What is Network Latency and How Does it Impact Data Transmission on an Internet / WAN Circuit?

Some sources say the average person makes a brain-numbing 35,000 choices each day. For an organization, one of those choices is which network provider and connection is best for their needs. 

The two main criteria are speed and cost, and for most enterprises, the decision is usually a balance of these factors.

A more informed decision-making process is also going to evaluate solutions based on diversity and redundancy. But here at Lightyear, there’s one metric we’ve seen businesses neglect time and time again. 

Latency is the dark horse of connectivity metrics, and as we move into an increasingly cloud-dominated business environment, it’s going to become even more relevant. See here for another blog post we wrote about packet loss and jitter.

What is Latency?

Latency is the time it takes for data to pass from one end point (individual device or user) on a network to another. 

Modern network technology is a marvelous thing, but sooner or later, it always comes up against the immutable laws of physics. Even fiber-optic networks are limited by the speed of light. If your endpoints are at opposite ends of the Earth, it’s physically impossible to transfer data between them as quickly as you would two endpoints in the same building.

This basic fact is one of the main reasons no-one’s going to be living in Mark Zuckerberg’s metaverse just yet (if ever).

It’s not like you can just add bandwidth, either. Here’s an analogy to explain. Imagine a delivery truck. If it has a top speed of 60 mph, and the next drop-off is 60 miles away, it doesn’t really matter if it’s traveling on a two-lane road or a 10-lane interstate – that truck’s going to take an hour to get to where it’s going.

Other factors can impact latency, too – things like network congestion, network complexity, and file-transfer method can all slow down your data.

What Kind of Applications Struggle with Latency?

Depending on the purpose and the types of data you’re transferring, latency can be a minor irritant or a major roadblock. Some mainstream applications can be latency sensitive.

Voice over IP phone calls and video conferencing. Human conversations in-person are a nuanced and complex exchange of verbal and non-verbal information. Subtle factors like cadence rhythm have a powerful effect on how we communicate.

When we speak over the phone or internet, we’re still habitually trying to recreate that experience. Even without latency effects, calls can leave us feeling dislocated and weary.

Add in the gaps, screen freezes, and awkward silences of a latency-prone connection, and you may be left wishing you’d done the whole thing on email instead.

Windows file transfers. Command protocols that govern the speed of packet transfer sometimes make Windows file transfers a laborious process, and any additional latency is liable to leave you feeling like you’ve been transported back to an IT environment from 2004.

Remote desktop/virtual desktop. If you’re providing IT support services via a latency-prone link, it could be a long and frustrating day.

Streaming media. Latency issues with streaming media can often be compounded by syncing issues, where audio and video data don’t arrive simultaneously.

Latency is also a major issue for many niche applications, particularly those that rely on emerging technologies. These can include the following.

• IoT controlled devices, like AI-powered cars or drones

• Stock Trading

• Online Gaming

• High-resolution image transfers (MRI, CT Scans)

• Publication Editing

There are also several applications where latency’s effects have limited impact. These include Customer Relationship Management (CRM) platforms, productivity tools like Office 365 or Google Workspace, E-commerce platforms, and billing or audit tools.

Latency and File Transfer Methods

As we mentioned above, file-transfer methods are another factor that can impact latency levels across your network. Let’s compare two popular methods, UDP and TCP/IP, to help us understand.

UDP operates to a much less rigorous standard of data transfer, with no reconciliation process for any lost data packets. Simply put, an endpoint sending data via UDP is just going to spray as much data as it can, as quickly as possible.

By contrast, TCP/IP is a more measured, sophisticated transfer method. Each packet is sent in order. When it arrives, the recipient’s endpoint responds, acknowledging the received packet. If no acknowledgement comes, the packet is retransmitted.

As you can imagine, a file transfer method like TCP/IP is going to have greater latency implications – and if you’re struggling with a particularly sluggish application, checking the file- transfer method is a good place to start troubleshooting.

Planning Your Network to Reduce Latency

The good news is many latency challenges can be avoided with proper network planning and the right topology.

There are some general rules you can use to keep latency at a minimum.

Keep your files and applications close to their users. If end users are going to need continuous access to certain applications and files, make sure you put these resources on a nearby server if possible.

Don’t rob Peter to pay Paul. If you have users on the west coast and the east coast, then make sure that the server location reflects that and provides the best results for everybody.

Keep it simple. View each additional network hand-off, protocol, or network device as a potential source of latency, and remove them wherever possible. 

Virtual Private Networks (VPN) are a common culprit – each VPN adds an average of 100ms of latency to the existing constants in the network, so if sites are connected by VPNs, and then you implement additional VPNs for each PC to connect to a server, it quickly mounts up.

Optimizing your budget and adding capacity systematically for maximum application efficiency is vital to network planning. Too many times, we’ve seen businesses adding capacity and spending their money on bandwidth when other, simpler changes (like endpoints too far apart) could provide an easier (and cheaper) solution.   The Lightyear Telecom Procurement Platform (and the team that run it) would be happy to review your network topology and assist with changes and procurement opportunities.