Ingo Kofler

An evaluation of TCP-based rate-control algorithms for adaptive internet streaming of H.264/SVC

Recent work in TCP video streaming indicates that multimedia streaming via TCP provides satisfactory performance when the achievable TCP throughput is approximately twice the media bit rate. However, these conditions may not be achievable on the Internet, e.g., when the delivery path offers insufficient bandwidth or becomes congested due to competing traffic. Therefore, adaptive streaming for videos over TCP is required and a number of rate-control algorithms for video streaming have been proposed and evaluated in the literature.\\ In this paper, we evaluate and compare three existing rate-control algorithms for TCP streaming in terms of the (PSNR) quality of the delivered video and in terms of the timeliness of delivery. The contribution of the paper is that, to the best of our knowledge, this is the first evaluation of TCP-based streaming in an Internet-like setting making use of the scalability features of the H.264/SVC video codec. Two simple bandwidth estimation algorithms and a priority-/deadline-driven approach are described to adapt the bit rates of, and transmit, the H.264/SVC video in a rate-distortion optimal manner. The results indicate that the three algorithms perform robustly in terms of video quality and timely delivery, both on under-provisioned links and in case of competing TCP flows. The priority-/deadline-driven technique is even more stable in terms of packet delays and jitter; thus, client buffers can be dimensioned more easily.

Improving Internet Video Streaming Performance by Parallel TCP-Based Request-Response Streams

TCP-based video streaming encounters difficulties in unreliable networks with unanticipated packet loss. In combination with high round trip times, the effective throughput deteriorates rapidly and TCP connection resets or stalls may occur. In this paper, we propose a client-driven video transmission scheme which utilizes multiple HTTP/TCP streams. The scheme is largely insensitive to unanticipated packet loss and thereby reduces throughput fluctuations. Since it is based on HTTP, the scheme can easily be deployed in existing network infrastructures. It fosters scalability on the server side by shifting complexity from the server to the clients. Certain features of request-response schemes allow maintaining fairness, despite of using multiple HTTP streams. Making use of TCP, the scheme inherently adapts to congested network links.

Evaluation of HTTP-based request-response streams for internet video streaming

Adaptive video streaming based on TCP/HTTP is becoming popular because of its ability to adapt to changing network conditions. We present an in-depth experimental analysis of the use of HTTP-based request-response streams for video streaming. In this scheme, video fragments are fetched by a client from the server, in smaller units called chunks, potentially via multiple parallel HTT P requests (TCP connections). A model for the achievable throughput is formulated.The model is validated by a broad range of streaming experiments, including an evaluation of TCP-friendliness. Our findings include that request-response streams are able to scale with the available bandwidth by increasing the chunk size or the number of concurrent streams. Several combinations of system parameters exhibiting TCP-friendliness are presented. We also evaluate the video streaming performance in terms of video quality in the presence of packet loss. Multiple request-response streams are able to maintain satisfactory performance, while a single TCP connection deteriorates rapidly with increasing packet loss. The results provide experimental evidence that HTTP-based request-response streams are a good alternative to classical TCP streaming

Design options and comparison of in-network H.264/SVC adaptation

This paper explores design options and evaluates implementations of in-network, RTP/RTSP based adaptation MANEs (Media Aware Network Elements) for H.264/SVC content streaming. The obvious technique to be employed by such an adaptation MANE is to perform SVC specific bitstream extraction or truncation. Another mechanism that can be used is description (metadata) driven, coding format independent adaptation based on generic Bitstream Syntax Descriptions (gBSD), as specified within MPEG-21 Digital Item Adaptation (DIA). Adaptation MANE architectures for both approaches are developed and presented, implemented in end-to-end streaming/adaptation prototype systems, and experimentally evaluated and compared. For the gBSD based solution, open issues like the granularity of bitstream descriptions and of bitstream adaptation, metadata overhead, metadata packetization and transport options, and error resilience in case of metadata losses, are addressed. The experimental results indicate that a simple SVC specific adaptation MANE does clearly outperform the gBSD based adaptation variants. Yet, the conceptual advantages of the description driven approach, like coding format independence and flexibility, may outweigh the performance drawbacks in specific applications.

An H.264/SVC-based adaptation proxy on a WiFi router

Recent advances in video coding technology like the scalable extension of the MPEG-4 AVC/H.264 video coding standard pave the way for computationally cheap adaptation of video content. In this paper we present our work on a lightweight RTSP/RTP proxy that enables in-network stream processing. Based on an off-the-shelf wireless router that runs a Linux-based firmware we demonstrate that the video adaptation can be performed on-the-fly directly on a network device. The paper covers design and implementation details of the proxy as well as a discussion about the actual adaptation of the SVC stream. Based on experimental evaluations we show that our approach can handle a reasonable number of concurrent sessions for a typical home deployment scenario. Furthermore, the paper covers possible applications in which adaptation on the network device can be beneficial.

Efficient MPEG-21-based adaptation decision-taking for scalable multimedia content

The MPEG-21 standard defines a framework for the interoperable delivery and consumption of multimedia content. Within this framework the adaptation of content plays a vital role in order to support a variety of terminals and to overcome the limitations of the heterogeneous access networks. In most cases the multimedia content can be adapted by applying different adaptation operations that result in certain characteristics of the content. Therefore, an instance within the framework has to decide which adaptation operations have to be performed to achieve a satisfactory result. This process is known as adaptation decision-taking and makes extensive use of metadata describing the possible adaptation operations, the usage environment of the consumer, and constraints concerning the adaptation. Based on this metadata a mathematical optimization problem can be formulated and its solution yields the optimal parameters for the adaptation operations. However, the metadata is represented in XML resulting in a verbose and inefficient encoding. In this paper, an architecture for an Adaptation Decision-Taking Engine (ADTE) is introduced. The ADTE operates both on XML metadata and on metadata encoded with MPEGs Binary Format for Metadata (BiM) enabling an efficient metadata processing by separating the problem extraction from the actual optimization step. Furthermore, several optimization algorithms which are suitable for scalable multimedia formats are reviewed and extended where it was appropriate.,

Towards QoS Improvements of TCP-Based Media Delivery

The amount of audiovisual data available on the Internet and thus of multimedia communication over todays networks is increasing at a rapid pace. Despite the availability of specific media transport protocols like RTP, most content providers make use of the well-established and reliable TCP protocol to deliver audiovisual content over the Internet. The reason is that TCP-based data delivery in general is much less complicated for the clients to be served and over todays networks traversed (including proxies and firewalls), than making use of UDP-based RTP connections. However, in case of network bandwidth fluctuations and packet losses, TCP-based media delivery may lead to annoying jerky playback at the client side, due to retransmissions and late arrival of media data. This papers deals with TCP-based perceptual QoS improvement mechanisms for increasing the media experience for the consumer under unstable network conditions. Our approach is based on media content adaptation (transcoding) to fit the actual network bandwidth continuously monitored by the sender. The proposed mechanisms are applied at the application level at the server side, leaving the existing TCP implementation untouched and therefore enabling transparent use of existing media players. An evaluation of a realistic use case is presented which underlines the efficacy of our approach.

Towards MPEG-21-Based Cross-Layer Multimedia Content Adaptation

Cross-layer designs are becoming more and more attractive within the multimedia community since multiple-play services pave their way towards consumer markets enabling mobility in various aspects. However, cross-layer designs so far have mainly focused on performance issues and do not provide much support in terms of interoperability which is a requirement for services envisaged as part of the Fixed-Mobile Service Convergence (FMSC) initiative. This paper presents a first attempt towards increasing the interoperability of cross-layer designs by adopting an open standard - MPEG-21 Digital Item Adaptation - for describing the functional dependencies across network layers. In this paper a three-step approach for multimedia content adaptation is presented that introduces an MPEG-21-based cross- layer architecture.

Implications of the ISO base media file format on adaptive HTTP streaming of H.264/SVC

HTTP streaming has gained significant attraction in the last few years. Currently many commercial as well as standardized streaming systems are already offering adaptive streaming. In most cases, the adaptation is achieved by switching between separately encoded video streams in different qualities. In contrast to that, this paper focuses on the applicability of scalable video coding based on the H.264/SVC standard for adaptive HTTP streaming. Recent work has already highlighted the conceptual advantages like better cache utilization, fine-grained bit rate scalability, and lower storage requirements. This paper discusses the actual realization and design options for implementing priority streaming using the ISO Base Media File Format (BMFF). We propose three different strategies for organizing the scalable video bit stream that consider both the possibilities as well as limitations of the ISO BMFF. The proposed strategies are discussed and evaluated both conceptually and quantitatively. For that purpose, we provide a detailed analysis based on modeling both the overhead of the file format and the HTTP encapsulation. The results for all three priority streaming strategies show that the limitations of the ISO BMFF result in a high relative overhead in the case of low bit rate content. However, when applied to high quality content, priority streaming of H.264/SVC can be implemented at a very low cost. Depending on the number of layers and the offered scalability dimensions, different strategies should be chosen to minimize the overhead. Based on the analytical model and the discussion, this paper provides guidance for selecting the most efficient strategy.

Improving IPTV services by H.264/SVC adaptation and traffic control

This paper presents a novel approach that combines both in-network, application-layer adaptation and network-layer traffic control of scalable video streams based on the H.264/SVC standard. In the IPTV/VoD scenario considered, an intercepting RTSP/RTP proxy performs admission control of the requested video, based on the signaled scalability information, and decides whether the content can be streamed without changes or in an adapted version. The proxy configures the network layer appropriately in order to separate the video stream from best-effort traffic on the same link. Rather than performing fixed bandwidth allocation, our proxy approach uses the Hierarchical Token Bucket (HTB) queuing discipline to allow for borrowing bandwidth between traffic classes. In that setting, two different allocation policies are introduced. The Hard Reservation Policy (HRP) performs admission control and adaptation on the video streams and does not modify video bandwidth allocation after admission. In contrast, the Flexible Borrowing Policy (FBP) restricts the admission control to the base layer of the SVC stream. The packets carrying MGS enhancement layer data are marked with priorities by the proxy and are handled at the network layer by a priority-based queuing mechanism. Both a qualitative comparison and an experimental evaluation of the two policies are given.