number on the outside.. genius!
ah, no wait, that's a letter isn't it.
Al.
sapstuff.com [sapstuff.com]
Re:If Slashdot used these thing (Score:3)
Methinks they would get splatdropped...
packet life? (Score:3)
This gives a new meaning to TTL for packets. =) How many uses do you get out of one of those poor pidgeons? It'd be interesting to see network traffic by looking up and seeing how big the cloud of birds is. What happens to the bird when you have a resend or a NAK?
or
or
Have to watch out that you don't get cats sniffing the packets or that the packet falls out of the sky from heat exhaustion. Seriously, this is an example of doing something because it could be done. What good is this in an age were people turn their nose up at 300 baud? They could have put that bird to some use afterwards and eaten it. =)
perfect chance.. (Score:3)
we can destroy the internet as it is know, and use these little pigeons (maybe penguins that fly would be better) as our tunnel...
this way, everything would be almost to slow todo, except for the old schoolers that are used to there old 200bps modems..
now, everyone else decides to drop the internet, and we only have several people using it now..
then we will let it grow... but what will happen? i reckon itl end up like it is now.. again.
(a) some people see a chance for it making money
(b) they set up a 'bsp' or bird service provider
(c) this bsp has a user base of about 20
(d) several other small bsp's start. they hub together in a very unordered and unplanned fation, where each bird goes to a particular address, in its memory. unfortunatly the memory of pigioeons only holds 5, pengiuns actually hold 7.
(e) someone decides that the Australian's (me) want bird access.. and trains some pigeons to travel between us and au. Unfortunatly, most pigeons die because of the over loaded packages they are carying, bird loss is at 90%
(f) after a decade or so, despite its slowness, bird net is rocking along, with a whole 1000. Since the memory of a bird is maxed at 7 address', they have to start again. and where back to square one.
RFC Author (Score:3)
A flying start (Score:3)
Other applications... (Score:3)
How ironic (Score:3)
Reuters implemented this in the 1800's (Score:3)
Re:RFC Author (Score:3)
Yes, he has been informed. Or maybe he found it here (or somewhere else himself).
According to David Waitzman himself, this is indeed the first implementation of his "joke" RFC. (However, someone in Australia asked him about it some years ago, but he never heard back from them again, so it's likely they cancelled.)
Re:Will Microsoft use falcons ... (Score:3)
Or they'll install doors at the end stations which only allow Microsoft pigeons through
If Slashdot used these thing (Score:3)
I certainly feel sorry for anyone who get slashdotted
I tried to look at 10.0.3.1 (Score:4)
Re:Reuters implemented this in the 1800's (Score:4)
Re:Bandwidth is not the problem, but latency is... (Score:4)
Re:Just imagine... (Score:4)
I guess you could rig the package to explode if tampered with. Maybe for long distance, we could use a modified avian carrier such as an albatross.
Troc
Great, someone ate the packets again.
Re:Buggy protocol (Score:4)
When they finally broke off from the flock, they all broke off more or less simultaneously. This created some chaos at the other end. Simply said, they forgot to close the door of the pigeon cage, so the 2 last pigeons of 6 escaped.
We could have had 6 replies in maximum, but I had to let it go to 9 packet sent to get all the replies back. 8 of those packets was sent to the remote sent, for the 9th, we didn't have a pigeon to carry it. And alas, loosing two pigeons at the remote site, we only had 4 return pigeons.
- Vegard
Mirror. (Score:4)
bandwith that can withstand som slashdotting.
However, there is a mirror at http://www.pvv.org/rfc1149/
- Vegard, member of BLUG and the CPIP WG.
Re:perfect chance.. (Score:4)
what?!? I think 200 Birds Per Second is pretty fast!!
Re:packet life? (Score:4)
Not a problem actually - TCP/IP takes care of packet order and loss.
This could actually be a genuiinely useful form of wartime or covert communications, as long as you use encryption on the link. Just imagine anyone trying to make sense of all the encrypted TCP/IP packets (literally) flying around, and imagine their frustration when any packet loss they manage to cause doesn't affect the reliability of the communication channel.
Coming soon hamster-IP for your ground-based communications!
Similar, more monolithic method (Score:4)
We called it "HCP/IP."
-Omar
Of course it's secure... (Score:4)
BTW, why are we seeing so many posts moderated so high recently? I only want to see a few posts per article, and now you get 15 at +5. Did Taco make everyone a moderator full-time or something?
not_cub
Re:On closer observation... (Score:4)
There is always room for improvement over existing protocols.
Re:Open Source? (Score:4)
A Hot fix is due out next week.
Re:I guess this one is for... (Score:4)
Let me say this once before everyone goes nuts:
I never said that RFC == standard.
Read what I wrote. I said "well on its way." I did not say "definitely will become." Just because I did not explicitly say "but may not reach that point" doesn't mean it that isn't implied.
RFCs tended to be well documented protocols and procedures that tend to head towards standards or at least widely used methods. Most protocols never even reach this point. If a person or group writes an RFC, they believe they have something worthy of a larger audience.
And yes, I am aware of the multitude of humorous standards in there (IMPS, RFC 2795 [ietf.org], comes instantly to mind, RFC 1097 [ietf.org] or "subliminal telnet messaging" being an earlier one).
Still, my point in that post was that many RFCs are widely used as if they were standards even though they are not stands. Internet Relay Chat is RFC's 1459, 2810, 2811, 2812, and 2813, all marked "Experimental" or "Informational". Their headers do state they are not information standards, but this has not stopped over 10 IRC networks, dozens of client programs and tens of thousands of users from using them. Likewise, RFC 1413 [ietf.org], a.k.a. the ident protocol, has been a proposed standard for seven years, yet is included in every UNIX-based operating system. Your secure shell products (SSH) use a protocol that has a working group [ietf.org], but they have not even reached the RFC point in the process!
Just because someone says something is not a standard does not mean it is not widely adopted. Personally, I want to implement RFC 2324 [ietf.org], better known as the Hyper Text Coffee Pot Control Protocol :)
Nice music.. (Score:4)
I can see it now.. Someone just started a bird-farm, instead of a server farm, to index MP3 requests..
wait..the RIAA has already sued the bird-farm owner..
Gives me a great new idea. If each one of us had 4 pigeons, and we pool them together, cant we beat Freenet or Gnutella ?
Did they... (Score:5)
Would a Denial of Service attack ... (Score:5)
Re:Sorry... (Score:5)
Re:Limitations (Score:5)
/Janne
Innovation (Score:5)
Alan Cox (Score:5)
Pictures fro mthe last event are available here [ii.uib.no]
Re:Will Microsoft use falcons ... (Score:5)
- They make a big stink, and a lot of noise, but aren't really effective at anything but dying.
The process will be simple: MS will purchase several of the air cannons used to launch frozen poultry at aircraft windows to test for breaking point, and use those to launch the MS TIEE (Turkey Internet Enhanced Enabler) stack at any and all CPIP packets which get too close to Redmond, or any other MS campuses. And probably any QWest property now as well.--
Spam? (Score:5)
--
Re:Bandwidth is not the problem, but latency is... (Score:5)
Not to mention out-of-order packet reception. I think we could have a rather large issue hunting for pigeon #35431 in the Great Flock. And of course, the risk of packet loss is much higher. Farmers out in their fields with shotguns have a lot lower chance of disrupting your ethernet connection but could take out pigeons #234, 54245 and 6644 with one good blast of 00 buck. Just imagine of the retransmission requests.. *shudder*
Suckey as it is, I'll stick with my csma/cd, thank you. Though it doesn't have that soothing 'coo.' :)
On closer observation... (Score:5)
I propose that once testing has been completed on a fully standards complient version of RFC 1149, testing on the implementation of RFC 2549 [ietf.org], or "IP over Avian Carriers with Quality of Service", should begin. This extention of RFC 1149 adds many important features, such as quality of service, security, and traffic shaping.
Man-in-the-middle-with-a-big-net Attacks (Score:5)
Re:Open Source (Score:5)
Take a bird, open some sauce and
then cook it.
Bandwidth is not the problem, but latency is... (Score:5)
As the saying goes: there is no higher bandwidth link than a UPS truck fully loaded with DVDs...
Re:Bandwidth is not the problem, but latency is... (Score:5)
TCP/IP is perfectly able to deal with packet loss and re-ordering, so no problem there. Of course, this might cause some performance drop, unless selective acknowledge (SACK) is used.
> Farmers out in their fields with shotguns have a lot lower chance of disrupting your ethernet connection but could take out pigeons #234, 54245 and 6644 with one good blast of 00 buck.
True enough. However, consider this: roadworkers have a lot lower chance of permanently stopping your pigeons from flying but could take out your spiffy new phiber optic link with an appropriately placed "dig" of their backhoe... And don't talk about backup connectivity: everybody knows that usually the backup fiber runs within the same duct as the primary.
These are not dropped packets... (Score:5)
Re:Sorry... (Score:5)
Another time related problem is the speed of attaching the datagrams to the pigeons and then reading them. We need pigeons with detachable legs so we can easily pop the data on and off of the transport mechanism (a bit like a physical implementation of mbufs, the pigeon being analogous to an Ethernet frame).
We could try little rockets or shooting them out of cannons (reduces latency) but this can lead to excessive packet loss if you try to push the birds past their limits. Routing is also an issue, you've got to give them some initial aim otherwise all that energy is just wasted. Topological routing is now appearing which could come in handy for this but does anyone have a driver for a turret mounted pigeon cannon handy?
Re:Mirror. (Score:5)
> bandwith that can withstand som slashdotting.
Not if they are using CPIP!
Ping Flood (Score:5)
(don't worry, not a goatse.cx link..)
zsazsa
Just imagine... (Score:5)
I also suspect CAT might cause a segment fault.
Sorry... (Score:5)
-if not me, someone else.
---
Open Source? (Score:5)
Potential security problem (Score:5)
That's not a packet Alan Cox is holding! (Score:5)
Pic here [linux.no].
"Hey Alan, stop bogarting. Pass it!"
Limitations (Score:5)
- Packet loss - hunting season could cause major disruptions to services.
- Lack of broadcast/multicast support
- Limited packet size
- Mandatory source-routing could be a problem (I always configure my Linux kernel to drop source routed packages)
- Mating season could cause abnormal round trip delays
On the positive side, packet collisions are history - unless someone implements CPIP in a LAN that is...Will Microsoft use falcons ... (Score:5)
Re:Buggy protocol (Score:5)
inet addr:10.0.3.2 P-t-P:10.0.3.1
collisions:0
RX bytes:88 (88.0 b) TX bytes:168 (168.0 b)
You know, I never thought of packet collisions as really funny. Now, somehow it seems much more entertaining.
-dbOpen Source (Score:5)
This is a great example of community acting in a fun and open way.
Not for greed, but because its fun.
A sense of community that I don't see in the rest of my life (Where I live, where I work).
It makes life worth living when you enjoy it.
Are you on the Sfglj [sfgoth.com] (SF-Goth EMail Junkies List) ?
Faster than a 300 bps modem :-) (Score:5)
We need RFC1149.b... (Score:5)
Re:Will Microsoft use falcons ... (Score:5)
The Pigeons vs Windows (Score:5)
A funny little anecdote: One of the last frames (i.e. pigeons) we released actually crashed into a neighbor's bathroom window, after which Alan Cox himself commented; "Oh no, windows causing problems again." - Just thought I'd share that. :)
(BTW, I'm not officially a BLUG member, but was cordially invited by a friend of mine - thanks Karlmag! And hi mom!;)
Re:Would a Denial of Service attack ... (Score:5)
Upgrading to RFC1149 (Score:5)
--
McCarrum!
Re:Bandwidth is not the problem, but latency is... (Score:5)
I think the quote is:
"Never underestimate the bandwidth of a station wagon full of tapes hurtling down the highway." -- Andrew S. Tanenbaum - Computer Networks
Re:On closer observation... (Score:5)
Nope, you are wrong: the duct tape is only a informal suggestion. Otherwise the statement would be: "A band of duct tape MUST BE used to secure the datagram's edges"
SCNR :-)