# Throughput

> Source: Courseiva IT Certification Glossary — https://courseiva.com/glossary/throughput

## Quick definition

Throughput is how much data can move through a network connection in a given amount of time. It is measured in bits per second, like megabits per second (Mbps) or gigabits per second (Gbps). Think of it as the actual speed you get when downloading a file or streaming a video, not the theoretical maximum your plan says. Higher throughput means more data can flow, so tasks like video calls or online gaming work better.

## Simple meaning

Imagine you have a water pipe connecting your house to the main water supply. The water pressure from the supplier is like the bandwidth, which is the maximum possible flow. But the actual amount of water that comes out of your tap when you turn it on is the throughput. If the pipe is old, has a kink, or is shared with many other houses, the actual flow (throughput) will be much less than the pressure (bandwidth). In networking, throughput is the real-world measure of how much data actually gets from the source to the destination successfully. It is affected by many things, like network congestion, the quality of cables, the distance data has to travel, and the performance of devices like routers and switches. For example, if you have an internet plan that promises 100 Mbps, but you only get 30 Mbps when you test it, your throughput is 30 Mbps. This is the number that matters for your actual experience. If you are trying to stream a 4K movie, you need a certain amount of throughput to avoid buffering. If your throughput is too low, the video stops and loads. Throughput is not the same as speed in the sense of how fast a single packet travels; it is about the volume of data per second. A network can have low latency (fast individual trips) but low throughput if the road is narrow. Understanding throughput helps you diagnose why your internet feels slow, even if your plan seems fast.

## Technical definition

In networking, throughput is defined as the rate of successful message delivery over a communication channel. It is a measure of how many units of information a system can process in a given amount of time, typically expressed in bits per second (bps), kilobits per second (Kbps), megabits per second (Mbps), or gigabits per second (Gbps). Throughput is distinct from bandwidth, which is the theoretical maximum capacity of a link. For example, a Gigabit Ethernet link has a bandwidth of 1 Gbps, but actual throughput will be lower due to protocol overhead, error correction, retransmission, and network congestion. Throughput is measured end-to-end, meaning it accounts for all devices, cables, and protocols in the path. In TCP/IP networks, throughput is heavily influenced by the TCP window size, which controls how much data can be sent before waiting for an acknowledgment. If the window is too small, the sender has to wait frequently, reducing throughput. The round-trip time (RTT) also plays a role; longer RTTs require larger windows to maintain high throughput. The formula often used is: Throughput = (TCP Window Size) / (Round-Trip Time). This is why on high-latency links, like satellite connections, throughput can be low even if bandwidth is high. Network devices such as routers and switches also introduce latency and packet processing delays that reduce throughput. Quality of Service (QoS) settings can prioritize certain traffic to improve throughput for specific applications. In wireless networks, throughput is affected by signal strength, interference, and the number of users sharing the access point. Standards like 802.11ac or 802.11ax advertise high theoretical data rates, but actual throughput is often 50-70% lower due to protocol overhead (headers, acknowledgments, guard intervals). When troubleshooting, IT professionals use tools like iPerf to measure throughput between two hosts, and network performance monitors to track utilization over time. Throughput is a critical metric for service level agreements (SLAs) where providers guarantee a minimum throughput. It is also key in capacity planning, where you must ensure that your network can handle peak demand without dropping throughput below acceptable levels. In data center environments, throughput is essential for storage area networks (SANs) and backup operations. A mismatch between throughput and application requirements can lead to timeouts and poor user experience.

## Real-life example

Think about a busy highway with multiple lanes. The number of lanes is like the bandwidth, which determines the theoretical maximum number of cars that could pass. But the actual number of cars that get from point A to point B each hour is the throughput. Now imagine a toll booth at the exit. Even if the highway has six lanes, if only one toll booth is open, cars have to queue up, and the number of cars getting through per hour drops significantly. That toll booth is like a slow router or a weak Wi-Fi signal. Also, if some cars get into accidents (like data packets that get corrupted and need to be resent), the traffic slows down even more. In real life, if you are driving to work, throughput is how many miles you cover per hour, but traffic jams, red lights, and construction zones all reduce your actual throughput. In IT, if you are downloading a large file, your internet plan says you have 500 Mbps bandwidth, but if your home router is old and slow, or if your neighbor is streaming video on the same connection, your actual download speed (throughput) might be only 100 Mbps. You can test this by running a speed test. The result you get is your throughput at that moment. It is the real-world measure that tells you if your network is performing well or if there is a bottleneck. Just like you would widen a road or add more toll booths to improve traffic flow, IT professionals upgrade hardware, adjust TCP settings, or use QoS to improve throughput for critical applications.

## Why it matters

Throughput matters because it directly affects the performance of every application that uses the network. For IT professionals, understanding throughput is crucial for troubleshooting slow connections, designing networks that can handle expected loads, and ensuring that users have a good experience. If employees cannot access cloud applications quickly, or if video conferencing is choppy, it often comes down to insufficient throughput. When you are setting up a network for a small business, you need to calculate the total throughput needed for all users and applications. For example, if you have 50 people doing video calls that each need 2 Mbps throughput, you need at least 100 Mbps of actual throughput, not just bandwidth. Internet service providers often advertise high speeds, but actual throughput can vary due to congestion at peak times. A network administrator must monitor throughput over time to identify trends and plan upgrades. Throughput is also a key factor in determining whether a network link is saturated, which can cause packet loss and retransmissions, further reducing throughput. Many cloud services charge based on data transfer, so you need to know your throughput to estimate costs. For home users, throughput determines how many devices can stream, game, and work simultaneously. A low throughput can cause buffering, lag, and frustration. In the context of security, low throughput can indicate a denial of service attack or a malware infection consuming bandwidth. Overall, throughput is not just a number; it is a health indicator for the network. It tells you if your network is working efficiently or if there is a problem that needs fixing.

## Why it matters in exams

Throughput is a fundamental concept tested across many IT certification exams because it is a core performance metric. In the CompTIA Network+ exam (N10-009), throughput appears in the context of network performance metrics, troubleshooting slow networks, and comparing Ethernet standards. Expect questions that ask you to differentiate between throughput and bandwidth, and to identify factors that affect throughput, such as latency, packet loss, and jitter. The exam may present a scenario where a user reports slow file transfers, and you must determine whether the issue is bandwidth, throughput, or something else. Also, wireless standards are tested with specific throughput numbers, like the maximum throughput of 802.11ac vs. 802.11n. In the CCNA exam (200-301), throughput is essential for understanding TCP performance, specifically the relationship between TCP window size, round-trip time, and throughput. You may get a question where you calculate approximate throughput given a window size and RTT. QoS configurations often aim to guarantee a minimum throughput for voice or video traffic. In the CompTIA A+ exam, throughput is tested in the context of internet connections, such as comparing cable, DSL, fiber, and satellite, and understanding what actual throughput a user can expect. For the AWS Certified Solutions Architect exam, throughput relates to EBS volumes, where you need to choose the right volume type based on IOPS and throughput requirements. The exam also tests throughput for VPC endpoints and VPN connections. Across all exams, common question types include: 1) Definition questions that ask you to select the best description of throughput. 2) Scenario-based questions where you are given a problem and must identify that low throughput is the cause. 3) Calculation questions where you determine throughput using TCP window size and RTT. 4) Comparison questions that ask you to identify which factor has the greatest impact on throughput. Knowing that throughput is actual data transfer rate, not theoretical maximum, and understanding factors like congestion, overhead, and protocol efficiency will help you answer correctly. Also, remember that in some exams, throughput is expressed in terms of goodput, which excludes protocol overhead.

## How it appears in exam questions

Throughput appears in exam questions in several common patterns. The first is direct definition: Which of the following best describes throughput? The answer will be something like the actual amount of data transferred per unit of time, not the theoretical maximum. Another common format is a scenario where a user is experiencing slow internet, and you must identify the most likely cause. For instance, a network technician runs a speed test and gets 20 Mbps on a 100 Mbps plan. The question might ask: What is the best explanation for this difference? The answer would involve factors like network congestion, router limitations, or TCP overhead. You might also see a question about wireless throughput: A company deploys 802.11ac access points. What is the expected real-world throughput? While the theoretical maximum is around 1.3 Gbps, real-world throughput is often 200-400 Mbps due to overhead and interference. Exam questions also ask about TCP throughput calculation: Given a TCP window size of 64 KB and an RTT of 40 ms, what is the approximate throughput? You would calculate: (64 * 1024 * 8) / 0.04 = about 13 Mbps. Another type involves QoS: Which QoS mechanism ensures a minimum throughput for voice traffic? The answer is shaping or policing that reserves bandwidth. There are also troubleshooting questions: A network administrator notices that throughput drops significantly during peak hours. What should be done first? Check utilization on the main link, then consider upgrading bandwidth or implementing QoS. In the CompTIA Network+ exam, you may have to read a network diagram and identify the bottleneck that is limiting throughput. For example, a 1 Gbps backbone connected to a 100 Mbps switch will limit throughput for devices on that switch. Ethical questions can also appear: A user asks why their file transfer is slow. As a technician, you measure throughput and find it matches the plan. The correct response is to explain that throughput is the actual rate and may vary. Finally, some questions combine throughput with other metrics like latency and jitter to test your understanding of overall network performance.

## Example scenario

You are working as a help desk technician for a mid-sized company. A user in the marketing department calls and says that uploading large video files to the company's cloud storage is taking much longer than usual. They used to be able to upload a 500 MB file in about 40 seconds, but now it is taking over two minutes. You suspect that throughput might be the issue. You start by asking the user to run a speed test from their workstation. The test shows a download speed of 150 Mbps, which is normal, but the upload speed is only 5 Mbps. The company's internet plan includes 200 Mbps download and 20 Mbps upload, so the upload throughput has dropped significantly. You check the network usage on the router and see that another department is running a large backup to an offsite server, consuming most of the upload bandwidth. This is causing the upload throughput for the marketing user to drop. You then explain to the user that the network is busy with other large data transfers that are using up the available throughput. You recommend that the marketing department schedule their large uploads for off-peak hours, or you could implement QoS rules to prioritize marketing traffic for uploads. After the backup finishes, the upload throughput returns to the expected level, and the user can upload files quickly again. This scenario shows that throughput is not fixed; it changes based on network load. As a technician, identifying and measuring throughput helped you pinpoint the cause of the slowdown. On the exam, you might be asked what tool you would use, or what metric to check first. The answer would be a speed test to measure throughput, and then check network utilization to see if the link is saturated.

## Common mistakes

- **Mistake:** Confusing throughput with bandwidth
  - Why it is wrong: Bandwidth is the theoretical maximum capacity of a link, while throughput is the actual data transfer rate. Mistaking the two leads to incorrect expectations about network performance.
  - Fix: Remember that bandwidth is the speed limit on a road, but throughput is how fast you actually drive, considering traffic and road conditions.
- **Mistake:** Assuming wired throughput equals advertised speed
  - Why it is wrong: Advertised internet speeds are often 'up to' values and do not account for overhead, congestion, or router limitations. Expecting full speed in all conditions leads to false diagnoses.
  - Fix: Always test throughput under different conditions and compare to expected ranges for the technology, not the theoretical maximum.
- **Mistake:** Forgetting that TCP overhead reduces throughput
  - Why it is wrong: TCP headers, acknowledgments, and retransmissions consume bandwidth. Ignoring this overhead makes you think the network is underperforming when it is normal.
  - Fix: Know that TCP throughput is typically 50-80% of the raw bandwidth due to overhead. Use tools like iPerf to measure actual TCP throughput.
- **Mistake:** Thinking that higher latency always means lower throughput
  - Why it is wrong: Latency and throughput are related but not directly proportional. You can have high latency and still achieve high throughput if the TCP window is large enough.
  - Fix: Understand the relationship: throughput = window size / RTT. High latency reduces throughput only if the window size is fixed. Modern TCP implementations can compensate.
- **Mistake:** Not considering the number of users or devices
  - Why it is wrong: Throughput is shared among all active users. A single device may show low throughput because others are using the network, not because the connection itself is bad.
  - Fix: Always check the total network utilization and count of active devices when troubleshooting low throughput on one device.

## Exam trap

{"trap":"The exam asks: A network has a bandwidth of 100 Mbps, and a user gets a throughput of 95 Mbps. What is the most likely reason for the difference?","why_learners_choose_it":"Learners often think there is a problem, like a faulty cable or interference, because the throughput is less than the bandwidth.","how_to_avoid_it":"Understand that it is normal for throughput to be slightly lower than bandwidth due to protocol overhead (headers, framing). A 5% difference is expected on a healthy network. Only a large discrepancy (like 50% or more) indicates a problem."}

## Commonly confused with

- **Throughput vs Bandwidth:** Bandwidth is the maximum data rate a link can theoretically support, while throughput is the actual rate achieved in practice. Bandwidth is like the highway speed limit, throughput is your actual speed accounting for traffic and stops. (Example: A fiber link may have 1 Gbps bandwidth, but if the server can only send at 200 Mbps, throughput is 200 Mbps.)
- **Throughput vs Latency:** Latency is the time it takes for a single packet to travel from source to destination, measured in milliseconds. Throughput is the amount of data transferred per second. Low latency does not guarantee high throughput if the pipe is narrow. (Example: A satellite connection has high latency (600 ms) but can still have decent throughput if the TCP window is large.)
- **Throughput vs Goodput:** Goodput is the application-level throughput, excluding protocol headers and retransmissions. Throughput includes some overhead, goodput is only the actual usable data. (Example: If you download a file, the goodput is the file size divided by time, while throughput includes the TCP headers and retransmissions.)

## Step-by-step breakdown

1. **Step 1: Define the communication path** — Identify all devices and links between the source and destination, including cables, switches, routers, and the internet service provider. Each component can affect throughput.
2. **Step 2: Understand the protocol overhead** — Protocols like TCP or UDP add headers to each packet. The overhead reduces the amount of data payload that can be sent per second. For example, TCP has a 20-byte header plus IP header, which consumes part of the link capacity.
3. **Step 3: Measure the round-trip time (RTT)** — RTT is the time for a packet to go to the destination and back. This matters because TCP uses acknowledgments. The longer the RTT, the longer the sender waits before sending more data, which can cap throughput.
4. **Step 4: Consider the TCP window size** — The TCP window size determines how much data can be sent before an acknowledgment is required. A small window on a high-latency link will severely limit throughput. Modern TCP uses window scaling to allow larger windows.
5. **Step 5: Factor in network congestion and errors** — Packet loss due to congestion or interference causes retransmissions, which consume bandwidth and reduce throughput. Network congestion from multiple users also forces packets to queue, introducing delay and possibly drops.
6. **Step 6: Measure actual throughput** — Use tools like iPerf or speed tests to send data and measure the transfer rate. This gives the real-world throughput, which accounts for all the factors above. Compare it to the expected bandwidth to identify if there is a problem.

## Practical mini-lesson

In practice, measuring and optimizing throughput is a daily task for network administrators. You start by identifying the baseline. For a typical office LAN, you should expect throughput to be close to the link speed, for example, 900+ Mbps on a Gigabit Ethernet connection using iPerf with TCP. But if you are testing over the internet, throughput will be lower due to ISP shaping, remote server limits, and the path complexity. When you run a speed test, understand that it measures throughput to a specific server, not to all destinations. So if a user complains about slow access to a particular cloud service, test throughput directly to that service if possible. To improve throughput, you can adjust TCP settings such as increasing the maximum window size or enabling window scaling. In Windows, you can use netsh commands to modify auto-tuning levels. On Linux, sysctl parameters like net.ipv4.tcp_window_scaling and net.core.rmem_max control buffer sizes. For wireless networks, reducing interference by changing channels, upgrading to 5 GHz, or using access points with MU-MIMO can boost throughput. Also, use QoS to prioritize critical traffic. For example, give voice traffic a higher priority to ensure stable throughput for calls, even when the link is busy. In data centers, use link aggregation (LAG) to combine multiple physical links into one logical link, increasing total throughput. Remember that throughput is not just about speed; it is about reliability and consistency. A link that achieves 90 Mbps steadily is better than one that spikes to 200 Mbps but drops to 10 Mbps frequently. When troubleshooting, always start with measuring throughput at different points. If the throughput is low at the user's device but high at the switch, the issue is local. If it is low everywhere, check the internet link. Finally, document peak throughput times and plan capacity accordingly. If average throughput is consistently above 70% of the link capacity, it is time to upgrade to avoid performance degradation.

## Memory tip

Think 'TAO' for Throughput: T is the actual Transfer rate, A is Actual not theoretical, O is Overhead reduces it.

## FAQ

**What is a good throughput for home internet?**

For general browsing and streaming, 25 Mbps per device is good. For 4K streaming, you need at least 25 Mbps per stream. For gaming, 10 Mbps is usually fine, but low latency matters more.

**Can I increase throughput by upgrading my router?**

Yes, an old or low-end router can be a bottleneck. Upgrading to a router that supports modern Wi-Fi standards (like 802.11ac or 802.11ax) and has faster processing can improve throughput.

**Why is my wireless throughput lower than wired?**

Wireless has more overhead, interference, and signal loss. Half-duplex communication, where devices take turns sending and receiving, also limits throughput compared to full-duplex wired Ethernet.

**How do I measure throughput accurately?**

Use a tool like iPerf to send a TCP stream between two endpoints and report the measured throughput. For internet tests, use a reliable speed test service that connects to a nearby server.

**Is throughput the same as download speed?**

In common usage, yes. Download speed is a measure of throughput. But technically, throughput includes both upload and download rates, and it is the actual data transfer rate, not the advertised speed.

**What does low throughput indicate?**

It can indicate network congestion, a faulty cable, outdated hardware, interference (on wireless), or a problem with the ISP. It might also be due to a misconfigured device or a malware infection consuming bandwidth.

**How does VPN affect throughput?**

VPN adds encryption overhead and can change the packet size, often reducing throughput by 20-50%. The VPN server's capacity and distance also affect throughput.

## Summary

Throughput is a critical networking metric that reflects the real-world data transfer rate, as opposed to the theoretical bandwidth. It is affected by many factors, including protocol overhead, latency, TCP window size, network congestion, and hardware quality. Understanding throughput helps IT professionals troubleshoot slow connections, design efficient networks, and manage capacity. In certification exams, throughput appears in definition, scenario, and calculation questions across multiple exams like CompTIA Network+, CCNA, and AWS. The key takeaway is that throughput is what users actually experience, and it is usually lower than the advertised speed. By learning to measure and interpret throughput, you can identify bottlenecks and optimize network performance. Remember that throughput is not static; it changes with network conditions, and monitoring it over time gives valuable insights. Whether you are a beginner or preparing for an advanced certification, mastering throughput will help you in real-world IT work and in passing your exams.

---

Practice questions and the full interactive page: https://courseiva.com/glossary/throughput