Pro Tips

Metered vs. Unmetered Bandwidth: Which is Best for a Bare Metal Server? 

Rackdog Team

bare metal server with unmetered bandwidth

Metered and unmetered bandwidth are two different ways infrastructure providers structure the cost and usage terms for data transfer on their servers.

For teams evaluating bare metal servers, the difference matters. Your bandwidth model can affect monthly costs, billing predictability as you scale, and how closely your team needs to monitor data transfer over time.

In this guide, we’ll break down how metered and unmetered bandwidth work, compare the advantages and trade-offs of each model, and help you determine which option is better suited for your bare metal workloads.

What is bandwidth? 

In infrastructure, bandwidth refers to the rate at which data can be transmitted over a network connection. Higher bandwidth means more data can be transferred in a given time period, allowing the server to support more traffic or move larger amounts of data faster.

In the context of bare metal servers, bandwidth describes how much data the server’s network port can send or receive per second. 

A server’s theoretical bandwidth capacity is determined primarily by its port speed. For example, a server with a 1 Gbps port can transmit up to 1 gigabit of data per second, or about 450 gigabytes over the course of an hour if the connection is fully utilized the entire time.

However, the actual amount of data a server can move may be lower than its theoretical maximum. Network conditions, the application running on the server, and provider policies can all affect whether the server is able to make full use of its available port speed.

What is metered bandwidth?

Metered bandwidth means a provider tracks the amount of data a server transfers during a given billing period.

With a metered bandwidth plan, a server typically comes with both an assigned port speed and a monthly data transfer allowance. The customer can transfer data at the assigned port speed and up to their allowance without additional fees. If usage exceeds the allowance, the provider typically handles it in one of two ways, depending on the plan terms.

1. Charge overages

Above the included allowance, additional data transfer is billed as overage. These rates are usually advertised or agreed to before the server is leased. Once the billing period ends, the usage meter resets.

For many bare metal providers, metered bandwidth allowances are based on outbound transfer, meaning overage fees apply to egress while inbound transfer is not counted. This is similar to many cloud pricing models, including hyperscaler platforms like AWS, Microsoft Azure, and Google Cloud Platform. 

2. Throttle the port

Alternatively, once the customer reaches the included data transfer allowance, the provider may reduce the server’s available port speed for the remainder of the billing period. 

This limits how much additional data the server can send or receive, helping the provider control network usage without charging overage fees.

What is unmetered bandwidth?

Unmetered bandwidth means there is no fixed data transfer allowance for the billing period. Instead of tracking usage against a monthly cap, the customer can transfer as much data as the assigned port speed allows without overage fees or throttling for exceeding a set allowance.

Unmetered bandwidth is still limited by port speed. A server with a 1 Gbps unmetered port can transfer data continuously, but it cannot exceed the rate supported by that port.

Providers may also apply fair-use or acceptable-use policies to protect network performance and ensure reliable service across customers. For that reason, it’s still worth reviewing the terms behind any unmetered plan to determine potential limits.

Metered vs. unmetered bandwidth compared

The difference between metered and unmetered bandwidth matters for teams choosing a bare metal server because each model affects the cost, predictability, and scalability of the deployment.

At a high level, the difference comes down to how much certainty you want around usage and billing.

Here's a quick comparison of metered vs. unmetered bandwidth:

Factor

Metered bandwidth

Unmetered bandwidth

How it works

Includes a set data transfer allowance for the billing period

Allows data transfer up to the assigned port speed without a fixed monthly allowance

Cost model

Often lower base cost, but overage fees or throttling may apply if usage exceeds the allowance

Often higher base cost, but usage does not trigger overage fees

Billing predictability

Less predictable if traffic varies or spikes unexpectedly

More predictable because monthly cost is not tied to total transfer volume

Operational burden

Requires closer monitoring of data transfer usage

Easier to plan around for high-volume or variable workloads

Best fit

Low-variance, lower-volume, or tightly controlled workloads

High-traffic, data-heavy, or variable workloads where predictable costs matter


Metered bandwidth can be a good fit when traffic is predictable and you want to avoid paying for unused transfer capacity. It offers a lower starting cost, but requires more attention to usage patterns and overage risk.

Unmetered bandwidth is usually better for workloads that move large amounts of data or experience unpredictable traffic spikes. The cost of the unmetered port is built into the server price, whether you use the full available capacity or not, but in turn you get simpler planning and more predictable monthly infrastructure costs.

The right choice depends on how much data your workload transfers, how consistent that usage is, and how important billing predictability is to your team.

Costs of metered vs. unmetered bandwidth compared

So which is cheaper, metered or unmetered bandwidth? 

The answer depends on the volume of data your workload transfers, how predictable that usage is, and which provider you choose.

Metered bandwidth costs

With metered bandwidth plans, there are three main factors that can impact bandwidth-related costs:

Port speed: Most providers scale server pricing with the speed of the port. A server equipped with a 1 Gbps port will typically cost less than a server with a 10 Gbps port. Providers often offer higher port speeds as an optional add-on. 

Data transfer allowance: This is the volume of data the server can transfer during a billing period before overage fees or throttling apply. Some providers offer multiple data transfer tiers, with higher prices for larger allowances. Others offer only a standard allowance included in the cost of a server. 

Overage rates: This is the rate charged for data transfer that exceeds the included allowance. Providers typically disclose this rate before commitment as a per-GB or per-TB figure. Any volume of outbound data transfer over the allowance is metered at this rate. 

Every provider structures metered bandwidth plans differently. To understand the true cost under this model, it’s important to estimate your expected data transfer volume and model how costs could change if usage exceeds the included allowance.

Unmetered bandwidth costs

The cost of an unmetered bandwidth plan is usually simpler to calculate and more predictable. A provider may include unmetered bandwidth in the cost of the server or offer it as a flat monthly add-on. 

In this model, bandwidth cost is tied to the server package and port speed rather than the total volume of data transferred during the billing period. As with metered plans, the provider may offer a standard port speed with options to upgrade to a higher speed for an additional cost. 

A real-world example

To help illustrate how metered and unmetered bandwidth can affect a bare metal server bill, let’s compare the cost of a hypothetical workload across increasing volumes of monthly data transfer.

In this example, we’ll break down the costs of two identical servers, one with metered bandwidth and one with unmetered: 

Metered bandwidth plan

  • 28-core server

  • 128 GB RAM

  • 1 TB storage

  • 10 Gbps port

  • 20 TB monthly data transfer allowance

  • $0.05/GB overage fee

  • Server cost: $184/month

Unmetered bandwidth plan

  • 28-core server

  • 128 GB RAM

  • 1 TB storage

  • 10 Gbps port

  • Unmetered data transfer with no overage fees

  • Server cost: $343/month

If we model those costs across different levels of data transfer: 


Volume of data transfer

Metered 

Unmetered

5 TB

$184

$343

20 TB 

$184

$343

25 TB

$434

$343

50 TB

$1684

$343

100 TB

$4184

$343


As you can see, the metered bandwidth plan is lower-cost at moderate levels of data transfer. But as overage fees increase, the total monthly cost can rise well above the cost of the unmetered plan.

In this example, a single 100 TB month on the metered plan would cost more than a full year of the unmetered plan. The breakeven point in monthly cost occurs at 23.18 TB of data transfer, less than 5 TB above the metered plan’s allowance. 

This highlights the importance of data transfer predictability when it comes to choosing between metered and unmetered bandwidth. 

The economics of a metered plan depend heavily on your workload maintaining data transfer at or below the plan’s included allowance.

Which bandwidth model is right for your bare metal server?

As with most infrastructure decisions, the right choice depends on the realities of your workload, budget, and operational needs.  

Metered bandwidth plans can be the right choice when:

  • Your monthly data transfer is relatively low or predictable.

  • You have reliable visibility into how much data your workload typically moves.

  • You want a lower base server cost and are comfortable monitoring usage.

  • Your workload is internal, backend-oriented, or not especially data-heavy.

  • Occasional overage fees would still cost less than paying for unmetered bandwidth every month.

Unmetered bandwidth plans are favorable when:

  • Your workload moves large volumes of data every month.

  • Traffic is unpredictable or prone to sudden spikes.

  • You want more predictable monthly infrastructure costs.

  • You would rather avoid monitoring data transfer allowances closely.

  • Your application is bandwidth-intensive, such as streaming, file delivery, or high-traffic web services.

  • Overage fees could quickly exceed the price difference between metered and unmetered service.

The downside of choosing the wrong model should be considered as well. Generally speaking, opting for a metered plan when a workload is better suited for unmetered can have larger consequences than the reverse.

Choosing a metered bandwidth plan for a workload that frequently spikes above its allowance can lead to significant overage costs or frequent throttling. In contrast, choosing an unmetered plan for a lower-volume workload just means paying a nominally higher amount for transfer capacity that goes unused. 

How does Rackdog approach bandwidth for bare metal servers?

Rackdog is a bare metal server provider specializing in high-bandwidth, low-latency infrastructure with an emphasis on transparency and cost predictability. 

Across 12+ global data center locations, Rackdog offers dedicated servers with unmetered bandwidth included. Customers can scale traffic without worrying about data transfer allowances, egress fees, or surprise overages. 

With port speeds available from 10 Gbps up to 400 Gbps, Rackdog bare metal is a strong fit for teams moving large volumes of data and looking to keep infrastructure costs predictable as usage grows. 

Get in touch with a Rackdog infrastructure expert to discuss your requirements and explore a bare metal solution that fits your workload.

Build with us

Ready to deploy your first server in seconds?