What is latency?
Latency is a key indicator of the quality of any business broadband connection and is not necessarily linked to broadband speed.
You can have expensive ultrafast broadband that downloads files in milliseconds but lags during video calls due to high latency.
This article explains latency, what causes it and how to reduce it, and gives latency figures for typical UK broadband connections and typical business cases.
What is latency?
Latency is the time delay between a user’s action and the response from the server or system. For example, if a user says “hello” on a video call, latency is the time it takes for the sound data to travel across the internet and onto the receivers’ headsets.
Latency is measured in milliseconds (ms) and can impact the performance of various online activities, especially those requiring real-time interaction, such as video conferencing, VoIP, online trading and payments.
What causes latency?
Latency can be caused by anything that slows the flow of data across the Internet, from physical blockers in broadband networks (cables, routers, modems) to programmed limiters in the network load balancing software handling the data transfer.
Here is a comprehensive list of factors that cause latency:
Factor causing Latency | Description |
---|---|
Distance | The physical distance data must travel between the sender and receiver. |
Network Congestion | High traffic on the network can slow down data transmission. |
Routing | The number of hops data must take through various routers and switches. |
Bandwidth (Speed) | Limited bandwidth can increase the time it takes for data to be transmitted. |
Hardware | The performance of network hardware (e.g., routers, modems) can affect latency. |
Processing Delays | Time taken by servers to process and respond to requests. |
Protocol (Software) Overhead | Additional time taken due to data encapsulation and handling by various protocols. |
Throttling | Programmed reduction in bandwidth by broadband providers for traffic management. |
Ultimately, broadband latency depends on how fast data can travel the required network distance. It’s the digital equivalent of an Amazon order travelling across the supply chain and into your mailbox.
Any traffic on the route (traffic on the motorway, delayed flights or ships) or issues in supply chain management (a bug in the system or human error when writing the destination address) will affect the “latency” of the order.
How is latency measured?
There are various ways of measuring latency because it varies over time and depends on the application used (see the specific factors affecting latency here). In any case, there are two approaches for measuring latency:
Generic broadband speed test
If you’re looking for a quick referential value for your latency at any given time, it is best to measure it off the cuff with a generic, one-click business broadband speed test. This test will provide you with a “ping” and “jitter” value to represent your latency:
- The “ping” value represents the round-trip time (RTT) in milliseconds for a data packet sent from your device to travel to a server and back.
- The “jitter” is the variation in packet arrival times, so it measures how latency changes during the short test.
This gives any employee or manager a generic representation of your network’s behaviour but lacks the depth required to form a coherent strategy for improving your business’s broadband.
Advanced latency monitoring
Businesses with complex or critical networks require sophisticated, longer-term monitoring to measure latency and what is causing any high latencies. This is outside the scope of this introductory article, but here are some aspects of this type of latency monitoring to consider:
Method | Description | Usage |
---|---|---|
Traceroute Command | Tracks the path packets take to reach a destination, measuring latency at each hop. | Detailed path analysis |
Network Monitoring Tools | Software like Wireshark, SolarWinds, and PRTG to monitor and analyse network traffic over time. | In-depth, long-term monitoring |
Application-Specific Tools | Built-in tools in applications like VoIP to measure and report latency. | Real-time application performance monitoring |
Synthetic Testing | Simulates user interactions and measures the latency for these simulated actions. | Predictive performance testing |
Real User Monitoring (RUM) | Measures latency based on actual user interactions with applications. | User experience analysis |
Service Level Agreement (SLA) Monitoring | Tracks and measures latency to ensure compliance with SLA terms. | SLA compliance |
For more information visit our guide to network monitoring tools.
What is low latency?
What is “low” latency for one business may not be low enough for another. The business’s latency threshold depends on what your business uses its broadband for.
For example, a financial institution conducting high-frequency trading requires consistently low latencies, under 5 milliseconds, which can be achieved through a combination of leased line broadband, broadband redundancy and SD-WAN.
On the other hand, a smaller business with non-critical digital requirements using it mostly for video conferencing will likely be satisfied with a latency of under 100 milliseconds, which can easily be achieved even by satellite broadband.
In any case, here are typical latency thresholds for typical and niche digital applications:
Low latency thresholds for typical business applications
Application | Ideal Latency | Acceptable Latency | Impact of Exceeding Thresholds |
---|---|---|---|
VoIP | < 20 ms | < 100 ms | Poor call quality, echoes, dropped calls. |
Video Conferencing | < 30 ms | < 100 ms | Lag, poor video and audio quality, disrupted communication. |
Web Browsing | < 20 ms | < 50 ms | Slow page load times, poor user experience. |
Cloud Applications | < 20 ms | < 50 ms | Slow data retrieval and processing, reduced productivity. |
Financial Transactions | < 10 ms | < 50 ms | Delays in trade execution, potential financial losses. |
E-commerce Checkout | < 20 ms | < 100 ms | Slow transaction processing, cart abandonment, lost sales. |
Streaming Services | < 30 ms | < 100 ms | Buffering, reduced video/audio quality, interrupted playback. |
IoT and Smart Devices | < 20 ms | < 50 ms | Delayed device responses, reduced automation efficiency. |
Low latency thresholds for niche business applications
Application | Ideal Latency | Acceptable Latency | Impact of Exceeding Thresholds |
---|---|---|---|
High-Frequency Trading | < 1 ms | < 5 ms | Delays in trade execution, significant financial losses. |
Autonomous Vehicles | < 1 ms | < 10 ms | Safety risks, delayed responses, potential accidents. |
Augmented Reality (AR) | < 5 ms | < 20 ms | Lag in visual overlay, reduced user experience and immersion. |
Virtual Reality (VR) | < 10 ms | < 20 ms | Motion sickness, disorientation, poor user experience. |
Telemedicine (Remote Surgery) | < 10 ms | < 20 ms | Delays in remote control, risks to patient safety. |
Industrial Automation | < 5 ms | < 20 ms | Inefficiency, potential equipment malfunction, safety hazards. |
Smart Grid Management | < 10 ms | < 50 ms | Delays in response, inefficiencies in power distribution. |
How can latency be reduced?
If your business is experiencing higher latencies than it requires, it is because:
- Outdated broadband: Your business has not upgraded its broadband capacity in a long time.
- Inadequate broadband: Your IT advisor might have chosen a broadband setup that doesn’t meet your current needs, or something in your broadband setup (hardware, software, or provider) is malfunctioning.
Therefore, the solution to reduce your latency depends on your specific issue. The first step is to measure your latency appropriately to pinpoint the problem and have a benchmark against which to compare it when fixing it.
Let’s explain each of these solutions in more detail:
1. Your broadband deal is outdated
Suppose you haven’t reviewed your broadband plan for a few years and have ignored any requests to engage with your business broadband provider. In that case, you are likely connected using an antiquated, high-latency technology that no longer fits this web3 era.
For example, many UK businesses are still signed up to old-fashioned ADSL connections that use the high-latency telephony copper wires laid out many decades ago.
While ADSL will be fully decommissioned during the “big switch off” in 2027, other increasingly redundant technologies like FTTC will continue to operate and have likely become insufficient for your actual broadband requirements over time.
Many firms fail to grasp how new technologies like smart devices, artificial intelligence (AI), virtual reality (VR), and cloud-based software and storage have massively increased bandwidth requirements, even when their headcount remains unchanged.
We recommend estimating your business broadband speed requirements and getting a deal with bandwidth surplus to your current needs. This ensures scalability for your business broadband and avoids any contract cancellation fees from having to change prematurely.
Compare business broadband deals to upgrade now.
2. Your broadband setup is inadequate or malfunctioning.
Each business has different network requirements, and there is no one-size-fits-all. Multiple network components, including hardware, software, and even your provider, may be substandard for your requirements.
Here are a few examples (non-exhaustive!) of what could be causing higher latency than what you should have:
Hardware limitations
- Routers: Old or inadequate business broadband routers may not support high-speed broadband connections, causing bottlenecks in data transmission and, therefore, latency
- Switches: Outdated or malfunctioning network switches that cannot handle gigabit speeds can slow down data transfer within your local network.
- Cabling: Using outdated or poor-quality cabling (e.g., CAT5 instead of CAT6 or higher) can unnecessarily introduce latency.
- Wireless broadband: Wireless connections have inherently higher latency because data travels through a gaseous medium, typically for longer distances. This is especially true for traditional satellite broadband (500-800 ms), LEO satellites like Starlink and OneWeb (25-40 ms) and somewhat for mobile connections like 4G and 5G (variable, 1 – 50 ms).
- Copper wires: Even solid, tethered connections can have relatively high latencies. Broadband connections relying on telephony copper wires, such as FTTC (10-20 ms) and Virgin’s cable broadband (10-20 ms), can be unsuitable for niche applications because their latencies can’t compete with full fibre business broadband connections (1-5 ms).
💡A note on 5G: The latency of 5G business broadband heavily depends on obstacles and antenna density. The latency of well-thought-out 5G private network for a business facility is comparable to full-fibre (1-5 ms)!
Software limitations
- Network configurations: Incorrect or suboptimal network configurations can lead to increased latency. For the most efficient, low-latency, and secure configurations, consider getting SD-WAN, cloud-based cybersecurity software, and zero trust network access
- VPNs: VPN connections force data packets to travel through additional nodes, causing the distance and complexity of the journey through the internet, and therefore the latency, higher.
- Outdated Firmware and Drivers: Ensure your network devices’ firmware and drivers are current. Manufacturers frequently release updates that can improve performance and reduce latency.
Provider limitations
- Performance: Providers offer different levels of service and have distinct packages. If your current one cannot offer the low latency your business needs, it might be time to consider switching business broadband providers.
- Throttling: Some providers may have more network congestion than others, and some may even throttle your bandwidth, leading to higher latency. Investigating and potentially changing to a provider with a less congested network can help.
- SLAs: Ensure your broadband provider offers service level agreements (SLAs) that guarantee low latency.
Comparing the latency of broadband technologies
Broadband technologies typically experience different latencies because of the various data transfer mediums (air, high atmosphere and space, copper wires, co-axial cables, and fibre optics).
While not the only factor, it is significant and worth summarising to give you an idea of what latency ranges you may expect:
Technology | Typical Latency Range (ms) | Medium | Notes |
---|---|---|---|
GEO Satellite (Traditional) | 500-800 | High atmosphere and space | High latency due to the long distance signals travel. |
LEO Satellite (Starlink/OneWeb) | 25-60 | Low Earth Orbit | Lower latency compared to traditional satellite services. |
SoGEA/ADSL | 24 - 28 | Copper wires | Highest latency tethered connection as it relies 100% on copper wires. |
4G/5G Broadband | 5-50 | Air | Variable latency due to weather, obstacles, antenna density and technology. |
FTTC (Fibre to the Cabinet) | 10-14 | Fibre optics and copper wires | Combination of fibre and copper introduces moderate latency. |
Virgin Cable Broadband | 12-15 | Co-axial cables and copper wires | Combination of co-axial and copper introduces moderate latency. |
Full Fibre (FTTP) | 4 - 8 | Fibre optics | Lowest latency as it uses fibre optics directly to the premises. |
Comparing the latency of different broadband providers
At Business Broadband Hub, we help businesses and households find better deals on their broadband. We’re often asked which business broadband provider has the lowest latency.
Helpfully, Ofcom produces detailed statistics on such things in their UK broadband performance report. The report finds that Openreach broadband providers outperform Virgin Media cable broadband as follows:
Average Openreach broadband provider – 8ms
Virgin Media Broadband – 15ms
The major broadband providers use the Openreach network: BT, Sky, TalkTalk, and Now.
Find the best broadband providers in your area using our easy comparison services:
Latency – FAQs
Our business broadband experts answer frequently asked questions regarding latency in business broadband in the UK:
What’s the difference between ping and latency?
A “Ping” measures the round-trip time (RTT) between your device and a target server. It is a generic way of testing latency and is available in generic broadband speed tests. See in what context it is best used here.
What’s the difference between ping and jitter?
While “Ping” is a way of testing latency, “jitter” is a way of testing changes in latency over time. High latency is problematic, and large variations in latency can present similar problems.
Usually, a high ping will be followed by a high jitter and vice versa. See in what context jitter is used here.
What is the difference between latency and speed?
Latency is somewhat a consequence of broadband speed, but other factors are at play. Take this example to illustrate the difference:
- Latency is like the travel time of a car journey between two cities, which depends on the distance and the car’s speed.
- Speed, in broadband terms, is like the width of the highway that determines how many cars can travel simultaneously.
Why is latency important in business broadband?
Latency is a key metric of any broadband connection because it determines the quality of real-time applications, such as those described in the ‘low latency’ section.
It is critical for typical business apps such as payment approvals, video conferencing, and niche apps such as virtual reality experiences and high-frequency trading.
How can users improve latency?
“Improving” latency is synonymous with “reducing” latency, as high latency (i.e., long time delays between sending and receiving data) must be avoided to have a high-quality connection.
Businesses experiencing high latency usually have inadequate broadband for their needs and should follow this latency-reducing advice.
What is deemed a good latency
“Good” latency is synonymous with “low” latency, as you want a short delay between sending and receiving data over the internet.
However, each application (such as video conferencing, smart devices or high-frequency trading) has a different latency tolerance, so what is “good” latency for one business may be “bad” for another.
See our low latency thresholds section for specific examples.