Network Coding: 

Linear Network Coding (LNC) has received considerable attention in recent years for its potential for achieving the theoretical upper bound (max-flow) of network resource utilization via the introduction of coding concepts at the network layer. It has been shown that with simple distributed linear coding in place of the usual forwarding at intermediate nodes, system throughput can be increased in several canonical network topologies.

Link Failure Monitoring Using Network Coding:

Monitoring of link properties within the Internet has been stimulated by the demand for network management tasks such as fault and congestion detection. This would help network engineers and Internet Service Providers (ISP) to keep track of network utilization and performance. The need for accurate and fast network monitoring method has increased further in recent years due to the complexity of new services (such as video-conferencing, Internet telephony, and on-line games) that require high-level quality-of-service (QoS) guarantees.

In this project we investigate novel use of network coding for a different purpose - to locate congested link(s) inside a network through end-to-end measurements at (external) boundary nodes. Current link monitoring schemes suffer from identifiability problems; i.e. they are unable to infer link status in many canonical network topologies. We show that network coding offers the promise of being able to identify congested links in any 'logical' network. In addition, it is possible to locate any congested link inside the network during an arbitrary amount of time by increasing size of transmitted packets. That introduces a speed/complexi

• You may find a presentation for this work here
• Mohammad H. Firooz, Sumit Roy, Linda Bai and Christopher Lydick, 'Link failure monitoring via network coding,' in Proc. IEEE 35th Conf. Local Computer Networks, 2010, pp. 1068-1075.
• Mohammad H. Firooz, Sumit Roy, Linda Bai and Christopher Lydick, 'Link Status Monitoring Using Network Coding,' preprint.

Network Coding Implementation

In this project, we investigate network coding throughput using a network simulator. We use imulator. We use OPNET Because of its wide acceptance as    a network modeling tool within the both academic and commercial communities. This is the first implementation of network coding within actual network simulator to the best of our knowledge. For that, we change current router model in OPNET to handle both ordinary and network coding packets flowing inside the network.

In addition, because of the inherent necessity of network coding to operate on unidirectional links, each interface within a router model is designated as a SEND or RECEIVE interface only for the network coded packets while operating regularly with non-network coded packets. For interfaces which only SEND NC packets, any received NC packets were discarded. Conversely, the RECEIVE interfaces intercepted and processed the NC packets.

Network coding - by definition- is intended for implementation in Layer 3, over IP frames. However, we employ network coding at transport layer (instead of IP layer) largely for convenience - it readily allows adding hidden data within the TCP/UDP frame in OPNET, which is invisible to the end-user and to the simulation statistics. As long as LNC performance is in concerned having it implemented at transport layer is the same as IP layer.

• Source Code is available here                                                                                                            
• You can download a step-by-step user guide here.

Data Dissemination using Network Coding

In this project we study role of network coding in data dissemination in a wireless network. We consider two following scenarios

1. VANET Networks: In this problem a specific node (called access point) contains a bunch of packets to share to all nodes in a linear network. Number of nodes in the network may be unlimited. We are interested in propagation speed, how fast information will delivered to an indivisual in the network. Clearly the goal is to maximize propagation speed.
2. In the second scenario we consider a wireless network with general topology. Each node in that network has a message to communicate with the rest of the network. We are interested in amount of time needed to disseminate information in the network when every node has all information.

• Mohammad H. Firooz, Sumit Roy, 'Collaborative Downloading in VANET using Network Coding,' IEEE ICC, (accepted), 2012.
• Mohammad H. Firooz, Sumit Roy, 'Data Dissemination in Wireless Networks with Network Coding,' preparing.

Network Tomography

Network Tomography via Compressed Sensing:

In (edge-oriented) network tomography, probes are sent between two boundary nodes on pre-determined routes; typically these are the shortest paths between the nodes. For some parameters like delay  an additive linear model adequately captures the relation between end-to-end and individual link delays, and can be written as

   y=Rx

where x is the n-by-1 vector of individual link delays. The r-by-n binary matrix R denotes the routing matrix for the network graph corresponding to the measurements and y is the measured r-vector of end-to-end path delays.

In Eq. (1), typically, the number of observations r << n, because the number of accessible boundary nodes is much smaller than number of links inside the network. Thus the number of variables in Eq. (1) to be estimated is much larger than number of equations (rank(R) < n) in the linear model, leading to generic non-uniqueness for any solution to Eq. (1),i.e., inability to uniquely specify links delay. A reasonable assumption to make the equation solvable is to assume x is k-compressible.

Over recent years a new approach for obtaining a succinct approximate representation of n-dimensional vectors (or signals) has been discovered. For any signal x, the representation is equal to Rx, where R is carefully chosen r-by-n matrix which is often referred to as measurement matrix. The main challenge in compressed sensing area is to design R with desirable properties, such as maximum possible compression or fast decoding time.

• You may find a presentation for this work here.
• Mohammad H. Firooz, Sumit Roy, "Network Tomography via Compressed Sensing", accepted to IEEE Globecom 2010.
• Mohammad H. Firooz, Sumit Roy, "Delay Estimation via Expander Graphs", preparing,

PHY/MAC-Layer Network Coding implementation on SDR

  Abstract: