Bandwidth vs Latency: How to Understand the Difference
By: Ernest Sampera on December 29, 2021
When it comes to network performance, one factor that consistently comes up for both consumers and businesses is speed. Everyone wants faster connections and better performance. Internet service providers are often eager to swoop in and offer higher bandwidth as a self-evident cure for every problem. After all, more bandwidth obviously translates to more speed, right?
Well...not necessarily. Network speed is impacted by a series of factors in addition to bandwidth. Throughput and latency often play just as much, if not more, of a role in performance. Here’s what everyone needs to know about bandwidth vs latency.
What Is Bandwidth?
Although people often equate bandwidth with network speed, the term doesn’t actually relate to speed directly. Bandwidth is a measure of how much data can be transferred over a communication band over a fixed period of time (usually one second). As the name suggests, it describes the “width” of the communication “band.” The earliest measurements were expressed in bits per second (bps), but modern networks have a much greater capacity. These connections are measured in megabits (Mbps) or even gigabits per second (Gbps).
Part of the confusion over network speed comes from the similarity between a bit, which is used to measure data transfer speed, and a byte, which is used to measure data storage. Bits are abbreviated with a lower case “b” while bytes are abbreviated with an upper case “B.” There are eight bits contained in every byte. A network connection with a bandwidth of 20 Mbps, then, will not be able to download a 20 MB file in one second. It will take eight seconds because the 20 megabyte file contains 160 megabits.
What Is Latency?
While bandwidth can certainly impact network speed, when most people complain about lag or buffering, they’re often talking about problems associated with latency. When a data packet is sent one location to another, it still needs to physically travel the distance over cables or some form of wireless frequency. Even with the fastest fiber-optic cables, however, data is still limited by the laws of physics and will never be able to exceed the speed of light. That means there will always be an upper limit to how fast data can move through a network even under ideal conditions.
And conditions are rarely ideal. While the bulk of the transmission may utilize fast network infrastructure, traversing the “last mile” before reaching either end of the connection often involves a hodge-podge of slower connections that reduce speed significantly.
Latency vs Bandwidth: What’s the Difference?
Network latency measures the lag between the moment a data packet is sent and the moment it is received and processed. In the early days of internet connections, latency was rarely an issue because bandwidth limitations masked how slowly data traveled through networks. The delay between a request being sent was generally much smaller than the amount of data that could move through the connections, so latency was all but invisible.
As higher bandwidth connections have greatly increased download speeds, however, latency is much more noticeable. For example, an image may only take 5 milliseconds to download, but latency may cause users to wait 100 milliseconds before they receive the first byte of data from their download request. It’s entirely possible, then, for a high bandwidth connection to actually perform slower than a lower bandwidth connection due to latency.
This is critical for organizations looking to maximize network and application performance. Sometimes, provisioning more bandwidth isn’t enough to improve speed. An optimized location strategy is often the most effective means of increasing speed and overall performance because moving assets closer to end users as part of an edge computing network means that data travels a shorter distance between points, resulting in lower latency and faster speeds.
Types of Latency
The term “latency” can apply to many different types of lags and delays in transmitting or processing data. Some examples of these types of latency include:
Disk Latency: A delay between when data is requested from a physical storage device and when that data is returned. It frequently applies to the rotational latency and seek time of physical hard drives.
RAM Latency: Also known as CAS latency, this is a measure of how many clock cycles it takes a RAM module to access specific data in one of its columns and make that data available.
CPU Latency: Measures the how many processor clocks are necessary for an instruction to have its data made available for use by another instruction.
Audio Latency: A common issue for musicians and podcasters, audio latency measures the amount of time it takes for an audio signal to be sent through an audio interface, which converts it from analog to digital, records it, and then sends it back through the system to an output.
Video Latency: Similar to audio latency, video latency measures the amount of time it takes for a frame of video to be transferred from the camera lens to the display.
For data center customers, however, the main type of latency to worry about is network latency, which measures how quickly data packets move through network connections over fiber optic cabling, copper wires, or cellular and wi-fi signals.
Types of Bandwidth
Since bandwidth measures how much data can travel through a connection at one time, there aren’t really different types of bandwidth. When organizations assess their bandwidth needs, the main decision they need to make is how much bandwidth they need to provision. Networks with heavy data processing needs, such as digital media, require a high bandwidth connection, while less data-intensive services can often get by quite easily with lower bandwidth capacity.
The big exception to this is burstable bandwidth, which is a special provisioning plan offered by many data center providers and cloud services. Burstable bandwidth works by providing an open network port that can deliver added capacity (also known as fractional commit) when the network needs it. This strategy is great for organizations that have relatively predictable data traffic that is occasionally hit by spikes in activity. Rather than scaling up to a higher bandwidth plan that will typically go underutilized, they can provision a burstable bandwidth plan that allows them to easily handle sudden increases in traffic.
Monthly Data Transfer and Throughput
Bandwidth is often confused with two other key data transfer terms: monthly data transfer and throughput. Monthly data transfer refers to the amount of data that travels through a network over the course of a month. A useful example for demonstrating this relationship is a length of water pipe. Bandwidth is analogous to the width of the actual pipe because it determines how much water can pass through it at one time. The monthly data transfer would refer to the amount of water poured through the pipe.
Throughput, on the other hand, measures how much data can be processed by a computer system in a network. If a device has a low throughput, no amount of bandwidth will be able to help it send and receive data any faster. This is the reason why some consumers become frustrated when their high bandwidth connections don’t deliver better performance. The problem isn’t with their network connection, but rather with hardware that lacks the processing power to take advantage of it.
While IoT devices can process a great deal of data locally, edge data centers can service more demanding processing needs without having to pass data on to larger hyperscale facilities located far away. This will greatly reduce latency by minimizing how far data has to travel. When implemented in conjunction with a high bandwidth network, edge computing architecture has the potential to greatly improve performance and help companies to provide much better services to their customers.
How Decreasing Latency Improves Business
Speed is essential for today’s network applications. Customers have little patience for lagging SaaS solutions and sputtering streaming services. By taking steps to reduce latency, companies can both increase the reliability of their services and bolster their brand’s reputation in the market. Given that most businesses expect to be competing primarily by delivering a better customer experience in the coming decade, reducing latency is critical to winning and retaining market share.
Benefits to Increasing Bandwidth
High bandwidth connections are absolutely essential for services that need to deliver data intensive services and media. As any video game player with a low bandwidth internet connection can attest, “The lag is real.” Given the uneven network infrastructure found across the country, having high bandwidth and burstable bandwidth plans in place where they’re available helps organizations to bolster their reliability and capacity as much as possible to protect against delays and outages.
Does Bandwidth Affect Latency?
Bandwidth may not be the primary determinant of speed for many customers, but it’s important to understand how it impacts network performance along with factors like latency and throughput. The relationship may not be direct, but their interaction has an important influence on speed. If a network is plagued by high latency connections, no amount of bandwidth is going to help it transfer data. Similarly, driving down latency with edge computing deployments may not deliver improved performance if bandwidth and throughput remain low. By working to improve all of these factors, companies can deliver better, faster services to their customers.
How the Right Data Center Provider Can Improve Both
The combination of flexible bandwidth options and edge computing strategies give colocation data centers a significant advantage when it comes to providing organizations with fast, reliable, and high volume networking services. Assets can be placed in locations that are closer to end users to minimize latency while also provisioning the right balance of bandwidth to control costs and maximize performance. As connectivity hubs with multiple direct cloud on-ramps, data centers can also use software defined networking services to create low latency multi-clouds that don’t clog up valuable bandwidth.
With multiple data centers located in key growth markets across the US, vXchnge provides outstanding colocation solutions for multiple industries. Backed by 100% SLAs and multiple redundancies, our colocation data centers are also committed to transparency with our award-winning in\siteintelligent monitoring platform. To learn more about how vXchnge can solve your organization’s latency and bandwidth challenges, talk to one of our solutions experts today.
About Ernest Sampera
Ernie Sampera is the Chief Marketing Officer at vXchnge. Ernie is responsible for product marketing, external & corporate communications and business development.