Konstantin Miller

Adaptation algorithm for adaptive streaming over HTTP

Internet video makes up a significant part of the Internet traffic and its fraction is constantly growing. In order to guarantee best user experience throughout different network access technologies with dynamically varying network conditions, it is fundamental to adopt technologies enabling a proper delivery of the media content. One of such technologies is adaptive streaming. It allows to dynamically adapt the bit-rate of the stream to varying network conditions. There are various approaches to adaptive streaming. In our work, we focus on the receiver-driven approach where the media file is subdivided into segments, each of the segments is provided at multiple bit-rates, and the task of the client is to select the appropriate bit-rate for each of the segments. With this approach, the challenges are (i) to properly estimate the dynamics of the available network throughput, (ii) to control the filling level of the client buffer in order to avoid underflows resulting in playback interruptions, (iii) to maximize the quality of the stream, while avoiding unnecessary quality fluctuations, and, finally, (iv) to minimize the delay between the users request and the start of the playback. During our work, we designed and implemented a receiver-driven adaptation algorithm for adaptive streaming that does not rely on cross-layer information or server assistance. We integrated the algorithm with a prototype implementation of a streaming client based on the MPEG DASH (Dynamic Adaptive Streaming over HTTP) standard. We evaluated the implemented prototype in real-world scenarios and found that it performes remarkably well even under challenging network conditions. Further, it exhibits stable and fair operation if a common link is shared among multiple clients.

A Control-Theoretic Approach to Adaptive Video Streaming in Dense Wireless Networks

Recently, the way people consume video content has been undergoing a dramatic change. Plain TV sets, that have been the center of home entertainment for a long time, are losing ground to hybrid TVs, PCs, game consoles, and, more recently, mobile devices such as tablets and smartphones. The new predominant paradigm is: watch what I want, when I want, and where I want. The challenges of this shift are manifold. On the one hand, broadcast technologies such as DVB-T/C/S need to be extended or replaced by mechanisms supporting asynchronous viewing, such as IPTV and video streaming over best-effort networks, while remaining scalable to millions of users. On the other hand, the dramatic increase of wireless data traffic begins to stretch the capabilities of the existing wireless infrastructure to its limits. Finally, there is a challenge to video streaming technologies to cope with a high heterogeneity of end-user devices and dynamically changing network conditions, in particular in wireless and mobile networks. In the present work, our goal is to design an efficient system that supports a high number of unicast streaming sessions in a dense wireless access network. We address this goal by jointly considering the two problems of wireless transmission scheduling and video quality adaptation, using techniques inspired by the robustness and simplicity of proportional-integral-derivative (PID) controllers. We show that the control-theoretic approach allows to efficiently utilize available wireless resources, providing high quality of experience (QoE) to a large number of users.

Utility Max-Min Fair Congestion Control with Time-Varying Delays

We present a framework for designing delay- independent end-to-end congestion control algorithms, where each end-user may have a different utility function. We only require that utility functions are strictly increasing. In this framework, we design an algorithm that maximizes the minimum utility value in the network, that is, the resulting resource allocation is utility max-min fair. To achieve this, we first extend the congestion control algorithm EMKC proposed by Zhang et al. [1], which aims at max-min fair bandwidth allocation. Our extension (xMKC) allows for arbitrary rate allocations in the steady state. We investigate xMKC analytically and prove local asymptotic stability with heterogeneous time-varying feedback delays in multi-link networks and global asymptotic stability with homogeneous time-varying feedback delays in single-link networks. Then, we propose uMKC (Utility Max-Min Fair Kelly Control), which achieves utility max-min fairness in its steady state. Based on the analysis of xMKC, we establish stability results for uMKC in the presence of time-varying feedback delays. Finally, we evaluate the performance of uMKC using NS-2 simulations [2].

Optimal Adaptation Trajectories for Block-Request Adaptive Video Streaming

Block-Request Adaptive Streaming (BRAS), in form of its most prominent representative HTTP-Based Adaptive Streaming (HAS), is about to become the dominating technology for video delivery over the Internet. One of the challenges in the development of BRAS clients is the design of mechanisms that dynamically adapt the streamed video quality to network conditions, in order to maximize users Quality of Experience (QoE). The main contribution of this paper is an approach to calculating optimal adaptation trajectories. This approach not only allows to benchmark the performance of any streaming client, it also provides the possibility to study the impact of the networking environment, and of configuration parameters such as the start-up delay, number of available video representations, etc., on the achievable streaming performance. Since, to the best of our knowledge, there exist no widely accepted or standard approach to measure QoE for BRAS, we alternatively maximize the average video bit-rate, minimize the number of quality switches, and impose a hard constraint on the absence of rebuffering events. Further, we evaluate two HAS clients, Microsoft SmoothStreaming and our own streaming client that supports the recently adopted HAS standard Dynamic Adaptive Streaming over HTTP (DASH), in an indoor Wireless Local Area Network (WLAN) emulated with a high degree of precision. We compare their performance with the optimal client, and explore the configuration parameter space of the DASH client. Finally, we evaluate the impact of start-up delays and number of available video representations on achievable streaming performance.

The Worst-Case Efficiency of Cost Sharing Methods in Resource Allocation Games

Resource allocation problems play a key role in many applications, including traffic networks, telecommunication networks, and economics. In most applications, the allocation of resources is determined by a finite number of independent players, each optimizing an individual objective function. An important question in all these applications is the degree of suboptimality caused by selfish resource allocation. We consider the worst-case efficiency of cost sharing methods in resource allocation games in terms of the ratio of the minimum guaranteed surplus of a Nash equilibrium and the maximal surplus. Our main technical result is an upper bound on the efficiency loss that depends on the class of allowable cost functions and the class of allowable cost sharing methods. We demonstrate the power of this bound by evaluating the worst-case efficiency loss for three well-known cost sharing methods: incremental cost sharing, marginal cost pricing, and average cost sharing.

QoE-Based Low-Delay Live Streaming Using Throughput Predictions

Recently, Hypertext Transfer Protocol (HTTP)-based adaptive streaming has become the de facto standard for video streaming over the Internet. It allows clients to dynamically adapt media characteristics to the varying network conditions to ensure a high quality of experience (QoE)—that is, minimize playback interruptions while maximizing video quality at a reasonable level of quality changes. In the case of live streaming, this task becomes particularly challenging due to the latency constraints. The challenge further increases if a client uses a wireless access network, where the throughput is subject to considerable fluctuations. Consequently, live streams often exhibit latencies of up to 20 to 30 seconds. In the present work, we introduce an adaptation algorithm for HTTP-based live streaming called LOLYPOP (short for low-latency prediction-based adaptation), which is designed to operate with a transport latency of a few seconds. To reach this goal, LOLYPOP leverages Transmission Control Protocol throughput predictions on multiple time scales, from 1 to 10 seconds, along with estimations of the relative prediction error distributions. In addition to satisfying the latency constraint, the algorithm heuristically maximizes the QoE by maximizing the average video quality as a function of the number of skipped segments and quality transitions. To select an efficient prediction method, we studied the performance of several time series prediction methods in IEEE 802.11 wireless access networks. We evaluated LOLYPOP under a large set of experimental conditions, limiting the transport latency to 3 seconds, against a state-of-the-art adaptation algorithm called FESTIVE. We observed that the average selected video representation index is by up to a factor of 3 higher than with the baseline approach. We also observed that LOLYPOP is able to reach points from a broader region in the QoE space, and thus it is better adjustable to the user profile or service provider requirements.

Efficiency and stability of Nash equilibria in resource allocation games

We study resource allocation games, where users send data along paths and links in the network charge a price equal to marginal cost. When users are price taking, it is known that there exist distributed dynamics that converge towards a fully efficient Nash equilibrium. When users are price anticipating, however, a Nash equilibrium does not maximize total utility in general. In this paper, we explore the inefficiency of Nash equilibria for general networks and semi-convex marginal cost functions. While it is known that for m ≥ 2 users and convex marginal cost functions, no efficiency guarantee is possible, we prove that an additional differentiability assumption on marginal cost functions implies a bounded efficiency loss of 2/(2m + 1). For polynomial marginal cost functions with nonnegative coefficients, we precisely characterize the price of anarchy. We also prove that the efficiency of Nash equilibria significantly improves if all users have the same strategy space and the same utility function. We propose a class of distributed dynamics and prove that whenever a game admits a potential function, these dynamics globally converge to a Nash equilibrium. Finally, we show that in general the only class of marginal cost functions that guarantees the existence of a potential function are affine linear functions.

Transport Optimization in Peer-to-Peer Networks

The peer-to-peer networking concept has revolutionized the cost structure of Internet data dissemination by making large scale content delivery with low server cost feasible. In a peer-to-peer network, the total upload capacity increases with the number of down loaders instead of staying constant as in a client-server architecture, making it highly scalable. Despite of its importance, the problem of efficient data transport in a peer-to-peer network is still an open issue, mainly due to its complex combinatorial structure. In the presented work, we formulate the problem of optimizing a peer-to-peer download with respect to its make span (time until all peers finish downloading)as a mixed integer linear program. Other than previous studies, we consider the case of arbitrary heterogeneous uplink and downlink capacities of the peers. Moreover, we do not consider the fluid limit case but allow the file to be subdivided in finitely many chunks. On the one hand, our results allow to infer the capacity of a peer-to-peer network, providing a benchmark for performance analysis of existing peer-to-peer protocols. On the other hand, we believe that our results build a step towards the development of efficient algorithms serving as a base for the design of data transport protocols leveraging the peer-to-peer concept.

Optimal file distribution in peer-to-peer networks

We study the problem of distributing a file initially located at a server among a set of peers. Peers who downloaded the file can upload it to other peers. The server and the peers are connected to each other via a core network. The upload and download rates to and from the core are constrained by user and server specific upload and download capacities. Our objective is to minimize the makespan. We derive exact polynomial time algorithms for the case when upload and download capacities per peer and among peers are equal. We show that the problem becomes strongly NP-hard for equal upload and download capacities per peer that may differ among peers. For this case we devise a polynomial time

Simulation Framework for HTTP-Based Adaptive Streaming Applications

(1+2\!\sqrt{2})