“Today we build, support, and use Internet applications and services on top of an extremely capable architecture not designed to support them.”

From the early days we started with communication, communication has mainly been focusing on connecting two end points. It started with telephony before IP communications was introduced modelled around client-server communication. Today we are less focused on a specific destination to connect, but more focused on information acquiring. As an ordinary network user we mainly focus on getting information, and thus less focused on which server delivering the information as long as we trust the delivered information.

Named Data Networking is an implementation and an architecture where the information given by name is the main focus as in contrast to IP with IP addresses presenting location. Hence, NDN might form the basis for the future Internet. NDN introduces a completely new way of accessing information in a network by addressing the name of the information instead of the IP address of the server that produces/stores the information. This allows for a better integration of the information infrastructure with the communication infrastructure and can potentially leverage a more efficient network used for information sharing. The main benefit of the architecture is data sharing. That is, a network where a larger part of the data is shared between more than one person such as with news and video.

NDN is a candidate for future autonomous systems. The main reason is that these systems consist of many nodes (UAV, UGV, USV, UUV) that depends on communication for collaboration to reach their common goal. This project will focus on challenges within autonomous systems to provide a network allowing collaboration and to acquire/share sensor information.

Autonomous nodes will collect sensor information and based on collected information calculate the next move. This process is a continued process and involves many problems in a large network in terms of consistency and more. Artificial Intelligence is one promising technology for future communication networks. Future communication networks will be large in terms of connected items, but also have a large variation of services. These elements, combined with mobility, increase the risk of a large configuration set, complex networks and instable networks. Artificial Intelligence used in a distributed autonomous network is therefore seen as a promising technology to mitigate configuration errors and to better utilize available network resources.

Despite great potential, the architecture is dependent on being explored and indicated challenges to be solved. In this project, assignments are given within some indicated challenging areas. The proposed topics are not covering all challenges, so students that see other challenges that they would like to follow are more than welcome to present the ideas. The overall goal of this project is to gain knowledge of NDN as a promising upcoming architecture for future Internet. By working on one of the given topics you will:

• Get good knowledge in present and future Internet challenges
• Deep Insight into how challenges can be mitigated or solved
• ..

This project is divided in a number topics. You can select one of them, or if you find the NDN interesting but does not find an interesting topic for what you are interested in, then feel free to contact and we will try to help.

Students searching for a master assignment are encouraged to take contact. Contact information can be find under Partners.

Below we have listed some topics, where master assignments can be selected, along with a short topic description.

Named Data networking uses names. Content or data is requested by names, but how we should arrange the use of names is not finalized. One suggestion is to adopt the naming systems used in DNS. DNS is using a hierarchically naming systems exemplified with its.uio.no. The last name “no”, is at the top level and where some DNS servers are responsible of keeping track of all names under no. Similar uio, has their own DNS and responsible for all names under UIO including its. Its.uio.no is then typically associated with one IP address. In IP networks, network destinations and services are typically given by name. IP is highly dependent on catalog systems to map service location and services to IP address. NDN, on the other hand is using the same name at the “application” layer and at the network layer. Hence, DNS and similar catalog systems is mitigated. Since all data is given by name, typically all sent content must be named with a unique name for the whole Internet. This provides many opportunities but also challenges.

Named Data Networking is not built on an end to end model exemplified with client server, but more on acquiring data independent of location. Data decoupled from location means that consecutive data requests might travel different paths, and hence traditional traffic control tools is not always suitable. Within this problem area, two assignments are provided. The first assignment is focusing on how to best get an overview of expected capacity on a service. The second assignment, focus on how to best control the traffic in terms of avoiding traffic congestion.

Future communication networks will likely hold a larger number of devices (IOT ), hold more services where larger services is split in subservices that further can be moved around to best utilize network resources and meet the expectation of the end users. This leads to a more flexible network, but also a more challenging network. In NDND a service can be divided in many subservices named differently, and data produced from the subservices can be stored on multiple locations.

Today majority of all network equipment’s are operating on IP technology. NDN is not based on IP, and thus need support for operating natively on the network hardware. Historically, new technology has been introduced by being run as an overlay over existing technology. The migration from telephony to IP is an example on this. In the early days, IP was run over telephony technology before being stepwise natively introduced to the network. Today, NDN can be run over IP, and IP can be run over NDN. However, the network equipment’s of today is not supporting NDN and is likely not going to support NDN in the near future. However, the trend of new network equipment’s is based on software. New technology P4 (Programmed…..) is a promising technology for future networks and especially for specialized networks such as emergency networks, oil platforms networks or networks dissimilar from the main stream networks.

All applications and services do have network requirements to operate without errors. To illustrate, some applications are more bandwidth sensitive while others are more delay sensitive. The differences in resource requirements need to be addressed by the network itself as all applications and their variation share the same resources. Hence, the goal of providing QoS at the network level is to utilize all network resources by arranging the resources to the applications. In case more resources are required than available, some resource request must be rejected by not admitting the request for resources. Traditional work within QoS in IP network is oriented either around traffic classes or individual flows. A traffic class can be exemplified by for instance video, where all applications sending video share the same resources within the network. QoS based on traffic classes is in IP networks often associated as Diffserv, while resource handling per individual flows is named intserv. NDN is not oriented around location but names and thus many of the traditional QoS methods are therefore not applicable in a NDN network. However, resource control is still of interest in NDN and especially in network with low resources or networks experiencing high resource variation such as emergency network. QoS in these networks play an important role.

NDN is at present time implemented and is continually tested in a testbed covering all continents.

Uninett is the national research IP network operator in Norway. Uninett provides universities, university colleges and research institutions with access to the global internet as well as access to a range of online services. Uninett also offers counseling and acts as secretary and coordinator in collaborative activities between the institutions interconnected by Uninett.

Uninett need to prepare for upcoming network technologies. Named Data Networks (NDN) is one potential future network architecture, howevere it is unclear how a network operator (like Uninett) may utilizes NDN in such a manner that revenue is generated.

The projects objective will be to survey potential “killer application” domains for NDN and investigate which business models may have relevance. Selected relevant business models should be analyzed and new novel models suggested.