Slashdot Mirror
Linux In the Family Room?
njcfm writes: "Ok, you've all heard and seen embedded Linux running in all sorts of useful devices; now with the backing of Intel it can run your home network too!" The story is about UPnP [?] for which Intel has released a Linux toolkit.
Until I have a robot that can make a good ham and cheese sandwich, I would have no use for a remote microwave... or a toaster (what, do I keep it filled with bread (getting stale)?). Windows, shades/drapes/blinds, lights... those sorts of things seem useful, but most kitchen appliances require human intervention to work anyway (prepare the food and place it in the unit - turn over halfway through - baste - etc...).
What would be the advantage of controlling my uWave from my office? Maybe if it had an automatic popcorn loading mechanism, but even that requires that you monitor the bag closely during the end of the popping. Decent quality video/audio streaming, networking and cooking from a $50-150 device? Hmmm...
"It's tough to be bilingual when you get hit in the head."
Platform independance.
;)
Okay, forget PnP, UPnP, whatever.. What we need is the ability to have universal drivers. I know it's probably been said a million times already, but untill such a beast arrives we will still be in the dark ages.
My thought of device independant drivers verges on the line of the idea behind Java. Every system has things that are in common with all other systems - something plugs into something else (be it a chipset, add-on card, etc) to make something happen (RNG on P2 mobos, sound card, video card, network card, etc). There would be a platform level loader that provides a common, standard interface to a higher-level driver. This driver would be directly portable across any platform for which there is a written loader.
Currently its like:
OS -> Driver -> Device
The OS talks to the driver and the driver controlls the device. The problem being that the Driver has to be in a language the OS can understand as well as the device.
Here's what I thought (maybe from reading simealr stuff)...
OS -> Loader -> Device
|
Driver
The loader is called by the OS and uses the instructions of the driver to control the device. The loader is essentially an interpreter for the code in the driver, and since the driver is independant of the OS a company can release ONE driver and hit Linux, BeOS, Windows, whatever... yes, providing that the OS provides an interface for this.
Is there something like this in the works? Linux would be a great place to piononeer a "universal device driver"... base it on Java, or make something new...
Bah, whatever
(if there is something like this I would be very interested in reading up on it)
Price, Quality, Time. Pick none. What, you thought you had a choice?
One thing to remember is that with this comes some major security issues. There will have to be some serious security holes totally patched, if that is even possible.
A question for everyone... What type of work is being done, or needs to be done for home security when the house is on the internet. Is there a company who's working on this? I would hate to have someone breaking into the house by disabling my security system from half way around the world.
Another question... Does this mean a central computer will control heat, locks, water, power, and ever thing down to childrens toys? I know this would be a good idea, but once again would this be worth the risk? I think not having some of those "features" would be safer for everyone, once again, what does everyone think, or anyone know places with information on these questions?
--Brandon
All this really amounts to is standard distributed ARP/DHCP protocol. Standards are good ;)
When a device comes up, it broadcasts "Hey, what IP's are you guys on? I'd like 198.111.21.42 with a hosts entry of 'Jim's Fridge', but I'm flexible! Oh, and by the way, I'm on Ethernet now, but I can use my SiR transciever too! Do I need two IPs?"
The rest of the devices come up, find an unused IP and assign the name "Jim's Fridge" to their hosts file. The SiR capable machines make a note to activate a connection for devices with that discovery information.
Simple stuff. Microsoft can't even screw that up!
.sig: Now legally binding!
It is important that technology like this be incorporated into the design and building of the house of today (not tomorrow). Did you forget to trun the stove off? Check it and if so, turn it off, from the other side of the world.
Sure, it poses a threat that someone could crack your toaster to burn your bread every morning, but I would rather have open source software, where such a hole would be noticed quickly, than a proprietary/bianary bundle which will release their security patches quarterly.
I'd rather have someone respond than be modded up.
I have a private home network as it is. I just use non-routable IPs. I gateway out when I need to. When I get DSL, I will just install a firewall and NAT.
All this seems to provide is another way from companies to sell me something I already have.
Is there something here I am missing?
"Trademarks are the heraldry of the new feudalism."
What percentage of Joe users do you think would be able to install and/or maintain any sort of home network environment with a dozen or more connected devices? Remember: they will probably have all these devices connected to the internet using some sort of TCP/IP protocol as well, so they will have two networks running simultaneously with a number of different devices made by different manufacturers trying to communicate with each other.
haha...HaHa...HAHAHAHAHAHAHA...cough, wheeze...
Sorry...almost hurt myself there. All I know is it was tricky for me to set up a network in my house using four PCs sharing one cable modem. Networking, firewall, cable network connection through a gateway... Now I'm not a systems guy, but I ain't PC illiterate either. Most people can't even figure out how to add some memory to their machines, much less set up a firewall. I can see it now...the kid down the street is laughing his ass off as he turns your microwave on and off at 5 minute intervals and has your refrigerator order four dozen Tombstone pepperoni pizzas!
Protocols and OS network interfaces have a long way to go before more than 1% of US households have the capability to manage any such animal as all these guys are envisioning...
It's not funny till someone gets hurt.
Interesting, if lopsided, view.
:-) Chances are it won't matter to the home user, as it'll "just work" and work fairly well (if you believe the UPnP claims).
1) Just because a company with a history of some half-baked concepts is implementing it, doesn't mean it will be bad. MSCHAPv2 is "secure enough" to foil most attacks. And their proprietary Kerberos increases the security if Win2k networks greatly, even if it's as interoperable with the original Kerberos as a dead pigeon.
2) I doubt they will do a "member of IP address" thing you mention. Likely it some proprietary network, like a NetBIOS that isn't tunneled (such a beast existed back in the 1980s). The "UPnP" specification probably contains this and other information beyond simple "handshaking."
As for the security issues, this product is targetted at private LANs which speak a probably non-TCP/IP protocol. While I personally would tend towards public key handshaking with blowfish transport encryption (as well as notification of the user of most actions), I tend towards the paranoid side WRT my encryption and privacy
---
--
Internet Explorer (n): Another bug -- that is, a feature that can't be turned off -- in Windows.
I dunno, this sounds like Jini.
Can anyone comment on the relative merits of the two technologies?
IANAC (I Am Not A Coder) but this just sounds like a rehash of the same concept, except it's supposed to work with Intel hardware only.
--
The real Webmaven is user ID 27463. I don't rate an imposter, because my ID is such a lame-ass high number.
One question: Will this get around the "where's the device driver?" problem that Linux faces? Or will every uPNP device still need a specifically developed Linux driver? (I suspect that later.)
-- Diana Hsieh
-- Diana Hsieh
GeekPress: The Weirder Side of Tech News
Intel made the announcement at a UPnP summit at Microsoft in Redmond, Wash.
:)
Hahahhahaha. Enough said.
Photos of bits of the past hiding in the present: afiler.com
I don't like the idea... PnP has never really worked, it has always been a mess. Adding TCP/IP to PnP will only make it a distributed mess.
A printer must still have its OS-specific drivers. A scanner? same story. Yes, the peripherials will tell you where they are, and that's a plus... But I don't think our problems will end when everything is automatically recognized.
Much to the opposite.
But anyway, users want to be fed in the mouth... Why don't they include a slave techie with each sold system?
See the announcement here, and the meeting publicity material on upnp.org.
This definitely bring linux one more step closer to universal use. Anything that can conceal the underworkings of the OS and network configurations is a definite plus (so long as it can be easily bypassed by those who know what they're doing).
With Gnome and KDE you can have a predefined desktop GUI ready to go on installation. Now you'll be able to get your printers and scanners hooked up automatically. All that's left (if this hasn't been done already) is something similar to MS Internet Connection tool that acts as mini dhcp/dns/proxy server to automatically set up the network.
Realistically, if Corel's office2000 proves itself (or some other similar office product) and network set up becomes transparent, Linux should gain a huge jump in use.
Even the peripheral adding will give many home users a much easier time installing the system (I know I had a heck of a time getting my network card working with linux on my first install...)
Yes, it doesn't do much for people who can already network and do manual installs, but for people who are content with things to just work right with no fuss/no muss, this a *good thing*
Of course I use Microsoft. Setting up a stable unix network is no challenge
Could she use linux to run her house? Sure. My TiVo runs linux and has one of the BEST interfaces I've seen on anything.
Although the names sound familiar, UPnP is not related to Plug and Play.
UPnP is also not a Microsoft technology, there are many companies that support the UPnP forum, and yes Microsoft is one of the bigger supporters.
Similar to both HAVi and Jini, UPnP is just one of the computing industries attempts at simplifying the network of "connected appliances" that will be filling your homes in the future.
Sony pushes HAVi because they want the CE devices to be king.
Sun pushes Jini because they want to make licensing money.
Microsoft and Intel push UPnP because they want an IP based, computer friendly, device discovery and control protocol that doesn't require the fees Sun is demanding.
Intel is releasing this Linux UPnP SDK because they don't want UPnP to be seen (by the world) as MS only, and they don't want to see Microsoft control the spec, by being the only OS that supports it (when Millennium ships).
Does anybody have any ideas for obfuscating or introducing security to these new "connectivity" options?
Just put a firewall between your network and the outside, and deny your fridge access!
Of course this means that AMD will come out with a version of your fridge that runs faster and cheaper.
And your Cyrx Toaster will have a tendacy to burn toast since it cooks it with the CPU instead of heating coils.
To me this is interesting, but I agree with some of the questions of practicality. The best I can see is for notification of events.
Would it be nice to have your fridge email you when it beaks down, so you know to empty it before everythign spoils. Or have your stove let you know your food is burning, while you are on your computer?
I sig therefore I am...
I can't believe this post disappeared from this thread (but I can't find it since I reloaded - bug? Troller grudge? lost packet?), so I'm responding here:
The original post said:
Jeez, that's what I really want. A product called - "You Pee 'N' Pee"
Don't knock it. An earlier version of this technology ("B.E.E.R.") has enjoyed thousands of years of popularity among geeks/early adopters, and the general/neophyte market segments. It's often one of the first protocols a student learns in college.
It has already found widespread integration with home entertainment centers (TV viewing, VCR movies, and music) and was a vital enhancement to live sports presentations even before the advent of widespread televised sports.
BEER is already widespread in the contemporary American kitchen, though underutilized in that setting. (It can take much of the tedium out of elaborate preparation, for examples) Still, would Thankgsgiving be Thanksgiving without Beer? Not in most households!
Beer is widely used for communications, so much so that some people rarely communicate without using it. It may be the most widely used peering method in modern life. I-buy-U (a beer) is one popular application and Eff-U/Eff-Me often relies on it as well.
While efforts to integrate it into automotive settings have been -er- 'disappointing', the active developer base who try every day to make this work considerable exceeds the total Linux user base. Changes to the automotive infrastructure have long been planned to make this a more secure activity.
There are, however, concerns. Integrating child protection features on the USE, PROVIDER, and SERVICE models has raised serious concerns. Tighter access controls have traditionally failed, and if you think NetNanny and CyberParent are worrisome, check out a nanny or parent on Beer.
If you can go to bed, knowing you did a valuable thing today, you're very lucky. If you can't... it's not bedtime
Intel's been helping with Mozilla for a long while now. For instance, look at all the bugs that have been assigned on Bugzilla to people with @intel.com addresses. It wouldn't surprise me in the least if Intel also had people working on other open source projects.
They have to be working with the source with this type of box, or as some people have pointed out, the living room box just wouldn't be stable enough. Perhaps we should stop thinking of Intel as a hardware company...their marketing mantra of "building blocks of the internet" includes software too.
-Merlyn42
The audience doesn't care if it's hard.
So, I'd say Linux has been doing fine in my family room for 3 ro 4 months now.
DrLunch.com The site that tells you what's for lunch!
Here's intels site for upnp:
http://www.intel.com/ial/upnp/
Unix is simple. It just takes a genius to understand its simplicity. -Dennis Ritchie
""UPnP allows (devices) to participate in the same economy ... with the end user not having to understand how to set up an IP address,"
"Microsoft is building UPnP into its future Windows releases, starting with its Windows Millennium Edition."
With the future of device connectivity being, "Now you can receive emails from your fridge! And so can the companies that stock it!" -- doesn't this pose a major security risk for 2 reasons.
1) Microsoft is implementing it.
2) Interaction between all devices is handled by a standard protocol and each device is seen as a member of the same IP address?
Does anybody have any ideas for obfuscating or introducing security to these new "connectivity" options?
Ace
There has been userland PnP for some time, and it's in the 2.3 series kernel. PnP was even enough to get devfs into the kernel.
So uPnP seems to be some sort of extension of regular PnP. Will this new stuff fit with what we already have, or will it require substantial ripup and roadmap pain?
The living have better things to do than to continue hating the dead.
"...we'll still have a problem with the drivers."
Maybe, but the SDK allows manufacturers a good tool for creating drivers for their upnp devices. I suspect that with the introduction of tools like this, manufacturers will be better at supporting linux since they don't have to do so much from scratch.
SLP (Service Locator Protocol) seems to already be a good way to detect and locate services automatically on a network. It is a defined standard in RFC2608 and is used as the basis many networking technologies already (NetWare 5, MS W2K) and can be easily and consistantly configured to support any service type imaginable (it uses URI/URLs as its encoding system). There is already two SLP libraries for Linux, one from Sun under the SCSL and one from Caldera under the GPL.
Actually, in my limited research, this looks like a really neat way to generally advertise services. You would be able to create something like the Network Neighborhood or an NDS tree (Using SLP and LDAP together) for UNIX. Being able to plug into a network and instantly find out exactally what hosts have what services available without any configuration on your part (info for master DA server can be sent via DHCP) without all the messiness that the SMB browser protocol has (SMB, against all odds, is actually considered a protocol! Ha!) is a Really Good Feature(tm).
-- Remember: Wherever you go, there you are!
Jeez, that's what I really want. A product called -
"You Pee 'N' Pee"
Almost as good as drinking a lot of water in a short period of time.
--
"Outlook not so good." That magic 8-ball knows everything! I'll ask about Exchange Server next.
Just what you've all been waiting for, the UPnP Toaster. Now you too can *actually* asign your toaster an IP address and have it make toast for you.
No longer is this just a joke, it's a way of life!
damn.. sometimes we just have some weird technology.
nerdfarm.org
Dacels Jewelers can't be trusted.
It may not be perfect at first, or ever, but it's still better than nothing.
M$ PnP was awfull at first, but it's gotten a lot better if you stick with name brand components. I find a lot of PnP issues envolve the end-user not reading the installation directions first--as it to preload new drivers, etc..
Now all we need is an opensource Linux version of 'OilChange' with no subscriptions fees. (For those of you who don't know, OilChange monitors your system's drivers and checks the manufacture's websites for periodic updates and will automatically download/install new drivers).
Iota
God is Real Unless Declared Integer