The Internet we have today was not what the early developers had in mind for its use. Today's Internet has weaknesses, suffers security attacks, and sometimes lacks performance reliability. The National Science Foundation (NSF) has funded five projects that look forward to a new Internet architecture - one that can meet the ever-expanding needs of businesses, education, government, and consumers.
The concept of the new architecture is not designed to be bolted onto the existing Internet. It is a technical and philosophical replacement. It is called Named Data Networking (NDN).
Named Data Networking Project
It attempts to respond to the fact that the patterns of communications are changing, and these changes need to be supported.
The NDN concept is to name data rather than naming its location. The current Internet secures the data repository. NDN instead secures the contents. This new design disconnects trust in data from trust in hosts, an approach which can support very different communication mechanisms. The project studies the technical challenges that must be addressed to validate NDN.
The NDN Project Overview lists some challenges, such as, "routing scalability, fast forwarding, trust models, network security, content protection and privacy, and fundamental communication theory." It continues: "The project uses end-to-end test bed deployments, simulation, and theoretical analysis to evaluate the proposed architecture, and is developing specifications and prototype implementations of NDN protocols and applications."
Back in 2010, four projects were funded (and later a fifth project was funded) for the investigation of the Internet of the future. The NDN project was funded under the NSF's Future Internet Architecture (FIA) Program. The FIA program is not the end, but an initial step toward delivering an improved future Internet.
"While the ultimate goal is the design and deployment of a network that serves all the needs of society, we realize that these projects are just the beginning of what it would take to create a full scale Future Internet," according to Darleen Fisher, program director for the FIA projects. "We expect that the knowledge obtained from this research will inform the development of future networks."
The principal investigators for each of the five projects are from higher education institutions: UCLA, Rutgers, University of Pennsylvania, Carnegie Mellon, and University of Massachusetts. Other higher educational institutions are members of each team.
Five Separate Projects
The five NDN projects cover:
• Named Data Networking, which moves the concept from being network centric--where the data is located--to what is the content.
• MobilityFirst focuses on mobility as the norm rather than the exception.
• eXpressive Internet Architecture (XIA) addresses the API for communications and network mechanisms.
• NEBULA is an architecture defining cloud data centers as the repositories of content and computational facilities.
• ChoiceNet project aims to develop a new architectural design to enable continuous innovation in the core of the network, using economic principles
Architectural Principles
The "Named Data Networking: Motivation & Details" document presents the principles that are used to guide the overall design. The diagram below from the document uses the image of an hourglass to show the first-generation Internet's focus on the universal IP layer.
The NDN vision replaces IP packets at the critical point in the infrastructure with something called "content chunks." The overview document linked above describes names, content chunks, and their role as follows:
"The names in NDN datagrams are hierarchically structured but otherwise arbitrary data identifiers. They can be used to name a chunk of data in a conversation, as the TCP/IP transport connection identifier plus sequence number does today, but they can also name a chunk of data from a YouTube video directly, rather than forcing it to be embedded in a conversation between the consuming host and youtube.com."
The basic transportation is pushed lower in the hourglass within the NDN architecture.
NDN retains the end-to-end capability that already exists with IP and expands on it. NDN does not have a separate transport layer. Transport functions are moved up to the applications. Security, rather than being attached to various protocols and applications, is integrated into the architecture.
NDN provides a data-centric security building block just above content chunks that functions by signing ALL named data. NDN defines the entire upper part of the hour glass as part of individual applications with file stream and browser functions.
What Does This Mean?
It will be a few years (at a minimum) before enterprises will have to contend with NDN. In the meantime, there will be NSF papers, new NDN websites, and vendor discussions about NDN. NDN could eventually affect all endpoints, network components, data centers, cloud services, application software, browsers, and software developers.
What is not clear is how the present Internet and all the connected devices and software would transition to NDN. There will have to some translation devices and/or cloud translation services.
There will be considerable push back against NDN because of the economic costs that will have to be incurred by everyone. When things work well, we do not want to change.
Consider that most of the PCs installed today do not use Windows 8 or 8.1. I heard a news program mention that 13% of the PCs and laptops on the Internet today still use Windows XP.
That being the case, it is worth following the progress of NDN, but it is not yet time to budget for the change to NDN.
