What Is Latency? A Deep Dive

I’ll admit that I love a great streaming hit. “From Only Murders in the Building” on Hulu to “The Perfect Couple” on Netflix, I’m always seated for a smash series to the point of having to unpeel myself from my couch. I blame it on having access to too many streaming services.

What truly makes these viewing experiences enjoyable is just that: the experience. Of course, the story and the cast are why I click. Yet, without a performant network, these TV shows would be rendered unwatchable.

Latency is the time delay, usually measured in milliseconds, between when a user takes an action in a system and when they get a response.

High latency is something no one wants to run into while browsing the web or using a network. Latency is a key metric for measuring the performance of a network. So you can see why high doesn’t mean good.

Latency is complex and can have many moving parts that can influence it. Below, I’ll dive deeper into what latency means, how to measure it, and how you can reduce it so you can improve your user experience for streaming and beyond.

Anyone can feel the effects of latency, whether you’re playing a game online or video chatting with a friend. The internet has become an integral part of our lives. Understanding latency can help you identify why certain networks or applications aren’t performing to your expectations. So let’s see why these delays happen.

Factors Contributing to Latency:

• Distance and propagation delay: A propagation delay is the time it takes for the last bit of data to reach its destination. This is where distance comes into play. The further you are from a server, the more time it takes for data to travel through the transmission medium, increasing the latency.
• Processing delay: When data is sent, it has to be processed by the receiver. The time it takes for a recipient’s device to process and decode that data is called a processing delay.
• Queuing delay: Once a data packet is received, it has to wait in a queue, also known as a buffer, before it can be processed by the recipient’s device. The amount of time it stays in that queue is called a queuing delay.
• Transmission delay: A transmission delay is the time taken to transmit data from the host server to the transmission medium. The medium is what’s used to send a data packet to the destination.

Delays can be caused by various factors, which we have seen above. A delay can happen due to a combination of the above factors, or it could be just one culprit affecting your user experience.

While doing research on latency, I found that latency can affect multiple layers of a system. Those layers include the network, application, storage, and communication protocols. I’ll explore in more detail how each of these layers is affected by latency below.

Network Latency

Network latency refers to the time taken for data to travel across a network. A network is a system of devices that is connected and can communicate with each other. A few examples of networks include your home WiFi, the internet, and mobile phone networks.

Growing up, I had to share a house and WiFi network with six other people. My family would run into latency issues, as we all used the WiFi at the same time, whether that was for streaming, work, or gaming. With a brother who loves to game, it was a frequent problem. So network latency was more like another relative; always around, if you get my drift.

But my home WiFi example also displays one of the main causes of network latency: congestion. Just like a highway backed by too many cars, too many devices on a network can cause high latency and slow performance.

If your network has too little bandwidth, it won’t be able to handle all the data or traffic coming its way.

Bandwidth is a network’s capacity to transmit data in a given amount of time.

The lower your bandwidth, the less data can transfer over a network connection.

Application Latency

We interact with applications, whether web- or mobile-based, all the time. Certain apps have become staples in my daily routine. From Apple Music to Notes, I use software applications to power through my day.

That also means my day is not immune to latency. Performance issues due to latency can occur at any moment with software applications since they all rely on data transmission to operate.

For instance, I hop on a Zoom call for a work meeting. If the application is experiencing network issues, I might encounter buffering of some type before I get on the call. The delay between when I clicked on the “join meeting” button and my screen fully loading is an example of application latency.

While using this video conferencing example, other forms of application latency, in this case, can include delays between audio and video and screen freezing.

Gaming is another industry that has a high emphasis on latency.

Low latency is integral to the performance of gaming applications and experiences.

Having played video games for years, lag is the bane of my competitive existence. They can turn a rather enjoyable gaming experience into a complete drag. Imagine playing Mario Kart and pressing the brake button, but your car brakes too late due to high latency, and you fly off the road. Now, that’s frustrating.

While application latency can occur due to network issues, this type of latency is more specific to the time delay between when a user takes an action and when the application completes the desired action.

Storage Latency

Storage latency is the time a storage device takes to send and complete a read or write request. So storage latency is basically the time delay between when you request data and when that data is made available to you.

Over the years, storage latency has drastically decreased as innovations emerged. For example, standard SSDs are much faster than their predecessor HDDs and can handle more traffic.

These innovations have made data retrieval much smoother and more seamless.

Lower latency occurs when requests traverse storage systems with more efficiency and speed.

SSDs conduct read/write operations at higher rates because they use flash memory and have no moving parts, while HDDs have limited speed due to their rotating, mechanical platters.

Latency in Different Communication Protocols

The Transmission Control Protocol (TCP) and User Datagram Protocol (UDP) are parts of the Internet Protocol (IP) suite. Every network has protocols, and the TCP and UDP are the most widely used communication protocols of the IP network.

Both protocols are used to facilitate communication between two points. However, the Transmission Control Protocol is more reliable than the User Datagram Protocol. Let me tell you why.

TCP ensures that the data being transmitted will be accurately reconstructed at its destination. This approach is needed to compensate for the inherent unreliability of the Internet Protocol.

On the other hand, while UDP can transfer data to more than one source, it doesn’t ensure the accurate reconstruction of that data once it reaches its destination, making it less reliable.

The upside with UDP, however, is that it has lower latency than TCP. Since TCP has more functionality, those various functions add latency to the data transmissions.

Suppose you live in Los Angeles and you’re using a server located in Tampa, Florida. That means you’re about 2,500 miles away from your server. In this case, distance can cause a major spike in latency. A person 100 miles away from the same server may have 5 milliseconds of latency, while you may have 40 milliseconds.

Understanding how latency is measured can help you when choosing servers or solutions. Below, we’ll look at measurement tools and how to interpret latency measurements.

Tools and Techniques for Measurement

Latency can be difficult to calculate on your own, with so many variables to consider. Hey, we’re not all mathematicians. Good thing there are techniques and solutions that make latency easy to monitor and measure.

Here are a few options you can consider for measuring latency:

• Ping is a command people can use to test the reachability of a device on a network. It uses the Internet Control Message Protocol to measure how long it takes for a message to reach a targeted host and return. It can also tell you if any bits of the data packet were lost and if the remote host is active or not.
• Traceroute is a command line interface tool that determines the path a packet takes from the source to its destination and the latency of network hops. It can help you find points of failure within the router system since it sends a packet through each router involved in a data transmission to see whether it can transfer the data.
• Network performance monitoring tools provide users with information about bandwidth use and traffic patterns. Analyzing these metrics can help you determine the causes and effects of your network latency.

Interpreting Latency Measurements

No connection is truly instantaneous. Fun fact: Data can only travel as fast as the speed of light in a vacuum.

But there are acceptable latency ranges. These ranges are based on perceived latency. Although latency is constantly present, depending on its severity, we often don’t consider or even notice it. When the milliseconds start to turn into seconds, that is when our brains begin to register the dip in speed.

Of course, different applications have different standards and user expectations. For example, I might be able to wait a few seconds for a Netflix show to load, but in gaming, those few seconds would ruin the entire experience.

I’ll list a few examples of acceptable latency ranges across industries below:

• Gaming: An acceptable latency range for gaming is 40 to 60 milliseconds (ms). Speeds over 100 ms will be noticeable to a gamer and considered lagging.
• Network: Anything under 100 ms is acceptable, but for optimal performance, 30 to 40 ms is best.
• Video conferencing: Anything at 150 ms or less is acceptable for an effective connection. But 100 ms or less is the best for real-time conferencing.

Low latency usually falls between the range of 20 ms and 100 ms for most online activities and network performance. If you experience latency at 100 ms or less, you don’t have to worry about a bad user experience.

You can liken sitting in traffic to experiencing latency. For instance, you could be two miles away from your destination — a trip that would usually take five minutes. But the long line ahead of you will cause you to reach your destination in 20 minutes.

High latency is similar. Whether due to congestion or distance, latency can make simple actions take much longer to come to fruition.

In Networking and Communication

Latency can hamper many things, including a business’s operations. Most importantly, latency impacts user experience and satisfaction, which can be detrimental to a business.

If you have an online store, you should always strive for low latency because it is integral to the user experience you’re building. So finding a host with low latency would help. But this is not only true for eCommerce.

Fixing latency issues can help all kinds of businesses, from gaming to VoIP solution providers. Low latency is especially essential for real-time applications.

When people use video conferencing or a game server, they expect the video and audio to match and the actions they take to happen in real time.

User expectations are only getting higher as performance continues to improve. So low latency will be the way of the future.

In Cloud Computing and Services

As we have seen in recent years, everything is going the cloud-based route. Cloud-based applications have entered every arena, from social media and entertainment to communications and eCommerce. In fact, most of our favorite apps are cloud-based, including Facebook, Zoom, and Netflix.

The cloud allows you to access applications from anywhere, which is one of the reasons why it’s so popular. But this also makes latency so much more important.

Although scalable, the cloud handles high-performing workloads, such as AI and real-time streaming, and needs auxiliary solutions to help it manage performance and increase responsiveness for end users.

Businesses that need to reduce latency in the cloud can start by choosing the right cloud solution.

Find a cloud region that is closest to your target audience to minimize distance delays. You can also employ CDNs and load balancers to better accommodate traffic and even distribute incoming requests.

In IoT and Edge Computing

The role of latency in the IoT industry is similar to that of its role in the cloud. Like the cloud, responsiveness is key for user experience and the performance of IoT devices.

One could argue low latency is even more crucial for the IoT industry. Let me explain why.

The Internet of Things solutions include autonomous vehicles, industrial automation, and healthcare monitoring. All of these things rely on quick responses to sensor data to function effectively. They also require real-time data exchange and immediate decision-making.

With all this real-time data processing, the IoT industry needs low latency to ensure minimal delays, fast response rates, and higher operational efficiency.

Also, most IoT applications operate on limited battery power. Higher network latency can cause increased power consumption, draining battery life.

Businesses can look to edge computing to reduce latency across their IoT networks. The edge brings network resources closer to the user, allowing them to process data at or near the generation point for improved latency and response times.

Latency is unavoidable, and I’m not just saying that because I’ve experienced it way too many times. It’s actually a fact. Data transmission has its limitations, so you will never have an instantaneous connection or zero latency.

But a business can optimize performance so much so that the loading becomes unnoticeable to us as users. I’ll share with you some strategies for reducing latency in your network, applications, and hardware solutions.

Network Optimization Techniques

One of the most widely used solutions for latency is a Content Delivery Network (CDN). CDNs use a network of geographically distributed servers to cache static and dynamic content in a location closest to the end user.

CDNs are excellent tools that allow businesses to optimize content delivery, improve loading times, and better manage traffic control. In addition to CDNs, businesses can invest in bandwidth management and traffic shaping.

Teams can increase their bandwidth to reduce network congestion and increase speeds. They can also employ the technique of traffic shaping, which can help them control the flow of data into and out of their network.

Traffic shaping allows you to prioritize mission-critical traffic to flow at optimal levels. This way, you can ensure higher performance for your most important applications.

Application-Level Optimization

At the application level, you can use various techniques to reduce latency. The more efficient your code, the more likely your application will perform better.

Here are some steps you can take to optimize your application’s code for performance:

• Keep your code as simple as possible
• Cache frequently used data
• Use the right data types and efficient algorithms
• Eliminate dead code
• Avoid unnecessary I/O operations

Besides code optimization, you can also leverage asynchronous programming and load balancing to decrease latency. Asynchronous programming is a coding technique that enables a program to essentially multitask.

This type of programming enhances the user experience by reducing wait times between functions and minimizing downtime.

Load balancers are nifty solutions that can help you drive performance without much work.

Hardware Solutions

Of course, you can also upgrade your hardware to solve your latency issues. This may be a more expensive solution, but it is effective nonetheless.

Your hardware is where everything you do flows from, so it is important to have that solid foundation. You can replace certain parts of your network equipment that are either outdated or underperforming with faster or more efficient solutions.

Consider faster storage solutions as well. You can be surprised by how much a switch between SSDs can increase performance and accelerate speeds. SSDs can make any operation quicker, from loading projects to launching apps. An NVMe SSD is a great starting point if you don’t have them already.

Latency management is essential for so many industries and applications. I’m sure anyone who owns a digital property would want it to work faster, whether for a website or social media app.

High performance has become commonplace, and we’ll only expect more of it as technology continues to advance. Below, I’ll explore a few real-world applications where low latency is not only desired but required.

Gaming

Online gaming has evolved. It has become more immersive, popular, and accessible. As more people flock to stream or engage in multiplayer games, they have come to demand and expect more from their internet services. This is where low latency comes into play.

Low latency is crucial for ensuring user satisfaction and enhancing the gaming experience. It allows users to play together in real time and reduces the delays between a user’s action and the gaming application’s response.

Financial Trading

For financial traders, speed is the name of the game. Due to high competition, traders have to be ahead in everything they do, from executing trades to reacting to market changes.

Since financial trading happens in real time, low latency is key for the industry. Trading firms constantly process ever-changing, real-time market data to get the best prices and capitalize on fleeting opportunities. But high latency can cause them to miss out on those opportunities and profitability due to slow performance.

Firms can use low-latency trading strategies to increase their competitive advantage. Optimizing their network infrastructure and using monitoring tools are two ways traders can ensure lower latency and quicker response times.

Streaming Services

If you’ve ever used a streaming service, you know how important low latency is to the user experience. Low latency ensures we get quick loading times and real-time exchange, whether gaming or TV shows.

Streaming platforms can use various strategies and techniques to minimize latency for their audiences. A few strategies include employing a CDN, leveraging low-latency communication protocols, and using adaptive bitrate streaming to adjust video quality.

There are also other preventative techniques providers can take, such as conducting regular network assessments to identify and mitigate bottlenecks.

Well beyond my streaming needs, low latency has the power to shift performance and output for the entire world. Latency is an important factor in many industries across the board, from industrial manufacturing and healthcare to gaming and financial services.

As we move deeper into the digital world, low-latency strategies will play a key role in ensuring performance and satisfaction. Hopefully, the rise of edge computing will take us to the next level and, one day, do away with lag forever (fingers crossed!).