iPlayer and catch-up TV G. Nencioni, N. Sastry, J. Chandaria, J. - PowerPoint PPT Presentation

Towards a greener and lower network footprint iPlayer and catch-up TV G. Nencioni, N. Sastry, J. Chandaria, J. Crowcroft Uni. Pisa, King’s College London , BBC R&D, Cambridge

About me N. Sastry  Rich Media + Social Networks  + Systems support for both   Data data everywhere…(very keen to share)  Video (meta)data  Vimeo (AAAI ICWSM 2012)  YouPorn (ACM SIGCOMM IMC 2013)  Gareth yesterday  BBC iPlayer (WWW 2013)  This talk  Social networks  Twitter (IEEE/ASE Social Informatics)  London Olympics + London Fashion Week  Pinterest + FB (AAAI ICWSM 2013 + Submission to ACM COSN)

http://www.watfordobserver.co.uk/nostalgia/memories/10099510.Coronation_treat_as_community_gathers _around_the_only_TV/ Early use of mass media N. Sastry Picture from the TV broadcast of the Coronation of Elizabeth II in 1953, Watford

Today’s “TV” viewing N. Sastry With Digital Media Convergence, TV is just another video app, accessed on-demand on the Web

What changed: Push  Pull N. Sastry  Superficially: audience to TV set ratio has decreased  At a fundamental level:  audience per “ broadcast” is lower  “Broadcast” time is chosen by the consumer  Traditional mass media pushed content to consumer  Current dominant model has changed to pull Generalizes to other mass media as well

Implications of the pull model N. Sastry  Traditionally, “editors” decided what content got pushed when  Linear TV schedulers use complex analytics to decide “primetime”  Users get more choice with the pull model  When to consume  What to consume (from large catalogue)  Unpopular/niche interest content also gets a distribution channel, not just what editors decide to showcase/bless as “publishable”  Cheaper to stream over the Web to a single user than to broadcast (e.g. to operate/maintain equipment like high power TV transmitters)  BUT: Cost of broadcast can be amortized across millions of consumers  Could be cheaper per user to broadcast than to stream

Understanding and decreasing the network footprint of Catch-up TV N. Sastry  How does pull model impact delivery infrastructure?  Can additional load of on-demand pulls be reduced by reusing scheduled pushes?  How do users make use of flexibility afforded to them?  Were/are editors good at predicting popularity?

Data to answer the questions N. Sastry  Nearly 6 million users of BBC iPlayer across the UK  32.6 million streams, >37K distinct content items  25% sample of BBC iPlayer access over 2 months

Understanding and decreasing the Network Footprint of Catch-up TV- WWW’13 • BBC proposes, consumer disposes! • Serials:~50% of content corpus; 80% of watched content! What users prefer to watch-I

Understanding and decreasing the Network Footprint of Catch-up TV- WWW’13 What users prefer to watch-II

Understanding and decreasing the Network Footprint of Catch-up TV- WWW’13 What users prefer to watch-III

Understanding and decreasing the Network Footprint of Catch-up TV- WWW’13 On-demand spreads load over time Linear TV schedulers seem to do a good job of predicting popularity! Impact of pull on infrastructure

Understanding and decreasing the Network Footprint of Catch-up TV- WWW’13 • BUT: iPlayer traffic is close to 6% of UK peak traffic • Second only to YouTube in traffic footprint • Compare to adult video, a traditional heavy hitter. Most popular adult video streaming sites have <0.2% traffic share • BUT: amortized per-user, broadcast greener than streaming * (using Baliga et al .’s energy model for the Internet) * All channels except BBC Parliament, which has few viewers  On-demand more suited to web/pull than linear TV Still, can we decrease its footprint, please?

Understanding and decreasing the Network Footprint of Catch-up TV- WWW’13 • DVRs have >50% penetration in US, UK • Many (e.g. YouView ) don’t need cable • Could also use TV tuner and record on laptop Yes, we can! But, people don’t remember to record always

Understanding and decreasing the Network Footprint of Catch-up TV- WWW’13 Speculative Content Offloading and Recording Engine Can we help users record what they want to watch?

Understanding and decreasing the Network Footprint of Catch-up TV- WWW’13 • Predict using user affinity for • Episodes of same programme • Favourite genres • We can optimise for decreasing traffic or carbon footprint • Decreasing carbon decreases traffic, but not vice versa • Turns out we only take 5-15% hit by focusing on carbon SCORE=predictor+optimiser

Understanding and decreasing the Network Footprint of Catch-up TV- WWW’13 • Assume finite/limited storage (32GB) • Sensitivity analysis because calculating energy per stream is difficult • We use model by Baliga et al (2009) Oracle saves: • Up to 97% of traffic • Up to 74% of energy • Savings relatively insensitive to choice of energy model parameters • SCORE saves ~40-60% of savings achieved by oracle • Green optimisation saves 40% more energy at expense of 5% more traffic Performance evaluation Compare SCORE relative to Oracle knowing future requests

Understanding and decreasing the Network Footprint of Catch-up TV- WWW’13 • Indiscriminately recording top n shows can lead to negative energy savings! • Personalised approach necessary, despite popularity of “prime time” content Not all of these savings come from predicting popular content

Summary N. Sastry  Characterising on-demand content consumption via 6 million users of BBC iPlayer  On-demand spreads load over time  Users have strong preferences over genre/duration/serials  If broadcast is efficient, we should find ways to use it!  SCORE: personalised content offloading engine for TV  Ideal future aware version saves 97% traffic, 74% energy  Our impl gets 40-60% of ideal, with very simple measures http://www.inf.kcl.ac.uk/staff/nrs