Tom Bostoen

Power-reduction techniques for data-center storage systems

As data-intensive, network-based applications proliferate, the power consumed by the data-center storage subsystem surges. This survey summarizes, organizes, and integrates a decade of research on power-aware enterprise storage systems. All of the existing power-reduction techniques are classified according to the disk-power factor and storage-stack layer addressed. A majority of power-reduction techniques is based on dynamic power management. We also consider alternative methods that reduce disk access time, conserve space, or exploit energy-efficient storage hardware. For every energy-conservation technique, the fundamental trade-offs between power, capacity, performance, and dependability are uncovered. With this survey, we intend to stimulate integration of different power-reduction techniques in new energy-efficient file and storage systems.

HTTP/2-Based Methods to Improve the Live Experience of Adaptive Streaming

HTTP Adaptive Streaming (HAS) is today the number one video technology for over-the-top video distribution. In HAS, video content is temporally divided into multiple segments and encoded at different quality levels. A client selects and retrieves per segment the most suited quality version to create a seamless playout. Despite the ability of HAS to deal with changing network conditions, HAS-based live streaming often suffers from freezes in the playout due to buffer under-run, low average quality, large camera-to-display delay, and large initial/channel-change delay. Recently, IETF has standardized HTTP/2, a new version of the HTTP protocol that provides new features for reducing the page load time in Web browsing. In this paper, we present ten novel HTTP/2-based methods to improve the quality of experience of HAS. Our main contribution is the design and evaluation of a push-based approach for live streaming in which super-short segments are pushed from server to client as soon as they become available. We show that with an RTT of 300 ms, this approach can reduce the average server-to-display delay by 90.1% and the average start-up delay by 40.1%.

Network-based dynamic prioritization of HTTP adaptive streams to avoid video freezes

HTTP Adaptive Streaming (HAS) is becoming the de-facto standard for video streaming services over the Internet. In HAS, each video is segmented and stored in different qualities. Rate adaptation heuristics, deployed at the client, allow the most appropriate quality level to be dynamically requested, based on the current network conditions. Current heuristics under-perform when sudden bandwidth drops occur, therefore leading to freezes in the video play-out, the main factor influencing users Quality of Experience (QoE). In this article, we propose an Openflow-based framework capable of increasing clients QoE by reducing video freezes. An Openflow-controller is in charge of introducing prioritized delivery of HAS segments, based on feedback collected from both the network nodes and the clients. To reduce the side-effects introduced by prioritization on the bandwidth estimation of the clients, we introduce a novel mechanism to inform the clients about the prioritization status of the downloaded segments without introducing overhead into the network. This information is then used to correct the estimated bandwidth in case of prioritized delivery. By evaluating this novel approach through emulation, under varying network conditions and in several multi-client scenarios, we show how the proposed approach can reduce freezes up to 75% compared to state-of-the-art heuristics.

HTTP/2-Based Adaptive Streaming of HEVC Video Over 4G/LTE Networks

In HTTP Adaptive Streaming, video content is temporally divided into multiple segments, each encoded at several quality levels. The client can adapt the requested video quality to network changes, generally resulting in a smoother playback. Unfortunately, live streaming solutions still often suffer from playout freezes and a large end-to-end delay. By reducing the segment duration, the client can use a smaller temporal buffer and respond even faster to network changes. However, since segments are requested subsequently, this approach is susceptible to high round-trip times. In this letter, we discuss the merits of an HTTP/2 push-based approach. We present the details of a measurement study on the available bandwidth in real 4G/LTE networks, and analyze the induced bit-rate overhead for HEVC-encoded video segments with a sub-second duration. Through an extensive evaluation with the generated video content, we show that the proposed approach results in a higher video quality (+7.5%) and a lower freeze time (-50.4%), and allows to reduce the live delay compared with traditional solutions over HTTP/1.1.

Analysis of disk power management for data-center storage systems

Up to one third of the electricity supplied to a data center is required for the operation of the storage subsystem. A typical data-center workload, characterized by short idle periods, prevents traditional disk power management (DPM) from saving energy. This paper starts with an analysis of DPMs traditional timer-based disk spin-down policy. Next, we examine how a multispeed disk adapts DPM to datacenter workloads. Finally, we determine how to shape the workload to enable DPM on conventional server disks. All analysis is based on an analytical model for the energy consumed by a disk during an idle period. In addition, we assume standby power to be non-negligible. The results are that (1) the competitive ratio of a threshold-based disk spin-down policy depends on the ratio of standby power to idle power, (2) the notion of break-even time can be generalized for multispeed disks, and (3) DPM saves most energy when mean idle time and idle-time variance are maximized. With this analysis, the authors intend to stimulate the design of new data-center file and storage systems that optimally exploit DPM to save energy.

Software-defined network-based prioritization to avoid video freezes in HTTP adaptive streaming

Summary nHTTP adaptive streaming (HAS) is becoming the de facto standard for video streaming services over the Internet. In HAS, each video is segmented and stored in different qualities. Rate adaptation heuristics, deployed at the client, allow the most appropriate quality level to be dynamically requested, based on the current network conditions. It has been shown that state-of-the-art heuristics perform suboptimal when sudden bandwidth drops occur, therefore leading to freezes in the video playout, the main factor influencing users quality of experience (QoE). This issue is aggravated in case of live events, where the client-side buffer has to be kept as small as possible in order to reduce the playout delay between the user and the live signal. In this article, we propose a framework capable of increasing the QoE of HAS clients by reducing video freezes. The framework is based on OpenFlow, a widely adopted protocol to implement the software-defined networking principle. An OpenFlow controller is in charge of introducing prioritized delivery of HAS segments, based on the network conditions and the HAS clients status. Particularly, the HAS clients status is obtained without any explicit clients-to-controller communication, and thus, no extra signaling is introduced into the network. Moreover, this OpenFlow controller is transparent to the quality decision process of the clients, as it assists the delivery of the segments, but it does not determine the actual quality to be requested. In order to provide a comprehensive analysis of the proposed approach, we investigate the performance of the proposed OpenFlow-based framework in the presence of realistic Internet cross-traffic. Particularly, we model two types of applications, namely, HTTP web browsing and progressive download video streaming, which currently represent the majority of Internet traffic together with HAS. By evaluating this novel approach through emulation in several multi-client scenarios, we show how the proposed approach can reduce freeze time for the HAS clients due to network congestion up to 10 times compared with state-of-the-art heuristics, without impacting the performance of the cross-traffic applications. Copyright

Minimizing Energy Dissipation in Content Distribution Networks Using Dynamic Power Management

The growing end-user demand for video services with superior quality on laptops, tablets, and smartphones spurs the deployment of telco content distribution networks (CDNs). Such CDNs provide scalable and bandwidth-efficient video delivery thanks to disk-packed cache servers deployed in the telcos data centers near the clients. However, a sustainable growth of these CDNs may be hindered by their lack of energy proportionality. In this paper we propose to apply dynamic power management (DPM) to the CDNs cache servers and their disks to increase the CDNs energy efficiency. We evaluate DPM using a CDN energy simulator driven by HTTP adaptive-streaming workload traces recorded by an operational CDN delivering IPTV to mobile devices. Even for a minimally-provisioned CDN, we observe a reduction of the energy dissipation by approximately 30% thanks to large cyclic load fluctuations characteristic of IPTV delivery.

Scalable network-based video-freeze detection for HTTP adaptive streaming

HTTP adaptive streaming (HAS) has become a key video delivery technology for mobile and fixed networks. Internet service providers and CDN (Content Delivery Network) providers are interested in network-based monitoring the clients Quality of Experience (QoE) for HAS sessions. In our previous work, we designed a HAS QoE monitoring system based on the sequence of HTTP GET requests collected at the CDN nodes. The system relies on a technique called session reconstruction to retrieve the major QoE parameters without modification of the clients. However, session reconstruction is computationally intensive and requires manual configuration of reconstruction rules. To overcome the limitations of session reconstruction, this paper proposes a scalable machine learning (ML) based scheme that detects video freezes using a few high-level features extracted from the network-based monitoring data. We determine the discriminative features for session representation and assess five potential classifiers. We select the C4.5 decision tree as classifier because of its simplicity, scalability, accuracy, and explainability. To evaluate our solution, we use traces of Apple HTTP Live Streaming video sessions obtained from a number of operational CDN nodes and traces of Microsoft Smooth Streaming video sessions acquired in a controlled lab environment. Experimental results show that an accuracy of about 98%, 98%, and 90% can be obtained for the detection of a video freeze, a long video freeze, and multiple video freezes, respectively. Excluding log parsing, the computational cost of the proposed video-freeze detection is 33 times smaller than needed for session reconstruction.

An HTTP/2 Push-Based Approach for Low-Latency Live Streaming with Super-Short Segments

Over the last years, streaming of multimedia content has become more prominent than ever. To meet increasing user requirements, the concept of HTTP Adaptive Streaming (HAS) has recently been introduced. In HAS, video content is temporally divided into multiple segments, each encoded at several quality levels. A rate adaptation heuristic selects the quality level for every segment, allowing the client to take into account the observed available bandwidth and the buffer filling level when deciding the most appropriate quality level for every new video segment. Despite the ability of HAS to deal with changing network conditions, a low average quality and a large camera-to-display delay are often observed in live streaming scenarios. In the meantime, the HTTP/2 protocol was standardized in February 2015, providing new features which target a reduction of the page loading time in web browsing. In this paper, we propose a novel push-based approach for HAS, in which HTTP/2’s push feature is used to actively push segments from server to client. Using this approach with video segments with a sub-second duration, referred to as super-short segments, it is possible to reduce the startup time and end-to-end delay in HAS live streaming. Evaluation of the proposed approach, through emulation of a multi-client scenario with highly variable bandwidth and latency, shows that the startup time can be reduced with 31.2% compared to traditional solutions over HTTP/1.1 in mobile, high-latency networks. Furthermore, the end-to-end delay in live streaming scenarios can be reduced with 4 s, while providing the content at similar video quality.

A machine learning-based framework for preventing video freezes in HTTP adaptive streaming

Abstract HTTP Adaptive Streaming (HAS) represents the dominant technology to deliver videos over the Internet, due to its ability to adapt the video quality to the available bandwidth. Despite that, HAS clients can still suffer from freezes in the video playout, the main factor influencing users’ Quality of Experience (QoE). To reduce video freezes, we propose a network-based framework, where a network controller prioritizes the delivery of particular video segments to prevent freezes at the clients. This framework is based on OpenFlow, a widely adopted protocol to implement the software-defined networking principle. The main element of the controller is a Machine Learning (ML) engine based on the random undersampling boosting algorithm and fuzzy logic, which can detect when a client is close to a freeze and drive the network prioritization to avoid it. This decision is based on measurements collected from the network nodes only, without any knowledge on the streamed videos or on the clients characteristics. In this paper, we detail the design of the proposed ML-based framework and compare its performance with other benchmarking HAS solutions, under various video streaming scenarios. Particularly, we show through extensive experimentation that the proposed approach can reduce video freezes and freeze time with about 65% and 45% respectively, when compared to benchmarking algorithms. These results represent a major improvement for the QoE of the users watching multimedia content online.