"Video streaming often outperforms DVD viewing because this newer technology promotes a shift away from older – less efficient – DVD players," Arman Shehabi of Lawrence Berkeley National Laboratory told environmentalresearchweb. "Streaming also helps move video viewing away from any personal vehicle driving. Jumping in your car to drive down to the video kiosk significantly increases the energy and carbon burden of watching a movie."

Transferring all DVD viewing in the US in 2011 to streaming services would have saved around 30 petajoules of energy and avoided 2 billion kg of carbon-dioxide emissions, Shehabi and colleagues Ben Walker and Eric Masanet from Northwestern University found. In total in that year, 192 PJ of energy was spent watching DVDs, emitting 10.4 billion kg of carbon dioxide.

The researchers used the Cloud Energy and Emission Research (CLEER) Model to compare the 2011 energy use and greenhouse-gas emissions of watching videos via internet streaming or on DVDs that were either rented at a "bricks-and-mortar" shop, rented online and received via the post, bought online or bought at a shop.

The three factors that most affected the carbon footprint of video viewing were the power demand of the end-user DVD player, the data transmission energy, and consumer travel for store DVDs.

"Some of the components in the video-viewing lifecycle, such as the data centres associated with video streaming or the manufacturing and shipping of DVDs, ended up having little influence on the total lifecycle energy," said Shehabi. "The devices used to actually view the videos require the lion's share of the energy use."

More efficient network transmission systems will be needed if video streaming is to remain the more sustainable option
Arman Shehabi

Data-centre energy use – both operational and embodied within the IT equipment – accounted for less than 1% of the total video-streaming energy use. "The nearly negligible relative energy contribution from data centres is due to the large capacity and high utilization of video streaming available from cloud-based IT equipment," the team wrote in Environmental Research Letters(ERL). The energy needed for data transmission was, however, key, the researchers found.

"The energy required to transmit video streaming – that is, the energy to build up and operate the network – accounts for a large portion of the video-streaming energy," said Shehabi. "As more complex video streaming comes into the market – and that video streaming requires greater bandwidth – network energy demand could become the dominant energy requirement. Our results are really a call to industry that more efficient network transmission systems will be needed if video streaming is to remain the more sustainable video-viewing option."

Video streaming required 7.9 MJ of primary energy and emitted 0.42 kg of carbon-dioxide equivalent per viewing hour, compared with 7.8–12.0 MJ of primary energy and 0.40–0.71 kg of carbon-dioxide equivalent per viewing hour for DVDs, depending on the viewing method.

"Streaming appears similar in efficiency to mail-delivered DVD viewing, but more efficient when compared with store-rented and store-purchased DVDs, owing to the impact associated with consumer driving," the researchers wrote in ERL.

In 2011, Netflix posted out roughly 2.2 million DVDs per day in the US, resulting in around 1.6 billion viewing hours. Store-rented DVDs saw around 3.6 billion viewing hours, meanwhile, and around 1.2 billion DVDs were purchased, which the team allocated equally between shop-bought discs and those received in the mail. In total there were roughly 17.2 billion hours of DVD viewing, around 5 hours per person per month; the estimate for internet streaming of full-length films and TV programmes was 3.2 billion hours.

"Over the past decade, there has been a lot of interest in the drastic growth of data centres and the energy use required to operate these buildings," said Shehabi. "But as internet services have become more pervasive, it's become apparent that the energy impact of services could extend beyond the direct energy use from data centres themselves. We built a system-wide model so that we could better understand the net energy impact of emerging data-centre services by comparing their energy demand with the traditional services they could replace."

Now Shehabi, Walker and Masanet are applying their model to other data-centre services to understand their lifecycle impacts and "help steer their development towards lower energy pathways".

Related links

Related stories