Hyunwoo Nam

Towards QoE-aware video streaming using SDN

Todays over the top (OTT) video service providers take advantage of content distribution networks (CDNs) and adaptive bitrate (ABR) streaming where a video player adjusts resolutions based on end-to-end network conditions. Although the mechanisms are useful to improve user-perceived video quality, they do not resolve the root causes of congestion problems. To pinpoint a bottleneck and improve video quality-of-experience (QoE), we leverage a software-defined networking (SDN) platform from OTT video service providers point of view. Our proposed SDN application is designed to monitor network conditions of streaming flow in real time and dynamically change routing paths using multi-protocol label switching (MPLS) traffic engineering (TE) to provide reliable video watching experience. We use an off-the-shelf SDN platform to show the feasibility of our approaches.

QoE matters more than QoS: Why people stop watching cat videos

With the proliferation of online video, measuring quality of experience (QoE) has become a pivotal aspect for the analysis of todays over-the-top (OTT) video streaming. To monitor video QoE, we introduce YouSlow that can detect various playback events (e.g., start-up latency, rebufferings and bitrate changes) from video players while a video is being played. Using YouSlow, we have collected more than 400,000 YouTube views from more than 100 countries. We measured the impact of these playback events on video abandonment and found that rebufferings incur abandonment rates six times higher than start-up latency, mostly caused by pre-roll ads. A single rebuffering event has three times the impact of a bitrate change. Even increasing the bitrate can raise abandonment rates by a factor of four compared to keeping the bitrate constant.

Intelligent content delivery over wireless via SDN

On the Internet, most traffic is generated by content delivery applications such as video streaming, P2P and web surfing. With the proliferation of mobile applications that offer on-demand access to services, the ability to track end-users perceived service quality has become a pivotal aspect for future networks such as 5G. Software-defined networking (SDN) and network functions virtualization (NFV) in the mobile core offer unprecedented flexibility for opportunistic traffic steering, leading to advanced levels of quality-of-service (QoS) control. We leverage the SDN concept to dynamically control network traffic over wide area networks (WANs) from edge nodes of wireless networks (e.g., a packet data network gateway in LTE), depending on changing network conditions and application types. Our SDN-based application-aware routing system allows mobile network operators to achieve better utilization of their networks, service providers to improve customer satisfaction, and end-users to experience desirable service quality for various network applications. As for a prototype, we attach 802.11g to Mininets virtual networks in order to show the feasibility of our approach.

YouSlow: a performance analysis tool for adaptive bitrate video streaming

Adaptive bitrate (ABR) technologies are being widely used in todays popular HTTP-based video streaming such as YouTube and Netflix. Such a rate-switching algorithm embedded in a video player is designed to improve video quality-of-experience (QoE) by selecting an appropriate resolution based on the analysis of network conditions while the video is playing. However, a bad viewing experience is often caused by the video player having difficulty estimating transit or client-side network conditions accurately. In order to analyze the ABR streaming performance, we developed YouSlow, a web browser plug-in that can detect and report live buffer stalling events to our analysis tool. Currently, YouSlow has collected more than 20,000 of YouTube video stalling events over 40 countries.

DYSWIS: Crowdsourcing a home network diagnosis

Existing failure diagnostic techniques for end users are insufficient to pinpoint the root causes of network failures due to their limited capabilities to probe other network elements. We present DYSWIS, an automatic network fault detection and diagnosis system for end-users. DYSWIS leverages user collaboration to distinguish important network faults from false positive indications, and diagnoses the root cause of the fault using diagnostic rules that consider diverse information from multiple nodes. Our rule system is specially designed to support crowdsourcing and distributed probes. We have implemented DYSWIS and compared its performance with other tools to prove that several network failures which are difficult to be diagnosed by the single-user probe can be detected and diagnosed successfully with our approach.

WiSlow: A Wi-Fi network performance troubleshooting tool for end users

Slow Internet connectivity is often caused by poor Wi-Fi performance. The main reasons of such performance degradation include channel contention and non-Wi-Fi interference. Although these problem sources can be easily removed in many cases once they are discovered, it is difficult for end users to identify the sources of such interference. We investigated the characteristics of different sources that can degrade Wi-Fi performance, and developed WiSlow, a software tool that diagnoses the root causes of poor Wi-Fi performance using user-level network probes, and leveraging peer collaboration to identify the physical location of these causes. WiSlow uses two main methods: packet loss analysis and 802.11 ACK number analysis. The accuracy of WiSlow exceeds 90% when the sources are close to Wi-Fi devices. Also, our experiment proves that the collaborative approach is feasible for determining the relative location of an interfering device.

Towards dynamic QoS-aware over-the-top video streaming

We present a study of traffic behavior of two popular over-the-top (OTT) video streaming services (YouTube and Netflix). Our analysis is conducted on different mobile devices (iOS and Android) over various wireless networks (Wi-Fi, 3G and LTE) under dynamic network conditions. Our measurements show that the video players frequently discard a large amount of video content although it is successfully delivered to a client. We first investigate the root cause of this unwanted behavior. Then, we propose a Quality-of-Service (QoS)-aware video streaming architecture in Long Term Evolution (LTE) networks to reduce the waste of network resource and improve user experience. The architecture includes a selective packet discarding mechanism, which can be placed in packet data network gateways (P-GW). In addition, our QoS-aware rules assist video players in selecting an appropriate resolution under a fluctuating channel condition. We monitor network condition and configure QoS parameters to control availability of the maximum bandwidth in real time. In our experimental setup, the proposed platform shows up to 20.58% improvement in saving downlink bandwidth and improves user experience by reducing buffer underflow period to an average of 32 seconds.

Towards dynamic MPTCP Path control using SDN

MPTCP (MultiPath TCP) boosts application network performance by aggregating bandwidth over multiple paths. However, it may cause poor performance due to the large number of out-of-order packets especially when the paths have different delays. To resolve this issue, we propose to dynamically add or remove MPTCP paths with the leverage of software-defined networking (SDN). The key idea is to track the available capacity of connected paths and pick the most appropriate paths depending on varying network conditions. To show the feasibility of our approach, we create an SDN platform using Mininet over Wi-Fi networks. Our analysis shows that a faster download and improved quality of experience (QoE) in adaptive rate video streaming is possible with our dynamic MPTCP path control mechanism using SDN.

Towards dynamic network condition-aware video server selection algorithms over wireless networks

We investigate video server selection algorithms in a distributed video-on-demand system. We conduct a detailed study of the YouTube Content Delivery Network (CDN) on PCs and mobile devices over Wi-Fi and 3G networks under varying network conditions. We proved that a location-aware video server selection algorithm assigns a video content server based on the network attachment point of a client. We found out that such distance-based algorithms carry the risk of directing a client to a less optimal content server, although there may exist other better performing video delivery servers. In order to solve this problem, we propose to use dynamic network information such as packet loss rates and Round Trip Time (RTT) between an edge node of a wireless network (e.g., an Internet Service Provider (ISP) router in a Wi-Fi network and a Radio Network Controller (RNC) node in a 3G network) and video content servers, to find the optimal video content server when a video is requested. Our empirical study shows that the proposed architecture can provide higher TCP performance, leading to better viewing quality compared to location-based video server selection algorithms.

A mobile video traffic analysis: Badly designed video clients can waste network bandwidth

Video streaming on mobile devices is on the rise. According to recent reports, mobile video streaming traffic accounted for 52.8% of total mobile data traffic in 2011, and it is forecast to reach 66.4% in 2015. We analyzed the network traffic behaviors of the two most popular HTTP-based video streaming services: YouTube and Netflix. Our research indicates that the network traffic behavior depends on factors such as the type of device, multimedia applications in use and network conditions. Furthermore, we found that a large part of the downloaded video content can be unaccepted by a video player even though it is successfully delivered to a client. This unwanted behavior often occurs when the video player changes the resolution in a fluctuating network condition and the playout buffer is full while downloading a video. Some of the measurements show that the discarded data may exceed 35% of the total video content.