Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
The Internet Networking IT

World's Smallest IPv6 Stack By Cisco, Atmel, SICS 287

B Rog writes "Cisco, Atmel, and the Swedish Institute of Computer Science have released uIPv6, the world's smallest IPv6 compliant IPv6 stack, as open source for the Contiki embedded operating system. The intent is to bring IP addresses to the masses by giving devices such as thermometers or lightbulbs an IPv6 stack. With a code size of 11 kilobytes and a dynamic memory usage of less than 2 kilobytes (yes, kilobytes!), it certainly fits the bill of the ultra-low-power microcontrollers typically used in such devices. When every lightbulb has an IP address, the vast address range of IPv6 sounds like a pretty good idea."
This discussion has been archived. No new comments can be posted.

World's Smallest IPv6 Stack By Cisco, Atmel, SICS

Comments Filter:
  • by Anonymous Coward on Wednesday October 15, 2008 @03:45PM (#25388111)

    You kid, but remote controlling consumer loads is indeed on the table. The idea is to make it less troublesome to turn the grid back on after a power failure by remotely disabling big consumers and reenabling them one at a time. The technology is based on powerline communications and uses the electric meter as a gateway.

  • Re:Sweet (Score:0, Informative)

    by Anonymous Coward on Wednesday October 15, 2008 @03:48PM (#25388167)
    what, you mean slashdot this: http://www.c64web.com/ [c64web.com] ? Cheers and My apologies to the website owner
  • by mustafap ( 452510 ) on Wednesday October 15, 2008 @03:56PM (#25388315) Homepage

    Any idiot ( even you ) could do a simple stack in much less than 11KB. It just depends on how much functionality you want to drop.

    I've an IPv4 webserver running here:

    http://mikehibbett.dyndns.org/ [dyndns.org]

    that's running* a Microchip stack on a PIC micrcontroller in about 16KB of code. I bet I could get that down to less than 1KB if I knock much of the functionality out. Want to have a bet on it?

    * it's running now. Not sure what a slashdotting would do to it.

  • Re:Linux? (Score:3, Informative)

    by c_oflynn ( 649487 ) on Wednesday October 15, 2008 @04:01PM (#25388411)

    Wrongo. The USB stick plugs into Linux and shows up as a network interface, instantly working with Linux (or Windows) IPv6 network.

    There is a quick demo at http://www.youtube.com/watch?v=yjztYx_F2Ko [youtube.com]

    And you can browse some of the documentation. It is fully IPv6 compliant, hence should work with anything...

  • by TheRaven64 ( 641858 ) on Wednesday October 15, 2008 @04:16PM (#25388695) Journal
    Note that this 11KB can go in ROM, and only 1.8KB needs to be RAM. The same guy wrote an IPv4 stack that is much smaller, but IPv6 is a huge protocol comparatively. Considering that the MTU size for most of the Internet is 576 bytes, implementing TCP on IPv4 or v6 in 301 bytes for RAM would be tricky...
  • by ChrisA90278 ( 905188 ) on Wednesday October 15, 2008 @04:49PM (#25389393)

    Why would anyone want to have a light bulb with a data connection?

    Car manufactures want this. Have you ever seen the wire bundles in cars? If the lights had network address all you'd need one just ONE wire ad the switch on the dash would broadcast the "turn on" command. Every light bulb and every switch could have just one wire.

    In your house it is the same. Light switches would simply issue commands and the light woould wait for a command to turn off/on. Housewiring would be hugely simplifed. and then automation would be easy. Room sensors could detect if people were there and kill the lights. All of this without mods to existing building.

    Think of how much you could save if a 20 story building used network lights? Millions.

  • by Anonymous Coward on Wednesday October 15, 2008 @04:51PM (#25389439)

    Actually that's a pretty bad idea....

    Although it might be useful to generate a report of which lightbulbs have burned out, it won't save any money because the biggest cost is labor. You know, the fool with the ladder. That's why bulbs are changed on a schedule, and all the lights in an area are changed at the same time, preferably before any (or many) of them go out. It won't make economic sense to wait until a bulb dies and then rush out to change it. Even when the price goes up by an order of magnitude when smart electronics are included.

  • Re:Lightbulb? (Score:3, Informative)

    by Machtyn ( 759119 ) on Wednesday October 15, 2008 @04:52PM (#25389457) Homepage Journal
    A firmware is an OS... an operating system. It operates the system for which it was designed. The firmware, if easily upgraded, is also easily hacked. If the thing has an IP address, it also means it has some remote interactivity, even if it is as simple as receive read (no write/execute) request, send data, buffer overflows can still be exploited, DDOS's can be used, etc. So that is another intrusion vector.
  • by cyclomedia ( 882859 ) on Thursday October 16, 2008 @05:30AM (#25396765) Homepage Journal

    >But I don't want them addressable from the world at large;
    >I want to go through a controller that verifies access and coordinates changes.

    You would already have this at home anyway, in the form of an LDAP server or at least coordinated LDAP rules. E.g. in my house the Xbox can read the network shares to play MP3s and AVIs but it can not delete them, this is controlled via LDAP and protects against hard-wired attacks, so even someone visiting who plugs their laptop into the socket behind the sofa can't delete those MP3s. And instead of a subnet mask you'd have a db of authenticated IPs and MACs alongside your basic user/password.

    You would also likely have a wireless setup too, not all the devices on your future home network (e.g. the lightbulbs) will be hard wired so they'd need authentication *anyway* to know who's allowed to turn them on and off. Instead of logging in and configuring each one you'd just have a button-press system, plug it in, turn it on, press the physical plastic "register" button on your main server box and it'll be set up. Kind of like the way you tether a cordless phone to its base station but with Public/Private encryption and mac/ip filtering. This controls wether or not the neighbors can get on your LAN at all, before we even start thinking about read/delete rules on shares.

    So there's already security in place, and the next logical step is to make these transparent and transient. At the moment when unplugging your laptop and taking it around to the neighbors you might get your home wifi still and be able to get on your LAN but what if you could visit the in laws in the next county, plug same laptop into *their* RJ45 socket and STILL be on your LAN?

    Remember - your LAN already has security, and an internet facing NAT+firewall will not protect you against a malicious wifi or plugged-in client in your house or 100m away from it.

    Back to the in-laws: In IPv4 space your laptop has to negotiate their router via UPnP (assuming they have it turned on) then find your home router via dyndns (which you have to use a demon to continually update) then connect to the only open port (which you have to set up and configure) on there to be redirect to a server inside your lan (which you've had to install and set up for this task specifically) to use some kind of virtual console or VNC session (which you have to setup and configure on your server and any client you expect to possibly take outside your house) to do anything you'd just be on your LAN, kind of.

    Again, remember, you *already had to do the configuring of your security rules* just to protect your house and 100m around it anyway, everything in the previous paragraph is just a pain in the ass *on top* of that.

    Instead, in IPv6 space your laptop will always be on your LAN - if you leave it on the train logge in (dumbass) you could revoke its MAC/IP from your LDAP as soon as you realise you've lost it, probably from your mobile phone. Your shares, printer and fridge will all visible just as if you were wired in in your lounge or wirelessly connected from the garden no matter where you are on the internet because the internet will be transparent, it's the way it's supposed to be.

  • Re:Sweet (Score:3, Informative)

    by Talderas ( 1212466 ) on Thursday October 16, 2008 @07:12AM (#25397435)

    Thermometers could have network access to report temperatures to a central server. You're talking about potentially simplifying devices to measure temperatures, or devices which traditionally never had network components because of power or other concerns. A hospital could use those chips in their digital thermometers to report a patients temperature to a server so that it's automatically timestamped and loaded into the patients record. With a push towards digitizing patient records, it would provide a simpler interface of inputting data.

    With light bulbs it doesn't seem as though it would be useful, since in most cases if a light bulb is broken, no big deal we'll replace it when it when we get around to it. However think about this from a security context. What if the light bulb can report when it's on or off? Combine that with motion sense lights, or just even tracking whether lights are on in a building could provide an opportunity to increase the level of security. The other option for use would be if the light bulb needs to be continuously on but typically isn't seen often by human eyes. If the light bulb is broke and not putting out light, then the chip reports that it's broken so that a maintenance person can get down there ASAP to replace the light bulb before the monster escapes.

FORTRAN is not a flower but a weed -- it is hardy, occasionally blooms, and grows in every computer. -- A.J. Perlis

Working...