Although vendor-written, this contributed piece does not promote a product or service and has been edited and approved by Network World editors.
Named Data Networking (NDN) may be in the early stages of development, but the strides being made are compelling. Gartner predicts NDN won't hit the “Plateau of Productivity” on their Hype Cycle for more than ten years, but when it does, a monumental shift in data organization and retrieval will come with it. We know a massive influx of data is on its way – NDN might be the key to managing the data, and Software Defined Networking (SDN) will be integral to helping NDN do just that.
The current Internet framework is built on the idea that communications is a conversation between a client and a server, an exchange of information between two points. NDN shifts the focus of networking from locations (origins and destinations) to the information itself. Networking will no longer be about connecting to a server to find information as much as requesting the information itself.
At the moment, the TCP/IP system identifies locations by numbers (your IP address). NDN seeks to imprint the data so the information can be retrieved from wherever it's held, instead of retrieving it from one specific location. This would mean the decentralization of data – instead of retrieving information from one point, you'll be able to get it from any computer that has it.
NDN was borne from an idea called Information-Centric Networking (ICN), which puts a great deal more importance on the information itself, instead of where it comes from. The goal of ICN and NDN is to build an architecture that's more efficient, flexible and secure, and that's something we'll all want as the public Internet becomes more congested.
As far as the design goes, it's a lot like P2P networking – you want information and you don't care where it comes from, so anyone with the information can give it to you. With NDN, a user sends a request, then the router stores the request and searches for devices with the requested information. It's a way to keep any one server from being overloaded, by spreading out potential streaming sites.
Efficiency, flexibility, security
When the current global network was designed, it was focused on moving small amounts of data on linear paths, and that design just won't cut it anymore. Legacy architecture is inefficient, inflexible, and insecure. Nobody could have predicted the current and near-future vastness of an Internet packed with streaming video, an Internet with over 5B devices connected.
Creating an Internet that's focused on information itself could lead to more flexibility, allowing us to optimize the flow of information and decongest the conduits funneling information around the world. This could help cut down on ISPs throttling bandwidth, as well as level the playing field for content producers that can't afford content delivery networks by simply eliminating the need for CDNs.
One of the most compelling reasons for NDN has to do with security. The current Internet's approach to security is centered around encrypting the connection instead of the information. NDN has the ability to embed security measures into the data itself. This means you don't have to worry about where the information is coming from, as long as the information itself has been labeled secure. NDN can tell you exactly where the data on the page came from, providing the user with both comfort and security.
How can SDN help NDN?
SDN is all about multipath routing optimization and flexibility. A challenge that will arise (and is already arising) for NDN is to figure out how to deal with a network stuffed to the brim with names. Each data set is going to have a panoply of names and the sheer number of data each user will interact with will be far greater than with TCP/IP. Another issue is that the name space is unbounded – there's no limit on how large it could get, potentially resulting in routing tables of unprecedented scale.
Fortunately, SDN can track data automatically and optimize subsequent searches. It can learn from previous routing attempts and sift through the extensive named data faster on each run. SDN controllers themselves could allow for prioritizing and influencing the named data that's accessible, as well. If a company wants to link certain names with the specific data they wish to correlate, and to maintain that information locally, they can do so.
Say a company relies heavily on weather data, they might want to prioritize local weather data over remote weather data on request. SDN can put together a tracking method for retrieving the information quickly, without starting from scratch each time.
As far as security is concerned, NDN will need to find effective and automatic ways to manage the data. SDN can assist with security by automatically tracking where information came from and whether or not that source is trustworthy. This is an adaptable system for maintaining security network wide, and will allow users to encode specific tails to the names that, in a sense, act as passwords. If you're worried about a certain piece of information being corrupted, have the sender encode the tail and set your SDN to check it as it comes to you.
Routing optimization is right in SDN's wheelhouse. NDN means data can have multiple locations and SDN can decide what location should be used to complete the data request for the user, as well as maintain network stability and information flow. Something needs to be responsible for figuring out the optimum route from the client to the desired data, and SDN would serve as a more advanced version of today's routing protocols. Furthermore, if the router already has the data in a network cache, SDN routers can figure out if the information is on its way. By aggregating all the requests for the same data, the router can optimize the routing path and spread the flow of traffic across multiple paths.
Even though widespread NDN is a long way off, it's worth taking steps now to optimize the current Internet architecture and to prepare for the transition to NDN. Employing SDN routing methods now not only improves the efficiency, flexibility, and security of the current Internet, but puts us all in a good position to ease the shift to a new network architecture.
Console is an enterprise software and interconnection platform that enables both network-to-network and enterprise-to-enterprise collaboration globally. Console also supports the CloudRouter Project, a collaborative open source project focused on developing a powerful, easy to use router designed for the cloud. This software-based router distribution is designed to run on physical, virtual and cloud environments, supporting software-defined networking infrastructure.
This story, "Named Data Networking is on the horizon, and SDN will help make it a reality" was originally published by Network World.