Tackling latency in next-gen gaming

Mathieu Duperré, CEO at Edgegap

Anyone that’s played a video game online has almost certainly experienced some kind of lag and connectivity issues. Despite huge infrastructure advances in the last few decades, latency remains a constant thorn in the side of gamers and detracts from the real-time experience that’s expected today.

Delivering a consistent experience to gamers playing on different devices with varying connection speeds – many of which are separated by thousands of miles – is a complex challenge. Massively popular online games like Roblox and Fortnite are just two of the many games which have benefited from years of investment into infrastructure in order to support millions of concurrent players. As the below chart from SuperJoost shows, multiplayer and online gaming is becoming the preferred way to play games amongst the most active gaming demographic, with all the technical challenges that this creates.

Games which can be played seamlessly across mobile, PC and console (so-called cross-play games) are also pushing the limits of what current internet infrastructure can deliver. Add in a new generation of streaming cloud gaming services like Stadia, Blacknut Games and Amazon’s Luna – plus Microsoft’s Game Pass and Sony’s revamped PlayStation Plus service, and you can see how the promise of console-quality performance over a broadband connection risks overloading networks that were never designed for this level of gaming.

So how can game companies, telcos and ISPs deliver on the performance promises being made to gamers? That’s where edge computing comes in.

Lag, latency and the Edge

When talking about latency it’s important to make it clear exactly what we mean. Latency refers to the amount of time it takes for game data to travel from one point to another. From the gamer’s perspective, it’s the delay between their command and seeing it happen in-game. How much latency a gamer experiences is dependent on the physical distance the data must cross through the multiple networks, routers and cables before it reaches its destination.

To use an extreme example, NASA’s Voyager 1 has made it about 14.5 billion miles from our planet so far, and it takes about 19 hours for its radio waves to reach us. Here on Earth, your latency is (hopefully) measured in milliseconds rather than hours; and gamers need around 30ms for the most optimal performance. Anywhere above 100ms can lead to noticeable lag and a frustrating experience.

This is where Edge computing comes in. As the name implies, Edge computing brings computation and data storage closer to the sources of data, placing it on the edge of the network where the performance gain is the greatest. As you’d expect, reducing unnecessary travel drastically speeds up the process providing an almost lag-free experience.

More players equals more chance for latency to be a problem

In the early days of gaming, local, couch play was part and parcel of the gaming experience. Today, a game where hundreds or even thousands of players are in the same session is nothing out of the ordinary, and there are Battle Royale games now, a whole genre of games where a hundred or more players are whittled down to a single winner.

The sheer scale of some online games dwarfs many of the most popular streaming services. Whilst Netflix remains the most successful streaming video site with 222 million subscribers, kids game Roblox has 230 million active accounts and Fortnite has over 350 million registered players. So if we assume these games reflect a growing trend, the demand on server networks is only going to increase, and gaming companies will have to look for more innovative solutions to continue meeting demand.

Cross-Platform

The ability for gamers on different devices and platforms to play and compete together is becoming an increasingly common feature of AAA multiplayer games like Apex Legends, Fornite and Call of Duty. EA Sports recently confirmed that FIFA 23 will be joining other heavy hitters in exploring cross-platform play. Considering the large amount of games on the market, and the various game modes for each game, studios are looking at crossplay to increase the amount of players who can play together. One of the main driver is to lower matchmaking time and prevent players from having to wait hours before opponents are ready to play with them.

From a latency perspective, different infrastructure across platforms means lag and downtime are far more likely. When it comes to cross-play, studios can’t use P2P (peer-to-peer) since console vendors don’t support direct communication (i.e. an Xbox can’t communicate directly with a playstation). On top of that, P2P may be limited by player’s home network (restrictive natting for example). That’s why studios typically use relays in a handful of centralised locations. Relays are seen as cheaper than authoritative server. They although have large flaws like making it harder for studios to prevent cheating, which is becoming more and more important with Web3 & NFT. This causes  higherlatency since traffic needs to travel longer distances between players. For example, when Apex Legends went cross-platform, players were inundated with frame rate drops, lags and glitches.

Edge computing allows studios to deploy cross-play games as close as possible to their players, significantly reducing latency. Which can negate some of the delay issues around differing platforms.

VR and the Metaverse

Despite hitting shelves in 2016, VR is only now slowly making its way into mainstream gaming. Advances in technology have gradually improved the user experience, while also bringing the price of hardware down and closer to the mass market – not to mention the metaverse bringing renewed attention to the tech. But latency issues still present a serious hurdle to wider adoption unless it’s addressed.

Latency impacts the player experience far more in VR than in traditional gaming as it completely disrupts the intended immersive experience. A 2020 research paper found latency of over 30-35ms in VR, had a significant impact on players’ enjoyment and immersion, which was far lower than acceptable margins on a controller.  But when it comes to the metaverse, achieving this might not be enough. Latency between headset and player has to be sub 5ms to prevent motion sickness.

In a recent blog, Meta’s VP, Dan Rabinovitsj, explained that cloud-based video games require a latency of around 75–150ms, while some AAA video games with high graphical demand require sub 35ms. Comparatively, Rabinovitsj suggests metaverse applications would need to reduce latency to low double or even single digits.

For better or worse, we’ve seen glimpses of what the metaverse has to offer already. Decentraland’s metaverse fashion week gave major brands like Dolce & Gabbana an opportunity to showcase virtual versions of their products. But attending journalists reported that the event was fraught with lag and glitches.

Gamers are a fickle bunch, so early adopters will simply move back to other games and platforms if they have poor initial experiences. Google’s Stadia promised to revolutionise gaming, but its fate was sealed at launch as the platform simply couldn’t compete with its competitors’ latency. Today, Google has ‘deprioritised’ the platform in favour of other projects.

If the metaverse goes to plan, it should encompass a lot more than traditional gaming experiences. But if it’s going to live up to players’ lofty expectations, akin to Ready Player One, more thought needs to be given to scalable and optimised infrastructure.

Unlocking next-gen gaming

The pace at which modern gaming is evolving is astounding, making the components discussed here work lag-free and as players expect will be a huge undertaking, and even more so when developers attempt to bring them all together in the metaverse.

The issue of latency may be less headline-grabbing than virtual fashion shows, NFTs and Mark Zuckerberg’s slightly unsettling promotional video, but the ability to seamlessly stitch all of these elements together will be critical in making the metaverse live up to expectations, and therefore, to its success.