Slashdot Mirror
DARPA Aims to Redo the Internet Protocol
Xaleth Nuada writes "The Defense Advanced Research Projects Agency's (DARPA) is looking to redo the entire Internet Protocol. With the DoD increasingly adopting network-centric warfare the shortcomings in the current IP have become resoundingly clear. Everything works fine for static hardwired networks. But not for dynamic wireless ones. The benefits for your average geek? How about REAL wireless networking? Easier network set-up? Increased wireless security protocol? Increased reliability in sending information?" Don't forget massive incompatibility and upgrade hassles. :)
Given the scale of the re-work proposals (replacing the Von-Nuemann architecture...), I'd be surprised if there wasn't some effort made to embed snooping and tracing into all packets transmitted. This *is* the DoD after all!
On the other hand, given how slowly IPv6 is making its way into the wider world, we probably don't have too much to worry about for the time being!
Simon
Physicists get Hadrons!
"Don't forget massive incompatibility and upgrade hassles."
I read that as:
"Don't forget about the sudden explosion of extended-temp jobs flooding the market as the Internet decides to change over..."
Job? I don't have time to get a job! Who will sit around and bitch about being broke and unemployed then?
Upgraded to IPv6. Sigh.
They'd best be careful, or this "Protocol 7" will inadvertently cause data from dead people to leak to the Internet...
Don't forget massive incompatibility and upgrade hassles. :)
Yeah man, but massive incompatability and upgrade hassles are what keep some of us employed! GO DARPA!
Excuse my speling.
Making The Bar Project
And when will this new Internet Protocol be rolled out...
shortly after IPv6 adoption?
I don't see Satan reaching for his winter parka just yet...
Easier activity tracing, easier monitoring, easing censorship of "bad" websites, easier disabling of internet access to undesirables.
DARPA did help lay the foundations for the Internet. They may be in a good position to bring positive innovation to the IP protocol. Just as long as enough of us /.ers can see through any hidden embedded packet sniffing credit card stealing email reading we're watching you protocols, we should be GREAT.
Im a former Marine myself, and I fondly remember what a nightmare it was just trying to get everyone to have the same crypto loads for existing voice communications hardware. Im really curious as to how they propose to keep the network secure. On the other hand, the possible benifits are huge. Distributed sensor networks in particular could be revolutionized by this.
"Hand me the bullet-shooty-thing and a box of little hurts" -Overheard on a USMC Rifle range
Let's just all pray the military dosn't call this SKYNET.
Yeah, heaven forbid we learn from our previous attempt and start fresh. We should aspire to do like Microsoft - maintain backward compatability above all other goals. Seems to work for them, right? It certainly makes things more secure...
.sigs are for post^Hers.
DARPA takes aim at IT sacred cows
By Joab Jackson
GCN Staff
ANAHEIM, Calif.--Now that the Defense Department is embracing network-driven warfare, it is taking a hard look at radically improving, or discarding altogether, some fundamental computer and network architectures.
Flaws in the basic building blocks of networking and computer science are hampering reliability, limiting flexibility and creating security vulnerabilities, program managers said this week at the Defense Advanced Research Projects Agency's DARPATech conference.
Among the IT holy grails that DARPA wants to see revamped are the Internet Protocol, the seven-layer Open Systems Interconnection model--which defines how devices communicate on today's networks--and the von Neumann architecture, the basic design style underpinning almost all computers built today.
Many military commanders have been slow to adapt IT for critical tasks because they sense the equipment is unreliable, said Col. Tim Gibson. He is a program manager for DARPA's Advanced Technology Office, which is leading efforts to radically redefine computer architecture.
"You go to Wal-Mart and buy a telephone for less than $10 and you expect it to work," Gibson said. Yet people usually do not expect the same of their computers. "We don't expect computers to work, we expect them to have a problem."
"If a commander expects a system to have a problem, then how could they rely upon it?" Gibson said.
Gibson cast some of the blame on the packet-based nature of Internet Protocol, which was not designed for foolproof delivery of messages. The protocol cannot guarantee delivery of e-mail, for instance.
"The packet network paradigm probably needs to change," Gibson said. "I'm not advocating throwing out the Internet Protocol completely, but we must absolutely have some mechanism for assigning network capabilities to different users and that capability has to scale to large numbers of devices automatically. The commander wants to be able to send a message and have it delivered, completely, accurately and on time."
Another limitation with the IP approach is the inability to dynamically build networks. The military wants to quickly set up ad hoc networks.
"Static networks are no good for tomorrow's battlefield, because everything will move around all the time," Gibson said. "What we need is dynamic scalability. Today's networks are stationary and have a static infrastructure that provides service to static end-nodes. Moving the node outside its standard service area requires reconfiguring something. Moving infrastructure always means reconfiguring something."
As a result, DARPA wants to fund development of new protocols or enhancements to the existing IP that will allow nodes, such as computers, to automatically sign on to networks in their vicinity.
Another aspects of the networking that DARPA wants to revise is the seven-layer OSI stack, long held as the basic foundation for building network protocols.
The OSI model was not designed for wireless communications devices, said Reggie Brothers, a DARPA program manager.
"The OSI model served us pretty well for the stable, predictable world of wireline communications," Brothers said. "Mobile networks are nothing like that. They are unpredictable and highly variable. We need to think of different layers of the stack to relate to one another directly, like a mesh, instead of one level up to the next."
The increased complexity of the network stack would let nodes enter a network quickly and without human intervention, Brothers said.
The von Neumann architecture will also come under scrutiny from DARPA.
"It is time to ask the harder questions about the ways of computer architecture we've been using for the past 30 years. Is it time to scrap the von Neumann architecture?" asked Anup Gosh, program officer for the Advanced Technology Office.
This architecture, which defines the basic essential parts of
Please, anything that's not encumbered by *anybody's* IP patents.
just like it has for IPv6.
People will only upgrade if it's absolutely painless or absolutely necessary, we should've learned this by now. I have friends that still use analog cell phones, just because it's easier not to switch.
-- atomly
SMTP is not a transport-layer protocol. TCP and UDP are the most common transport-layer protocols that ride over IP - although many others exist.
There are certainly some valid arguments for looking at other transport protocols (the lack of mobility features in TCP/UDP, for instance), but SMTP is not one of them since it's an application-layer protocol.
DARPA invented the Internet Protocol before, and within a few decades the technology was widely deployed. Unfortunately, this time around, things won't be so easy.
Before, it was competing against a vacuum. Now, it's competing against ubiquitous IP. They may develop some cool stuff that works on a battlefield, but it will never get widespread usage, commoditization, and economy of scale that IP has. If they come up with new features that work great, somebody will find a way to get similar functionality built on top of good old IP.
IP isn't perfect, but it's good enough that there's no way to displace it, given its free nature and level of entrenchment=.
Gibson cast some of the blame on the packet-based nature of Internet Protocol, which was not designed for foolproof delivery of messages. The protocol cannot guarantee delivery of e-mail, for instance.
Who is this guy really? Thats not what IP is for - foolproof delivery of packets is handled by connection-orientated protocols like TCP. Sure, it might not get the *right away*, but the flexibility of packet based routing is something that has made networks as reliable as they are today (despite the huge amount of moaning that people do about them).
Mind you, as people have pointed out before, IPv6 has been waiting in the wings for a while now, and a military request for change might be the kind action needed to kick other people into gear.
Voice is data, video is data, they all run over IP and therefore can be considered data just like anything else.
What we don't have is security built into IP. IPSec is a good beginning, but its more of an afterthought. Not nearly as good as what they could do if security were an integrated part of the native IP protocol.
we must absolutely have some mechanism for assigning network capabilities to different users
Sorry, but the network capability of running a web server hasn't been assigned to you. You are blocked at the protocol layer.
Sounds like they don't want the Internet to be a network of ends anymore and control who can do what with the network. Nice experiment, this unrestricted free speech on the Internet, but we've decided we don't want you to have that. Be consumers, not producers.
Oh, say does that Star-Spangled Banner entwine / The myrtle of Venus with Bacchus's vine?
I'm not sure why the von Neumann architecture is such a security problem. I mean, the problem with computers not working isn't how they're built per se--turing machine, post machine, hell use cellular automata--it's that the mathematical theory says "it is impossible to write code (in general) that is guaranteed to be bug free". You could change the von Neumann archiecture, sure, but you could just as easily 'write an interpreter' (though with hardware) for the architecture. Either way, if you're writing code, you're going to have bugs.
They blame the packet nature of the network for lots of the problems but I see not other perposal given. How on earth do you build a network as large as the internet based on a non-packet archetecture? I am studing computer science right now at school and haveing completed two telcom courses and nobody has ever discused a conection-oriented technology that or even a conection-oriented concept that could cope with a network as large as the internet with as many hosts. Do any of you in slashdot land have a clue how they might even start to go about doing this? The other posibility is its a new twist on a conectionless network but how on earht is that possible with out some sort of packet archetecture to send over it, otherwise you'd have no way to change path with conditions and changeing conditions are UNAVOIDABLE on any network I have ever seen.
Repeal the 17th Amendment TODAY! Also Please Read http://www.gnu.org/philosophy/right-to-read.html
Yes, but the serious question is whether or not this so-called IPv7 will incorporate the Schumann resonance, tap into the collective unconsciousness of mankind, spontaneously create a little girl complete with family, and allow its creator to become some sort of god-like revenant.
Maybe I'm just watching too much anime...
If it's for-profit but free, you're not the customer -- you're the product (e.g., the Slashdot Beta's "audience").
It's time to go back to basic networking class...
The OSI Networking Model is a 7-layer system that can be used interchangably, layers run on top of each other... for example, HTTP specifies that it use TCP which wraps around IP over any physical protocol. It doesn't care if you're using WiFi or a hardwired connection.
So, what this is saying is that IPv4, and even IPv6 are protocols that were written with wires and not wireless in mind. There are tweaks that can be made to the next version of the Internet Protocol and maybe even TCP and UDP to make them work better when on wireless without giving too much up when used on a wired physical link. This is the process of figuring out what changes should be made for next time.
TCP works poorly in a wireless environment because of the congestion control. When packets get lost, it assumes it's because of congestion and starts backing off, which slows down the connection even more. That's not always the case in wireless because packets can get lost due to interference and a number of other scenarios that do not exist for wired connections.
EvilCON - Made Famous by
It sounds to me more like some general had a brief introduction to computing theory, but didn't relate it to any real current technology.
The alternative to Von Neuman (Code and Data in the same memory) is to have code and data in seperate memory areas. This makes it very difficult to make computers where the code can change. Sure, there's no buffer overflows, but there's no security patches either. It might be fine for embedded devices, but I'll not have it on my desktop. The Page (or Segment) executable flag of more modern memory management units does the job fine, without all the hassle.
The OSI model is already not used anywhere except to compare proposed network models to; it's way too complex.
He talks about replacing packet switching so that messages are delivered on time & with certainty. Presumably he means some kind of virtual circuit switching, but he also talks a lot about constantly shifting ad-hoc networks. Circuit switchinfg & ad-hoc networks don't mix well. You have to know what the path is going to be before you can reserve it. It's probably better to just turn on the QoS and AH already implemented in IPv6.
They could mean ad hoc wireless networking. If they are looking for something that could help them communicate in the field, ad hoc wireless networking has great applications for them--basically, an ad hoc network does not have predefined hosts, access points, or what have you. Every node in the network communicates with the nodes around it (they could be a mixture of some wireless nodes and some wired nodes). There is no predefined leader, but the nodes themselves pick which nodes will act as temporary leaders to keep routing information, among other things. There are many different algorithms for determining these leaders, and the leaders can be changed if necessary due to nodes moving, entering an area, or leaving an area.
More information can be found here (Google's html version here.)
I am clearly fatter than you.
The article seems to have two different main points. Firstly that the entire networking model (7 layers) is inappropriate for "reliable" networks. Secondly they suggest that the entire model for building computers is wrong, and that somehow they need to use hardware to isloate programs.
The issues they address in the first point were issues which I felt were meant to be addressed by IP6, has/will it fail? I always understood IP6 as being designed to (optionally) have secure connections, qos and an ip address structure to allow for floating nodes. Would IP6 not stand up to delivering messages in network time for the entire US military structure?
The second issue seems simple to me, yes it will be much more reliable if you use a seperate computer for each task and allow them to communicate, but can you tolerate the lack of flexibility and is it even possible to do anything meaningful without adding lots of parts and weight (the more parts, the less reliable). I can imagine building a chip which actually contains 8 386s and 32M or ram split into 4M per 386, then have the disk controller map the device in an 8 way split so they can't touch each others data, a network chip could act as a switch to all the information, providing qos etc. buses to expansion could be mapped to cpus, but is it worth it or are you better off building two different but functionally identical systems so if one fails the other shouldn't? Also it's still one machine, as soon as you actually split it out into a meaningful number of machines weight, size and handling all become a problem. It would be lovely if you could sew tiny bluetooth enabled cpus w/mem into all the army gear and then they cluster together into a super cpu which reads the soldiers thumbprinted data device to figure out what to do, but would that actually require any sort of fundamental shift in how computers are made to achieve?
To me this article simply states that they haven't managed to build a good enough network yet, and want some cash to do it, and that they haven't managed to build a reliable os/app combination to deal with their needs yet either! Just the talk of "One of the limitations inherent in this approach is that when an application malfunctions, it can affect other programs" made me think they need to look harder at their OS. I will be surprised if the end result isn't IP6 (perhaps a modified army version) but you never know! I wonder what OS they'll go with though?
Never underestimate the dark side of the Source
Oh, the moaning, oh, the bitching.
:)
Has it occurred to anyone else that DoD might not be out to reform the Internet in any way? They are out to build a network model to serve their own needs, but they have no need to reform the rest of the world.
Now, if they make this revolutionizing new network protocol/infrastructure public other people might want to adopt it because it's neat. But me being a hardened cynic, this will most likely only find use in privately owned networking ponds...Kinda like a certain version pf IP we all know of
--Mike---
Flaws in the basic building blocks of networking and computer science... "It is time to ask the harder questions about the ways of computer architecture we've been using for the past 30 years. Is it time to scrap the von Neumann architecture?"
This is the only interesting part of the article. I couldn't care less what they do with the OSI layers. As long as someone writes about it as well as Stevens wrote about TCP/IP, it'll take me a month of reading and programming to get under my belt. We all learned Pascal, then C++, then C++ again when the standard came out, then Java, and Lisp, and Smalltalk, and Perl, andd Python, and C#, and a half-dozen more languages as the need came up. Now, you have to learn a few new networking layers and protocols. No big deal -- you should be pretty damned familiar with learning different implementations of stuff you already understand.
But, replacing the von Neumann architecture means changing just about everything I know. That's big. Everything is von Neumann. All the computational models, all the theory, all the basic underpinnings of what I know... it's all pretty much out the window once von Neumann goes. It's not just a dozen evenings at home with a book and reference implementation to relearn all of that stuff, either. It's relearning nearly all the Computer Science I know, and probably learning a whole bunch of new Maths to go with it.
That's gonna hurt.
Slashdot is jumping the shark. I'm just driving the boat.
DARPA and the military aren't interested in rebuilding the internet, they are interested in rebuilding IP.
They want to rebuild IP because they have a need for a better system. They need secure, reliable, ad hoc networking so that battle groups can communicate with each other.
These are private WANs. Not the Internet! The Military is not going to send real time battlefield data across the public internet, and real time battlefield data is what this thing is all about. The military launches and rents satellites for that sort of thing, they don't send it across uunet.
When they create a WAN, they have to have some mechanism to talk. Right now it might be IP, but in the future they want it to be something else. Something better for THEM.
The US Military couldn't care less if the rest of the world, or the internet itself, started to use whatever they come up with.
As far as those attacking technical limitations, when they started working on the original internet I'm sure everyone was saying, "Fault tolerant distributed networking with dynamic routing? That's impossible, why are they bothering" The point of DARPA is to do science and advance the field beyond current knowledge.
They may succeed, and they may fail. But they shouldn't just not try.
Map the cells in the state tables to appear as conventional RAM to the host, and reprogramming becomes as easy as a memory write. Bad cell?, just route around it. The fact that it's all state driven allows you to build an automated rerouter almost trivially.
post Von Neuman computers are going to be wicked fast, if they can build IO to keep up with them.
--Mike--
Now that I have read the article, I finally concluded it's full of shit. I'll break it down bit by bit:
... the seven-layer Open Systems Interconnection model''
..."''
``Among the IT holy grails that DARPA wants to see revamped are
Well, they can't. It's just a model, an abstraction. It's not like networks are actually built by looking at the OSI model and carefully following it. It's more like you build your network infrastructure and protocols, and then the OSI model says that you can call your wires the physical layer, the software that does something with the network the application layer, etc.
``Many military commanders have been slow to adapt IT for critical tasks because they sense the equipment is unreliable''
Well, that's their judgment, but what does it have to do with the Internet protocol?
``"We don't expect computers to work, we expect them to have a problem."''
I guess many people do, but I don't. I buy my computer and expect it to work. If it doesn't, I'll return it and get a working one or my money back.
``Gibson cast some of the blame on the packet-based nature of Internet Protocol, which was not designed for foolproof delivery of messages. The protocol cannot guarantee delivery of e-mail, for instance.''
Right he is. Reliability is in TCP, and this is why most application protocols build on TCP. The unrealiability of IP is there on purpose, so we don't have the overhead of TCP when it's not needed, and that if we come up with a better alternative to TCP, we can use that instead without having to throw away IP. Conversely, we can exchange IPv4 for IPv6 and implement TCP on top of that. It's called modular design, and generally considered a Good Thing.
``"The packet network paradigm probably needs to change," Gibson said. "I'm not advocating throwing out the Internet Protocol completely, but we must absolutely have some mechanism for assigning network capabilities to different users and that capability has to scale to large numbers of devices automatically. The commander wants to be able to send a message and have it delivered, completely, accurately and on time."''
Ok, fine, so you need a real-time protocol. I can see how that wouldn't work with IP's best-effort (read: unreliable) delivery, without further guarantees. However, there is nothing in IP that says it _has_ to lose packets. If you find a way to guarantee timely delivery of packets (my bet is that you can't), then you can layer IP on top of that. Of course, you don't _have_ to use IP, but if you opt for a different protocol, that doesn't mean that I have to drop IP too.
``Another limitation with the IP approach is the inability to dynamically build networks. The military wants to quickly set up ad hoc networks.''
I don't think that's true. Just like there is nothing in IP that _prevents_ guaranteed delivery, there is nothing in it that prevents building dynamic networks, either.
``"... Moving the node outside its standard service area requires reconfiguring something.
Yes, necessarily. However, the implication seems to be that IP somehow cannot handle this. Again, there is nothing in IP to prevent this. You could simply broadcast a message to discover nearby access points, and attach to the one with the strongest signal. Periodically, or when the signal gets weak, you broadcast again.
``As a result, DARPA wants to fund development of new protocols or enhancements to the existing IP that will allow nodes, such as computers, to automatically sign on to networks in their vicinity.''
Like ZeroConf? That would be a Good Thing. More power to them.
``The von Neumann architecture will also come under scrutiny from DARPA.''
I won't comment on that. I don't know what exactly the Von Neumann architecture is, and besides it is off-topic in my discussion on network protocols.
Please correct me if I got my facts wrong.
Most people seem to miss the fact that the R in DARPA stands for Research. Research is not done by accepting the status quo. If ARPA had not invested in the original network research, who knows were we would be today!
TCP/IP is not perfect for every use. If DARPA can find a better set of protocols to slide into layers three and four of the OSI model, more power to them.
Internet protocol suite
For every problem there is a solution that is simple, obvious and wrong.