Unicast Sessions Research Articles

Self-organizing algorithms for cache cooperation in content distribution networks

Introduction: The on-demand delivery of video content is anticipated to show tremendous growth over the next few years, driven by the huge popularity of user-generated video clips and the expansion of VoD libraries. Even stronger growth is likely to be fueled by the proliferation of IPTV services with personalized features such as CatchUp/PauseLive TV and NPVR capabilities. The common characteristic of these ‘time-shifted’ TV services, as well as VoD libraries and sites like YouTube, is that users can select from a vast collection of content material at any time they want. This runs counter to the broadcast paradigm embraced in conventional TV networks, and raises a need for unicast sessions, increasing the overall bandwidth demands by orders of magnitude. Caching strategies provide an effective mechanism for mitigating the massive bandwidth requirements associated with large-scale distribution of personalized high-definition video content. In essence, caching strategies exploit storage capacity to absorb traffic by replicating the most popular content closer to the network edge rather than storing it in a central location. The reduction in the traffic load translates into a smaller required transport capacity and capital expense as well as fewer performance bottlenecks, thus enabling better service quality at a lower price. In the present paper we develop light-weight distributed cooperative content placement algorithms so as to maximize the traffic volume served from cache, and thus minimize the bandwidth cost. As a canonical scenario, we focus on a cluster of distributed caches, connected either directly, or via a parent node, and formulate the content placement problem as an integer linear program in order to obtain a benchmark of the globally optimal performance. Under certain symmetry assumptions, the optimal solution of the relaxed linear program is found to have a rather simple structure. Besides interesting in its own right, the knowledge of the optimal structure offers useful insight for the design of low-complexity cooperative content placement and eviction algorithms. The performance of the proposed distributed algorithms is proved to be within a small constant factor from the performance of an optimal centralized scheme, while requiring only limited knowledge of global popularities, even in asymmetric scenarios. In contrast to [1, 2], we will focus on specific topologies, motivated by real system deployments, and use the insight from the linear program and the simple structure of its optimal solution to develop efficient distributed content placement algorithms with far more favorable performance ratios than the ones in [1, 2].

Optimized multipath network coding in lossy wireless networks

Network coding has been a prominent approach to a series of problems that used to be considered intractable with traditional transmission paradigms. Recent work on network coding includes a substantial number of optimization based protocols, but mostly for wireline multicast networks. In this paper, we consider maximizing the benefits of network coding for unicast sessions in lossy wireless environments. We propose optimized multipath network coding (OMNC), a rate control protocol that dramatically improves the throughput of lossy wireless networks. OMNC employs multiple paths to push coded packets to the destination, and uses the broadcast MAC to deliver packets between neighboring nodes. The coding and broadcast rate is allocated to transmitters by a distributed optimization algorithm that maximizes the advantage of network coding while avoiding congestion. With extensive experiments on an emulation testbed, we find that OMNC achieves more than two-fold throughput increase on average compared to traditional best path routing, and significant improvement over existing multipath routing protocols with network coding. The performance improvement is notable not only for one unicast session, but also when multiple concurrent unicast sessions coexist in the network.

Cross-layer optimization for wireless multihop networks with pairwise intersession network coding

For wireless multi-hop networks with unicast sessions, most coding opportunities involve only two or three sessions as coding across many sessions requires greater transmission power to broadcast the coded symbol to many receivers, which enhances interference. This work shows that with a new flow-based characterization of pairwise intersession network coding (coding across two unicast sessions), an optimal joint coding, scheduling, and rate-control scheme can be devised and implemented using only the binary XOR operation. The new scheduling/rate-control scheme demonstrates provably graceful throughput degradation with imperfect scheduling, which facilitates the design tradeoff between the throughput optimality and computational complexity of different scheduling schemes. Our results show that pairwise intersession network coding improves the throughput of non-coding solutions regardless of whether perfect/imperfect scheduling is used. Both the deterministic and stochastic packet arrivals and departures are considered. This work shows a striking resemblance between pairwise intersession network coding and non-coded solutions, and thus advocates extensions of non-coding wisdoms to their network coding counterpart.

Understanding and Exploiting the Trade-Offs between Broadcasting and Multicasting in Mobile Ad Hoc Networks

We find that current group communications protocols are far from "one size fits all", they are typically geared toward and optimized for particular scenarios. Multicasting, in general, works well if the density of group members is sparse and in low mobility; broadcasting, in contrast, works well with a high density of group members and in high mobility. Due to the dynamics of the network, one strategy may be preferable to the other at different times and in different localized regions. In this paper, we first quantify the trade-offs between broadcasting and multicasting and evaluate the suitability of a strategy in various scenarios of deployment. Based on the lessons learned, we design a protocol that adapts in response to the dynamics of the network. We named our protocol Fireworks. Fireworks is a hybrid two-tier multicast/broadcast protocol that provides efficient and lightweight multicast dissemination and self-adapts in response to variations in the density and distribution of group members to provide efficient performance. Fireworks creates pockets of broadcast distribution in areas with many members, while it creates and maintains a multicast backbone to interconnect these dense pockets. Fireworks offers packet delivery statistics comparable to that of a pure multicast scheme but with significantly lower overheads. We also show that Fireworks has a lower level of degrading influence on the performance of coexisting unicast sessions than either traditional multicast or broadcast methods

On the capacity of multiple unicast sessions in undirected graphs

Li and Li conjectured that in an undirected network with multiple unicast sessions, network coding does not lead to any coding gain. Surprisingly enough, this conjecture could not so far be verified even for the simple network consisting of K/sub 3,2/ with four source-sink pairs. Using entropy calculus, we provide the first verification of the Li-Li conjecture for this network. We extend our bound to the case of an arbitrary directed bipartite network.

Measurement-based optimal routing on overlay architectures for unicast sessions

We propose a measurement-based routing algorithm to load-balance intradomain traffic along multiple paths for multiple unicast sources. Multiple paths are established using overlay nodes. The distributed algorithm is derived from simultaneous perturbation stochastic approximation (SPSA) and does not assume that the gradient of an analytical cost function is known. Instead, we rely on (potentially) noisy estimates from local measurements. Using the analytical model presented in the paper, we first show the almost sure convergence of the algorithm to the optimal solution under a decreasing step size policy (as with a standard SPSA model). Motivated by practical concerns, we next consider the constant step size case, for which we establish weak convergence. We provide simulation results to demonstrate the advantages of our proposed algorithm under various network scenarios, and also present a comparative study with MATE (an existing optimal routing algorithm).

Congestion Control for Fair Resource Allocation in Networks With Multicast Flows

We consider the problem of congestion control in networks which support both multirate multicast sessions and unicast sessions. We present a decentralized algorithm which enables the different rate-adaptive receivers in different multicast sessions to adjust their rates to satisfy some fairness criterion. A one-bit ECN marking strategy to be used at the nodes is also proposed. The congestion-control mechanism does not require any per-flow state information for unicast flows at the nodes. At junctions nodes of each multicast tree, some state information about the rates along the branches at the node may be required. The congestion-control mechanism takes into account the diverse user requirements when different receivers within a multicast session have different utility functions, but does not require the network to have any knowledge about the receiver utility functions.

Configuring sessions in programmable networks

The provision of advanced computational services within networks is rapidly becoming both feasible and economical. We present a general approach to the problem of configuring application sessions that require intermediate processing by showing how the session configuration problem can be transformed to a conventional shortest path problem for unicast sessions or to a conventional Steiner tree problem for multicast sessions. We study both a capacity-constrained version of the problem and an unconstrained version and show, through a series of examples, that the method can be applied to a wide variety of different situations. Particularly, we show how to extend Dijkstra’s shortest path algorithm for use on the constrained version, and show that this approach can make significantly better use of network resources.

Optimization-based congestion control for multicast communications

This article outlines an approach for multicast congestion control based on an economic model that has been successfully applied to unicast congestion control. In this model, congestion signals are interpreted as prices and congestion-controlled sessions as utility maximizing agents. A naive extension of the unicast model fails to achieve a reasonable balance between providing the incentives necessary to promote the use of multicast and ensuring that multicast sessions do not interact too aggressively with unicast sessions. We extend the model by introducing a rational definition of multicast utility. The revised model provides a basis for multicast congestion control protocols that provide incentives to use multicast but are necessarily unfair to unicast traffic. We show, however, that the degree of unfairness can be controlled by appropriately setting a design parameter with a limiting case of strict fairness.

The multi-session bridge

The Multicast BackbONE (MBone) is heavily used to carry Internet video conferences. The low cost and ease-of-use have made the MBone and its associated tools such as sdr, vat, vic, and wb very popular. But HEP video conferences often require many-to-many sparse network topologies, an arrangement that cannot easily take advantage of MBone efficiencies. Moreover, many HEP sites do not support multicast and thus researchers at those sites are cut off from important meetings. In addition, the topology of the MBone can cause poor performance or lack of access for some sites. To solve these problems, HEPNRC has developed the Multi-Session Bridge (msb) and its companion tool vtc. The msb allows an arbitrary number of multicast and unicast sessions to be connected in one or more seamless video conferences. Multiple msb's can also be connected to form a conference virtual network. The msb also extends MBone sessions to non-MBone sites through unicast streams. The msb is being extended to provide secure communications.