Koushik Kar

Construction of an efficient overlay multicast infrastructure for real-time applications

This paper presents an overlay architecture where service providers deploy a set of service nodes (called MSNs) in the network to efficiently implement media-streaming applications. These MSNs are organized into an overlay and act as application-layer multicast forwarding entities for a set of clients. We present a decentralized scheme that organizes the MSNs into an appropriate overlay structure that is particularly beneficial for real-time applications. We formulate our optimization criterion as a degree-constrained minimum average-latency problem which is known to be NP-hard. A key feature of this formulation is that it gives a dynamic priority to different MSNs based on the size of its service set. Our proposed approach iteratively modifies the overlay tree using localized transformations to adapt with changing distribution of MSNs, clients, as well as network conditions. We show that a centralized greedy approach to this problem does not perform quite as well, while our distributed iterative scheme efficiently converges to near-optimal solutions.

Minimum interference routing of bandwidth guaranteed tunnels with MPLS traffic engineering applications

This paper presents new algorithms for dynamic routing of bandwidth guaranteed tunnels, where tunnel routing requests arrive one by one and there is no a priori knowledge regarding future requests. This problem is motivated by the service provider needs for fast deployment of bandwidth guaranteed services. Offline routing algorithms cannot be used since they require a priori knowledge of all tunnel requests that are to be rooted. Instead, on-line algorithms that handle requests arriving one by one and that satisfy as many potential future demands as possible are needed. The newly developed algorithms are on-line algorithms and are based on the idea that a newly routed tunnel must follow a route that does not interfere too much with a route that may he critical to satisfy a future demand. We show that this problem is NP-hard. We then develop path selection heuristics which are based on the idea of deferred loading of certain critical links. These critical links are identified by the algorithm as links that, if heavily loaded, would make it impossible to satisfy future demands between certain ingress-egress pairs. Like min-hop routing, the presented algorithm uses link-state information and some auxiliary capacity information for path selection. Unlike previous algorithms, the proposed algorithm exploits any available knowledge of the network ingress-egress points of potential future demands, even though the demands themselves are unknown. If all nodes are ingress-egress nodes, the algorithm can still be used, particularly to reduce the rejection rate of requests between a specified subset of important ingress-egress pairs. The algorithm performs well in comparison to previously proposed algorithms on several metrics like the number of rejected demands and successful rerouting of demands upon link failure.

Optimization based rate control for multirate multicast sessions

Multirate multicasting, where the receivers of a multicast group can receive service at different rates, is an efficient mode of data delivery for many real-time applications. We address the problem of achieving rates that maximize the total receiver utility for multirate multicast sessions. This problem not only takes into account the heterogeneity in user requirements, but also provides a unified framework for diverse fairness objectives. We propose two algorithms and prove that they converge to the optimal rates for this problem. The algorithms are distributed and scalable, and do not require the network to know the receiver utilities. We discuss how these algorithms can be implemented in a real network, and also demonstrate their convergence through simulation experiments.

A scalable low-overhead rate control algorithm for multirate multicast sessions

In multirate multicasting, different users (receivers) within the same multicast group can receive service at different rates, depending on the user requirements and the network congestion level. Compared with unirate multicasting, this provides more flexibility to the user and allows more efficient usage of the network resources. We address the rate control problem for multirate multicast sessions, with the objective of maximizing the total receiver utility. This aggregate utility maximization problem not only takes into account the heterogeneity in user requirements, but also provides a unified framework for diverse fairness objectives. We propose an algorithm for this problem and show, through analysis and simulation, that it converges to the optimal rates. In spite of the nonseparability of the problem, the solution that we develop is completely decentralized, scalable and does not require the network to know the receiver utilities. The algorithm requires very simple computations both for the user and the network, and also has a very low overhead of network congestion feedback.

Routing restorable bandwidth guaranteed connections using maximum 2-route flows

Routing with service restorability is very important in multiprotocol label switched (MPLS) networks, and is a necessity in optical networks. For restoration, each connection has an active path and a disjoint backup path. The backup path enables service restoration upon active path failure. For bandwidth efficiency, backups may be shared. This requires that at least the aggregate backup bandwidth used on each link be distributed to nodes performing route computations. If this information is not available, sharing is not possible. Also, one scheme in use for restorability in optical networks is for the sender to transmit simultaneously on the two disjoint paths and for the receiver to choose data from the path with stronger signal. This has the advantage of fast receiver-initiated recovery upon failure but it does not allow backup sharing. We consider the problem of efficient dynamic routing of restorable connections when backup sharing is not allowed. Our objective is to be able to route as many connections as possible for one-at-a-time arrivals and no knowledge of future arrivals. Since sharing cannot be used for achieving efficiency, the goal is to achieve efficiency by improved path selection. We show that by using the minimum-interference ideas used for non-restorable routing, we can develop efficient algorithms that outperform previously proposed algorithms for restorable routing such as routing with the min-hop like objective of finding two disjoint paths with minimum total hop-count. We present two new and efficient algorithms for restorable routing without sharing, and one of them requires only shortest path computations. We demonstrate that both algorithms perform very well in comparison to previously proposed algorithms.

Optimization Based Rate Control for Multipath Sessions

In this paper, we consider the rate control problem for multipath sessions with the objective of maximizing the total user (session) utility. This problem provides a framework in which flow control and routing are jointly optimized. We consider two cases of this problem, and develop two different rate control algorithms for these two cases. The first algorithm is an end-to-end rate control algorithm which requires, on the part of the user, explicit knowledge of the paths that the user uses. The second algorithm is a hop-by-hop rate control algorithm which does not require the user to keep track of the paths it uses. Both the algorithms are distributed and do not require the network to know the user utility functions. We analyze the convergence properties of these algorithms, and discuss how they can be implemented in a real network. Both of these algorithms are computationally simple, and have very low communication overhead.

Load balancing in large-scale RFID systems

A radio frequency identifier (RFID) system consists of inexpensive, uniquely-identifiable tags that are mounted on physical objects, and readers that track these tags (and hence these physical objects) through RF communication. In this paper we, therefore, address this load balancing problem for readers - given a set of tags that are within range of each reader, which of these tags should each reader be responsible for such that the cost for monitoring tags across the different readers is balanced, while guaranteeing that each tag is monitored by at least one reader. We show that a generalized variant of the load balancing problem is NP-hard and hence present a 2-approximation centralized algorithm. We next present an optimal centralized solution for a specialized variant. Subsequently, we present a localized distributed algorithm that is probabilistic in nature and closely matches the performance of the centralized algorithms. Our results demonstrate that our schemes achieve very good performance even in highly dynamic large-scale RFID systems.

A simple rate control algorithm for max total user utility

We consider the rate control problem with the objective of maximizing the total user utility. It takes into account the possible differences in user requirements, and also provides a framework for achieving a wide range of fairness objectives. We propose a simple algorithm for achieving the optimal rates for this problem. The algorithm can be implemented in a distributed way and does not require the network to know the user utility functions. In our algorithm, the network communicates to the user the number of congested links on the users path, and the user (end-host) adjusts its rate accordingly, taking into account its utility function and the network congestion feedback. We show through analysis and experimentation that our algorithm converges to the optimum rates.

MPLS traffic engineering using enhanced minimum interference routing: an approach based on lexicographic max-flow

The main contribution of this paper is a new algorithm for dynamic routing of bandwidth guaranteed tunnels. The algorithm is a generalization of the minimum-interference based algorithm presented in Kodialam and Lakshman (2000). The routing objective is to satisfy as many potential future requests as possible in an environment where requests arrive one-by-one, and there is no knowledge of future arrivals other than their potential ingress-egress points. An application of these algorithms is for the routing of MPLS bandwidth-guaranteed label-switched paths (LSP). The presented algorithm performs better than the previous scheme, and other previously proposed algorithms, on several metrics like the number of rejected demands and successful re-routing of demands upon link failure. Also, the new algorithm is more robust in performance to implementations where most requests are routed with a shortest-path computation, and the computations required for determining minimum-interference are performed only infrequently.

Scalable low-overhead rate control algorithm for multirate multicast sessions

In multirate multicasting, different users (receivers) within the same multicast group could receive service at different rates, depending on user requirements and network congestion level. Compared to unirate multicasting, this provides more flexibility to the user, and allows more efficient usage of network resources. In this paper, we address the rate control problem for multirate multicast sessions, with the objective of maximizing the total receiver utility. This aggregate utility maximization problem not only takes into account the heterogeneity in user requirements, but also provides a unified framework for diverse fairness objectives. We propose an algorithm for this problem and show, through analysis and simulation, that it converges to the optimal rates. In spite of the non-separability of the problem, the solution that we develop is completely decentralized, scalable and does not require the network to know the receiver utilities. The algorithm requires very simple computations both for the user and the network, and also has very low overhead of network congestion feedback. Moreover, the algorithm does not require the network links to maintain per-flow state, and is suitable for deployment in the current internet.