Network and Engineering
What is Latency?
Network and Engineering
What is Latency?
Jan 29, 2025
Comcast is rolling out a new, ultra-low lag connectivity experience for Xfinity Internet customers for apps like Apple’s FaceTime, Meta’s mixed reality headsets and What’s App, NVIDIA’s GeForce NOW cloud gaming, or Valve’s Steam cloud games platform. We spoke with Jason Livingood, VP, Technology Policy, Product and Standards, who has spearheaded Comcast’s industry-leading work on the technology responsible for delivering this ultra-low-lag experience about what it means for customers and the future of their connected experiences.
Q: This launch is an exciting development in Comcast’s quest to deliver the best connectivity experiences to customers. But can we take a step back and ask—what is latency?
A: Latency – referred to as “lag” or “delay”—is the measure of the time it takes for data to travel from a customer’s home to a server on the internet and back. So, think of how fast a game controller command in your home would take to travel up to NVIDIA’s GeForce NOW gaming servers and back.
Keeping that delay as low as possible is critical for any application where a user is interacting with a screen or other device. This can be gaming, video conferencing, web browsing, how fast Siri or Alexa replies to a voice prompt, and even video stream startup time.
A variety of factors impact latency, including how far the data—which we refer to as packets—travels (within your region vs. across the entire country), the functioning of your home Wi-Fi, user hardware and software, etc. That is why latency isn’t uniform and user experiences vary from home to home.
You really notice latency when using applications that require an interactive experience, like gaming. From Fortnite to Call of Duty to Gran Turismo, you can easily envision how even slight delays in controller inputs and on-screen changes can be the difference between winning and losing. You will certainly notice it when using Microsoft Teams, Apple FaceTime, and Zoom – you want no delays to interrupt your conversation’s back-and-forth, and no audio or video drops.
Q: Okay, so if it’s about how quickly data travels, how is latency different than speed?
A: Great question. Speed and latency are unique performance indicators of broadband quality, but they’re often confused. We have been conditioned to think that speed is “fast” on the internet. Actually, as latency gets lower and lower, you will perceive things happening faster and faster, as the round-trip times get lower. So, lower latency will actually feel “faster” than a connection with a lot of delay.
And “speed” is really a misnomer. “Speed” refers to bandwidth—the total capacity that can be transferred over a connection at one time. Higher speed—or more bandwidth—means a file download or upload can finish in less time. But at the massive speeds we provide to customers today, it’s unlikely they perceive the incremental difference between very high bandwidth levels.
In fact, the applications our customers use have shifted from simple downloads to applications that are interactive in nature, which is where latency really matters. Folks will often blame “speed” when their Internet buffers or a video freezes but those are all issues that occur when latency is very high and variable.
Q: Alright, I understand the difference between the two, but why does latency matter so much? Why isn’t speed enough?
A: So, what’s interesting about latency is that it’s really application, or use, dependent. If you’re downloading a large gaming file, latency is irrelevant because a straightforward download doesn’t demand any real-time responsiveness. Those downloads tend to happen in the background anyway—we are not often staring at a screen waiting for a download to complete.
On the other hand, when we are looking at a screen and tap or click something, we want something to happen immediately, and that depends on having low latency. Low latency is critically important for applications like video conferencing, gaming and e-learning—things involve continuous real-time input and interactivity, where consistently low latency has a huge impact on the overall customer experience.
If you’ve felt the frustration of being left frozen-faced on a Zoom call or lost a showdown in Call of Duty due to a slow keystroke command or even if you have seen a “buffering” icon when streaming —you’ve been impacted by high latency.
And latency is as important to application providers as it is to our customers, which is why we’ve been able to partner with some true industry leaders—like Apple, NVIDIA, Meta, and Valve—to bring our low-lag experience to customers.
Ultimately, latency is one of the biggest opportunities for improvement when it comes to customer experience as latency-dependent applications, like video conferencing, cloud gaming, conversational AI, and AR/VR continue to grow in usage. The benefits our customers are receiving today will only get better as our ultra-low lag connectivity experience continues to expand its impact.
Q: How is Comcast able to lower latency on its network? How does it work?
A: Our ultra-low lag connectivity experience is the product of considerable development and investment we’ve made in our network to be able to deploy what’s called Low Latency DOCISIS, a CableLabs technology standard that implements the Internet Engineering Task Force’s (IETF) Low Latency Low Loss Scalable Throughput (L4S) open standards. The L4S standards specify how network links, like your Internet gateway, can implement a new processing function for data packets that are marked latency-sensitive and as a result achieve an ultra-low lag experience for customers.
As part of these standards, app developers mark latency-sensitive traffic—such as video conferencing, gaming, and virtual reality—so that LLD can then manage the low latency flows to optimize application performance and quality for customers who use those applications. We are thrilled to have Meta, Apple, NVIDIA, and Valve as collaborators at launch following our initial trial work and are eager to work with any and every app developer that supports the new IETF standards. Application marking is entirely voluntary and available for use with no special cost, agreement, or proprietary APIs.
