Youtube traffic characterization: a view from the edge

By Phillipa Gill, Martin F. Arlitt, Zongpeng Li, Anirban Mahanti. ACM/SIGCOMM Internet Measurement Conference, 2007

1,040 citations recorded by the source corpus.

This paper presents a traffic characterization study of the popular video sharing service, YouTube. Over a three month period we observed almost 25 million transactions between users on an edge network and YouTube, including more than 600,000 video downloads. We also monitored the globally popular videos over this period of time. In the paper we examine usage patterns, file properties, popularity and referencing characteristics, and transfer behaviors of YouTube, and compare them to traditional Web and media streaming workload characteristics. We conclude the paper with a discussion of the implications of the observed characteristics. For example, we find that as with the traditional Web, caching could improve the end user experience, reduce network bandwidth consumption, and reduce the load on YouTube's core server infrastructure. Unlike traditional Web caching, Web 2.0 provides additional meta-data that should be exploited to improve the effectiveness of strategies like caching.

Subfields: Artificial Intelligence, Computer Networks and Communications

View Paper: https://www.semanticscholar.org/paper/0e047cd129be9f0b3a81f75ebf26e7afcdbcaa82

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Networks and Communications

Citations in the corpus

Confused, timid, and unstable: picking a video streaming rate is hard (2012, background): “Other work looks at network characteristics of video streaming traffic, rather than focusing on the client or viewer experiences [11, 19, 24].”
Understanding the impact of network dynamics on mobile video user engagement (2014, background): “First, we quantify the e.ect that 31 di.erent network factors have on user behavior in mobile video.”
Mahimahi: Accurate Record-and-Replay for HTTP (2015, methodology): “Beyond its widespread use as an application-layer protocol for loading Web pages, HTTP is now used for mobile apps [22], video streaming [14], and instant messaging [19].”
YouTube everywhere: impact of device and infrastructure synergies on user experience (2011, neutral): “We investigate the users behavior and correlateit withthe systemperformance.”
Caching performance of content centric networks under multi-path routing (and more) (2011, background): “V concludes the paper.”
Understanding online social network usage from a network perspective (2009, background): “[13] study patterns of access to YouTube from a campus perspective.”
A versatile and accurate approximation for LRU cache performance (2012, methodology): “Examples of content types with reported Zipf law behaviour are web pages [1], [14], files shared using BitTorrent [7], YouTube documents [9], [2], video on demand movies [19].”
Predicting the popularity of online content (2008, background): “Youtube is the pinnacle of user-created video sharing portals on the Web, with 65,000 new videos uploaded and 100 million downloaded on a daily basis, implying that that 60% of all online videos are watched through the portal [11].”
LIPSIN: line speed publish/subscribe inter-networking (2009, background): “Second, YouTube video popularity also shows a power-law distribution in a campus network [17].”
Statistics and Social Network of YouTube Videos (2008, background): “work complements works in [25], [26], [28] by crawling a much larger set of the videos and thus being able to accurately measure their global properties and, in particular, the social”
From QoS to QoE: A Tutorial on Video Quality Assessment (2015, background): “It has been confirmed in many papers that user access has a clear and consistent daily or weekly pattern [55], [145], [146], [148].”
Key differences between Web 1.0 and Web 2.0 (2008, background): “Recent work [11, 31] examining popularity of YouTube in campus environments showed that local and global popularity of video clips are significantly different.”
Understanding the impact of video quality on user engagement (2011, background): “Inthe spirit of Herbert Simon’s articulation of attention economics, the overabundance of video content increases the onus on content providers to maximize their ability to attract users’ attention [36].”
Scalable and Private Media Consumption with Popcorn (2016, background): “Since media objects are typically large (at least a few MBs [5, 35]), we assume that ` n.”
Less pain, most of the gain: incrementally deployable ICN (2013, background): “Many measurement studies have observed heavy-tailed or Zipf distributions (i.e., the i th popular object has a request probability proportional to 1 i α for some α > 0 ) in request popularities (e.g., [7, 20]).”
Characteristics of YouTube network traffic at a campus network - Measurements, models, and implications (2009, background): “The synthetic traces can be used as input to simulations that could, for example, show how various distribution infrastructures perform when the viewer population and the request arrival rates for YouTube videos grow in the near future.”
On dominant characteristics of residential broadband internet traffic (2009, background): “More recently, various efforts have aimed at understanding from passive measurements how the rapid advent of “Web 2.0” applications has changed HTTP traffic patterns [44], as well as Web based applications such as YouTube [ 20 , 57] and online social networks [21, 36].”
Analyzing the Video Popularity Characteristics of Large-Scale User Generated Content Systems (2009, background): “analyzed YouTube traffic generated by a particular collection of clients, providing a detailed view of local UGC service usage [33].”
Network characteristics of video streaming traffic (2011, background): “[17] observe that 80% of the video interruptions in a campus network are due to lack of user interest.”