NETWORKING

  • 10/10/2014
    8:00 AM
  • Rating: 
    0 votes
    +
    Vote up!
    -
    Vote down!

Named Data Networking Would Eliminate IP Addresses

Cisco, Huawei, and other industry leaders have teamed up with universities to develop a completely new Internet architecture with no dependency on IP.

The newly formed Named Data Networking Consortium, where companies and research institutions can collaborate, is aimed at developing a new system of connectivity that promises a future without servers or IP addresses.

Named Data Networking (NDN) is a potential Internet architecture designed as a distribution network. To access the NDN network, a user needs to install some software applications and the protocol stack. This serves as a software router, working together with a core component of the network, the NDN Forwarding Daemon, to allow communication without IP addresses or hardware servers.

The consortium has attracted industry leaders such as Alcatel-Lucent, Cisco Systems, Huawei Technologies, Panasonic Corporation, and Verisign. Its members also include eight American and six international universities, including UCLA, devoted to creating the framework for NDN as the future standard of networking.

The current status of a basic NDN network testbed can be seen at the Washington University in St. Louis website.

The networking world is still looking for wider adoption of IPv6, but the new NDN protocol could change the way data is being transmitted, especially for corporate networks. NDN can be very attractive for organizations looking to share data over public networks and ensure the delivery and security of content.

Because data transmitted using NDN doesn't use IP addresses, traditional packets cannot be identified in transit; only the sender and recipient know which data needs to be reconstructed. This provides an additional layer of security over other methods, such as encryption.

NDN also is the right architecture to implement solutions such as Network Coding, a new approach to transmitting information over the Internet. This approach, using mathematical equations instead of sending packets, can be 5-10 times faster than other methods.

NDN is a clean-slate design in that it is a completely new architecture and has no dependency on IP. As the consortium website states, "NDN is an entirely new architecture, but one whose operations can be grounded in current practice. Its design reflects our understanding of the strengths and limitations of the current Internet architecture."

NDN aims to remove the restriction that packets can only name communication endpoints. "As far as the network is concerned, the name in an NDN packet can be anything: an endpoint, a chunk of movie or book, a command to turn on some lights, etc.," the consortium website says.

Figure 1:
Internet and NDN hourglass architectures.(Image: NDN Consortium)
Internet and NDN hourglass architectures.
(Image: NDN Consortium)

It likely will take several years to build the basic infrastructure for NDN to be a viable commercial option for enterprises and consumers. However, as often happens with revolutionary technologies, if the benefits are clear and the NDN consortium can attract other organizations, we could start seeing NDN applications in the marketplace relatively soon.

The basic Internet infrastructure we use today was defined more than 30 years ago. In the next decade, we could see a handful of new architectures taking shape that can creatively disrupt the core of the communications system. NDN, together with IPv6, 5G technologies, and the Internet of Things, will change the way we communicate and connect to our world.


Comments

No more IP

Pablo, thanks for this update. The flawed nature of IP is something I've heard rumblings about for years, but this is the first concrete evidence of a possible replacement. It's definitely exciting but difficult to imagine the actual migration process. When they say it is "grounded in current practice," does that mean it will run on existing hardware?

Re: No more IP

Susan, the current NDN protocol is designed to be compatible with the existing Internet as a "universal overlay": NDN can run over anything, including IP, and anything can run over NDN, including IP.

Obviously, like IPv6, it needs to be compatible with existing hardware. When NDN reaches a critical adoption size new hardware specisfically designed for it will be available, as the presence of hardware vendors such as Cisco and Huawei suggest.

Re: No more IP

This is a great concept!  Not only would it help with a lot of the insecurities that come with normal traffic thanks to what seems to be a better encrypted transportation layer, but it could also help with streamlining security tools down the road and creating better defined policies depending on the types of information that can be pulled from the traffic.

Re: No more IP

IPv6 never really caught on like it was projected to. Now will the NDN move people from v4?  It will be a few years until we can even venture at a guess i would say but it could be.

 

Re: No more IP

I am not much familiar witth NDN but yes this sounds interesting, curious to understand how address management will work here, will it be easy or complex.

Re: No more IP

Paul, it will take a few years to have commercial applications of NDN in the market. But I'm hopeful that we'll see NDN and similar concepts being developed across universities all over thw world, making a strong case for adoption.

Re: No more IP

"it will take a few years to have commercial applications of NDN in the market. But I'm hopeful that we'll see NDN and similar concepts being developed across universities all over thw world, making a strong case for adoption."

Pablo, it means still in research phase. Hope this will be good for the networking community.

Re: No more IP

it means still in research phase. Hope this will be good for the networking community.

Mynet, yes! there is much more reasearch to be done, but the support of industry heavyweights such as Cisco and Alcatel-Lucent means they are taking this new approach seriously.

Re: No more IP

"yes! there is much more reasearch to be done, but the support of industry heavyweights such as Cisco and Alcatel-Lucent means they are taking this new approach seriously."

Pablo, thanks for the update. Approximately how long it will take to get commercialized.

Re: No more IP

Pablo, Paul makes a good point about resistance to IPv6. Do you see NDN coexisting with IPv6, or eventually replacing it altogether?

Re: No more IP

Susan, I believe we need IPv6 now! NDN needs several years to be a suitable replacement and, as I said in a previous comment, can work over IP, and IP can work over NDN.

The current infrastructure of the Internet is IP based and trying to replace it is impossible right now. IPv6 needs to be accepted as the standard for many years. Other solutions such as NDN could start replacing IP, or complementing it, when the market conditions and the technology are more mature.

Re: No more IP

Hi everyone,

Hi Pablo -- Many thanks for the topics and this clear explanation. Also, these kind of topics help us to understand again that whatever the massive adoption of a specific technology it remains irreplaceable and open our eyes on news perpectives! Really, that's a great info for me !!! Thanks again.

I remember the OSI(Open Systems Interconnection) protocol, the famous CNLP (Connectionless Network Protocol) despite all its advantages and fundations, now it's IP and we talk IP. Also IS-IS (Intermediate System to Intermediate System) with it all advantages (i want to mention one Node based routing protocol) we convaincing him to join IP and we give him a new name Integrated IS-IS. I remark that we still make effort the maintain IP at the top by providing solutions to solve of course its problem (i'm referring to technologies such MPLS, RSVP, IPv6,...) and the interesting part is that some of vendors in NDN technology remain doctors of IP. At this time where all communications seem to be based on IP and telecoms operators try and are working to provide IP connections End to End by eliminating old powerful techno (ATM, X.25..) , i think that if it must be adopted, it will just complete IP and very very difficult to believe into its chance to replace IP.

"The current infrastructure of the Internet is IP based and trying to replace it is impossible right now. IPv6 needs to be accepted as the standard for many years."  I totally agree with you Pablo, and this technology must be very strong and must demonstrative to achieve its purpose even the big vendors are its founders!  after all one never know, that's technology and upset is well known. (but i'm not too much optimistic).

I wonder to know what Routing protocol we plan to use for NDN ?

Yes, very impressive to see Cisco and ALU be part of this project!!! But, please, Pablo, i'd like to know  if others big vendors such as Ericson, Juniper are part of the team ?

I ask myself if things still continue in this way, isn't a kind of instability ?

Great topic & Many Thanks !

Re: No more IP

It does look promising. Alot will depend on how cleanly it works.

Re: No more IP

NDN would make the internet much more resilient and enable data sharing even if the infrastructure is down. I can imagine it will be useful to gain real time sensory data to build a map if, for instance, a datacenter was down due to a natural disaster. Good for the small business as well that do not have the resources to deploy or rent a datacenter for running their app.

security

The security aspects of this project are really compelling. Anything that can bolster what we have now for Internet privacy and security would be very welcome. Pablo, are other security companies involved other than Verisign?

Re: security

Free flowing information and content produces a lot of economic value, it is an excellent thing. I wonder if it would be theoretically possible to block information/content that is not linked to an IP address or server. For instance, a social media app built upon NDN and a hypothetical country wanting to block the app.

speed

I wonder if there are any concerns with the speed of data flowing across NDN. Would "IPless" be slower?

 

Domain Name Routing

Interesting. I happened to work for a professor of Santa Clara University in 1997, Dr. Hasan Alkhatib. He was working on similar concept. The technology he developed was called Domain Name Routing, which was basically the same concept as NDN. The technology allowed network elements to be addressed using human-friendly domain name instead of IP address. He was trying to tackle IPv4 address depletion problem. DNR technology was compatible with IPv4 network and allowed unlimited addresses based on domain name structure. He was trying to promote DNR at the time for a few years but was unsuccessful. Now, maybe, there are other incentives for this to happen.

Re: Domain Name Routing

Peter, thank you for sharing that story. I belive Dr. Hasan Alkhatib was not succesful at that time because the industry was focusing on solving the IPv4 problem with a similar technology.

NDN is now getting attention because we need more flexible protocols, especially for the IoT.

Re: Domain Name Routing

You are right. The industry was focusing on IPv6 at the time, which was endorsed by several industry heavyweights.

Interesting but...

I love new technology and the excitement that comes along with various university projects NDN in the latest in a long line of these 'next-gen' ideas. Remember OpenFlow and how it was going to fundamentally change everything, we are over 7 years into it and still owkring out the kinks. I am not trying to be negative but realistically things like this will take a very long time to happen and will not happen unless there is an economic driver - you reference IoT as being this type of driver but there are many ways to solve that issue that would leverage the current infrastrcuture and its addressing schema a good example is LISP. In my opinion great new ideas are always welcome but enthusiamsm must be tempered with a bit of reality...

Re: Interesting but...

Thanks for weighing in on this MarkB111, and for the dose of reality. While I don't think anyone here is expecting anything soon from this initiative, it's always good to have perspective, and your example of OpenFlow is a good one. 

Re: Interesting but...

"While I don't think anyone here is expecting anything soon from this initiative, it's always good to have perspective, and your example of OpenFlow is a good one. "

Marcia, eventhough we know that nothing Is going to be happen in near future; we wish to happen it at earlier as possible. I think it's an industrial requirement from system admin point of view.

Re: Interesting but...

@MarkB111> I'm glad somebody mentioned LISP, because as I read the initial article I kept thinking that this sounded terribly like LISP, but with domain names in top of it rather than yet more IP addresses for the end points.

Re: Named Data Networking

Like some of the other commenters here, I think this technology sounds pretty cool, but I'm not so sure how likely it is to catch on. The benefits sound substantial (and as someone else said, your point about IoT is spot-on, Pablo). Likewise, I won't deny that could plug up some glaring holes with traditional network protocols in the security area, but I'm a little skeptical of even that - I'm of the mind that if you can engineer something, you can reverse engineer it. Nothing will be secure forever.

Minus that, I'm just not sure the payoff here will be alluring enough to cause a large enough scale shift. We sort of end up with the case that to become popular, NDN would need to already be of comparable quality to what's out there already, but for it to reach that level of quality, it would have already be popular and have a lot of money put into it - a classic catch-22. Tech companies tend to be pretty risk-averse; if they're not going to see results tomorrow, it's hard to get them on board with something if what they have already is working fine.

Re: Named Data Networking

I find NDN fascinating and probably how the Net should have been constructed to begin with - which makes me wonder if it is really worth the effort  ?  Especially now since we have IPv6 addresses just waiting for someone to use them.

Hilarious

The hilarious thing about this to me is that, the way things are going, IPv6 will be obsolete LONG before it sees significant adoption.

Re: Hilarious

@Joe Stanganelli> Wait ... are you saying that IPv6 is going to see significant adoption? ;-)  I'll take a double of whatever you're having, my friend!  *grin*

Re: Hilarious

@jgherbert: By "significant," I meant statistically so.  ;)

Questions, questions, questions...

So I've read this post. I've also gone off to named-data.net and tried to read through their project summary and FAQ to get a better understanding of what's going on. At every turn I found myself saying "wait, I need some more detail on that". At one moment it looked like an overlay, then it looked like a full IP replacement, then somewhere in between; it doesn't have a dependency on IP, but it can use the same routing protocols to distribute name data instead - yet with the exception of ISIS, the routing protocols typically require IP to run. Waahh. My head hurts. More data needed. I can see I'm going to have to start digging into this one more deeply!

Re: Questions, questions, questions...

@jgherbert: Congratulations.  With all the research and attempts at real understanding you've done, you're now more qualified to be a tech journalist than many tech journalists!  :)

Of course, the above dig is not levied at Pablo -- whose work is always excellent, well researched, and thought-provoking.  :)

Routing on NDN

Hello, when I am trying to make the difference between IP and NDN, How can NDNoverIP work if NDN is based on named data however IP is based on IP address? and can they coexist together?