Live streaming records shattered again: How soon will we reach 100m?

New OTT video live streaming records seem to be more than a yearly occurrence these days. But while the official viewing stats can be a little hard to come by, what our North American readers may be surprised to learn is that it’s not the NFL, NBA, English Premier League, UEFA Champions League, or even League of Legends gaming that leads the way in terms of the largest live streaming events to date. It’s cricket.

And this year, cricket’s live streaming records have tumbled several times. It began in May with JioCinema’s stream of the Indian Premier League (IPL) finals, which recorded 32 million ‘concurrent peak viewers.’ However, during the fall, the ICC Cricket World Cup saw that record smashed four times on the Disney+ Hotstar platform throughout the tournament. Unsurprisingly, the host nation – India – featured each time, culminating in the 59 million concurrent peak viewers that watched the India vs. Australia final on November 19. Most of these streaming viewers came from India, although the Disney+ Hotstar streaming service is also available in Indonesia, Malaysia, and Thailand, all of which have a significant Indian diaspora community. To do all of this – manage a live streaming event at such a scale and deliver quality experiences to viewers – is an outstanding achievement and milestone.

It seems inevitable that we will eclipse the 100 million concurrent streams barrier over the next few years, especially if one of the big quad-annual sporting events, such as the 2028 Olympic Games or 2026 FIFA World Cup, or perhaps even one of the upcoming Super Bowls, is delivered via one single unified streaming service.

Yet, the feel-good story of new streaming records being broken could be written by ChatGPT – just change the winner and event with a bigger number. This is why it’s significant that a number of other exciting developments are happening behind the scenes; each has challenges that need to be overcome but also holds the potential to transform the streaming experience significantly better at scale.

Tackling the issue of latency capacity

For live sports streaming, one of the biggest changes in recent years is the significant reduction in latency compared to legacy experiences. In 2018, Sports Illustrated wrote a great article explaining why a watch party for an NBA game amongst a group of friends was painful due to the different lag across various devices. Five years ago, we were in an era where OTT streaming was some 20 seconds behind the live game. Even cable TV was lagging significantly compared to the speed at which posts on the game appeared on social platforms, such as Twitter (now X). The issues stemmed from multiple factors such as encoding delays, chunk transmission delays, and added delays for advertising insertion, with the net result of live sports streamers hoping not to hear of the next goal, basket, or touchdown via the cheers of neighbors watching a live broadcast feed.

This issue of delay is not so noticeable today through the wider adoption of technologies like DASH and HLS that use smaller chunks, better players that can render out at lower quality levels, and more efficient CDNs – especially services like Open Caching that utilize the network edge as the origin server for video distribution.

The future of Multicast-Assisted Uncast Delivery (MAUD)

Another great example of how streaming delays are being tackled can be seen through the work of BT, as highlighted by its presentation at IBC2023, where a team from BT Research and Network Strategy outlined a groundbreaking project around Multicast-Assisted Unicast Delivery (MAUD). Its innovative approach uses multicast to assist with the unicast delivery of content to handle large peaks in network traffic.

The trial, performed with Qwilt’s Open Edge Cloud, deployed a MAUD test device in the home network of about 90 BT trialists, some of whom used the BT Sport app to request live BT Sport content, while others ran a robot-client on the MAUD device in the home. The ultimate goal of the trial was to ensure content preparation and client applications were unchanged, with the latter being unaware of the use of multicast, which could simply be considered a transparent means of optimizing network performance. In other words, it is one multicast stream across the network to the edge rather than millions of individual streams through the core.

This approach of delivering the video payload by multicast and the header by unicast and combining them in the Edge Proxy helps to achieve several aspects of the vision. The initial test of MAUD was a resounding success, and due to the ease of integration with the Qwilt Open Edge Cloud, BT plans to develop the MAUD system further, run larger trials, and ultimately deploy it as a network-optimization, traffic offload, cost-reduction technology.

Virtualizing CDNs

In a parallel trial, BT also worked with Qwilt earlier this year to test a live virtual CDN (vCDN) concept to help solve capacity issues. Unlike traditional CDNs, which are tightly coupled with underlying hardware, with a vCDN, software-based caches can be flexibly instated on commodity hardware at selected points, deeper in the network, to serve a particular ‘edge’ geography. This enables more flexible workload management as software-based caches can support unexpected surges in demand for different types of content, such as catch-up TV or streamed video from a major sporting event. The trial has proved how vCDN enables high levels of cache efficiency, meaning fewer network ‘hops’ are required to deliver the same amount of content.

It is worth noting that Airtel currently uses Qwilt’s Open Edge Cloud technology in India for its content delivery services, and, furthermore, Airtel was one of the multiple CDNs used to deliver the Cricket finals. MAUD and vCDN are two emerging approaches that have also integrated an innovative use of the Qwilt edge to reduce latency further and efficiently use core network capacity. Several other projects are also quietly underway with other pioneers that aim to achieve similar outcomes.

As more broadcast output becomes streamed with the growing popularity of FAST and more specialist VOD services, the aggregate benefit of building out an inherently more efficient delivery mechanism will benefit viewers and operators transporting content. So, when we reach a new record of 100 million concurrent streaming viewers, we’ll acknowledge the moment and then turn our attention to preparing for 1 billion.