EDUKATE SINGAPORE · GOLD STANDARD PRACTICAL CASE STUDY 220
100 Mbps is not 100 megabytes per second
A fictional digital class expects a 100 MB file to download in one second on a 100 Mbps connection. The speed is stated in megabits per second, while file size is in megabytes. Because a byte has eight bits, the file contains 800 megabits and the idealised time at a sustained 100 Mbps useful-data rate is eight seconds.
All class, file and measured performance details are hypothetical. The broadband cover illustration is conceptual, not an image of a real speed test.
The calculation at a glance
- File size: 100 MB = 100 million bytes = 800 million bits using decimal SI prefixes.
- Nominal connection rate: 100 Mbps = 100 million bits per second.
- Idealised lower-bound transfer time: 800 Mb ÷ 100 Mb/s = 8 seconds.
- Equivalent rate: 100 Mbps ÷ 8 = 12.5 MB/s under the same simplified unit convention.
- Real transfers can take longer, and a 100 MiB file is different in size from 100 MB.
Chapter navigation
- 1. The plan says 100 Mbps, the file says 100 MB
- 2. Capital B and small b are not decorative
- 3. Calculate the theoretical lower bound from the stated quantities
- 4. Decimal MB and binary MiB are also different
- 5. Why eight seconds is not a promise to the customer
- 6. A twelve-second observed transfer answers another question
- 7. Bandwidth, latency and throughput have different jobs
- 8. Build a correct comparison before judging a transfer
- 9. A simple exercise catches the factor-of-eight error
- 10. The lesson transfers to budgeting and service descriptions
- 11. The final decision is to correct the unit, not blame the system
1. The plan says 100 Mbps, the file says 100 MB
A fictional computer-skills class has an Internet connection advertised at up to 100 megabits per second, commonly written 100 Mbps. A teacher chooses a file described as 100 megabytes, or 100 MB, and tells students it should download in exactly one second because both quantities contain the number 100.
The plan’s speed and file’s size use different base units. Mbps is megabits per second; MB is megabytes. Under the standard definition, one byte contains eight bits. A 100 MB file therefore contains 800 megabits using decimal prefixes, before considering network and application overhead.
The idealised transfer time at a sustained 100 Mbps data rate is 800 Mb divided by 100 Mb/s, or eight seconds. Real transfers may take longer for many reasons. Neither one second nor exactly eight seconds is a defensible guarantee for an actual download.
The school, file and measured transfer times are invented. The broadband illustration is conceptual, not a photograph or benchmark of a specific telecommunications service.
2. Capital B and small b are not decorative
The US National Institute of Standards and Technology’s byte glossary defines a byte as eight bits in the relevant contemporary computing standards. The shorthand b denotes a bit and B denotes a byte. Confusing the two produces an eightfold error.
A data rate of 100 Mb/s is one hundred million bits each second under decimal SI prefixes. A file of 100 MB is one hundred million bytes. At eight bits per byte, it has eight hundred million bits. Both quantities involve ‘mega’, but their underlying units differ.
Students should write complete units before dividing. Dividing 100 MB by 100 Mb/s without converting bytes to bits leaves inconsistent units and gives a misleading number. Convert the file to bits or the rate to bytes per second, then check that the units cancel into seconds.
The Gold Standard begins with labels that survive copying into a presentation, quotation, speed-test screenshot or lesson worksheet. A capital letter can change the answer more than an apparently complicated formula.
3. Calculate the theoretical lower bound from the stated quantities
Using decimal prefixes, 100 MB equals 100,000,000 bytes. Multiply by eight to get 800,000,000 bits. A connection carrying exactly 100,000,000 bits per second of useful file data would transmit that amount in 800,000,000 ÷ 100,000,000 = 8 seconds.
This is an idealised minimum under the chosen conditions. It assumes continuous useful throughput at the full advertised numerical rate, no protocol overhead counted against it and no time for connection setup, processing, server response or other delays. Actual applications rarely meet every assumption.
The same arithmetic can be checked by converting the rate: 100 Mb/s divided by eight equals 12.5 MB/s. A 100 MB file divided by 12.5 MB/s again gives eight seconds. Two independent paths leading to the same number provide a useful dimensional-consistency test.
If the worksheet instead predicts one second, ask which factor has disappeared. Usually it is the eight bits per byte, but teachers should verify the file-size convention and stated transfer rate before deciding.
4. Decimal MB and binary MiB are also different
The NIST reference on binary prefixes distinguishes decimal megabytes from binary mebibytes. One MB is 1,000,000 bytes. One MiB is 1,048,576 bytes. Different software or documentation can use these labels explicitly or sometimes confusingly.
If the fictional file were 100 MiB rather than 100 MB, it would contain 104,857,600 bytes, or 838,860,800 bits. Dividing by 100,000,000 bits per second yields about 8.39 seconds under the same idealised assumptions.
The difference between eight seconds and 8.39 seconds is much smaller than the eightfold error from confusing bytes with bits. It still matters when precise storage and transfer estimates are being compared. Always identify whether the displayed size is decimal or binary before reporting a tight numerical result.
Students need not memorise every prefix at first. They should learn that labels carry definitions, and changing a unit without changing the number can alter the quantity being described.
5. Why eight seconds is not a promise to the customer
Broadband plans describe a rated or advertised connection characteristic under stated terms. The observed speed of a particular file transfer can depend on network congestion, Wi-Fi quality, the remote server, local device performance, protocol design and other users sharing the connection.
Header and control traffic also uses capacity. A file of 800 million payload bits may require additional transmitted bits or time for reliable transport. The exact overhead varies with protocols, packet sizes, losses and network conditions. A simple classroom division does not model all of them.
Consequently, the calculated eight seconds is an arithmetic benchmark for a fully utilised idealised 100 Mbps useful-data rate, not an entitlement to receive every 100 MB file in eight seconds. A real network diagnosis would measure relevant throughput and conditions rather than compare one observed transfer with a simplified lower bound.
The school should avoid using the model calculation to accuse a provider of poor service or to promise that a download finishes at a fixed instant.
6. A twelve-second observed transfer answers another question
Suppose the fictional 100 MB file actually takes twelve seconds to download. Its average application-level useful data rate is 800 megabits divided by twelve seconds, approximately 66.7 megabits per second, assuming the stated file size and timing are accurate.
That observation can be useful. It says the file payload arrived at an average effective rate of about 66.7 Mbps across the timed interval. It does not prove that the access link’s physical rate was 66.7 Mbps or that a provider’s service is defective. Other parts of the end-to-end path can constrain useful throughput.
An appropriately controlled test may compare wired and wireless connections, devices, servers, times of day and transferred file sizes. Its results should preserve the test conditions. A short single transfer may be influenced by startup behaviour and caching, while a longer or repeated test may provide a different practical picture.
The Gold Standard distinguishes a network service specification, theoretical calculation and observed application result. They are related but not the same quantity.
7. Bandwidth, latency and throughput have different jobs
Bandwidth or nominal data rate is commonly used to describe a maximum or rated carrying capacity. Latency describes delay involved in communication. Throughput describes data successfully transferred over an interval, and useful application throughput can be lower than the raw transmission rate.
A file transfer might begin slowly because of round-trip communication delays or remote-server processing, then send faster once its connections are established. A classroom worksheet that assumes an instant constant 100 Mbps stream does not include such dynamics.
For a user, the meaningful question may be ‘How long until the file is available?’ rather than ‘What is the advertised link rate?’ Both matter, but a useful estimate should state assumptions and acknowledge the end-to-end system.
Students should avoid treating every speed number in a settings screen as equivalent. A Wi-Fi link rate, a broadband service rate and a measured file-copy average may have different bottlenecks and measurement boundaries.
8. Build a correct comparison before judging a transfer
A reliable classroom table has columns for file size with units, assumed useful throughput with units, calculated ideal time, observed time, device and test context. This prevents a later reader from copying one number without its reference conditions.
For the invented lesson, the size is 100 MB (decimal), ideal throughput is 100 Mb/s, theoretical time is eight seconds and observed example is twelve seconds. The difference of four seconds is a real difference between the model and observation, not proof that a telecom contract was breached.
Teachers can add a second file of 50 MB and ask learners to predict four seconds under the same idealised rate. If the observed result is very different, students should investigate fixed overhead, variable throughput and conditions rather than force all measurements to fit a constant-speed model.
The evidence plan should also avoid using confidential student files as test objects. Publicly permitted, synthetic or otherwise suitable test data can demonstrate transfer timing without creating privacy risks.
9. A simple exercise catches the factor-of-eight error
Write the two quantities on separate cards: 100 Mbps and 100 MB. Have one group convert the rate to MB/s and another group convert the file size to megabits. Both groups should arrive at an eight-second idealised duration under the explicitly simplified assumptions.
A third group receives a deliberately misleading slide saying ‘100 equals 100, therefore one second’. Its assignment is to identify the unit mismatch and rewrite the caption. Another group calculates the 100 MiB example and explains why approximately 8.39 seconds is a different answer from eight seconds.
Older students can examine a twelve-second file transfer and compute a useful-data average near 66.7 Mb/s. They should list reasons why this number is not identical to the physical line rate. Younger learners can practice comparing bits and bytes without needing detailed networking protocols.
The assessment rewards clear units and justified assumptions, not merely quick arithmetic. Good technical reasoning is the ability to explain why the answer changes when a relevant definition changes.
10. The lesson transfers to budgeting and service descriptions
An organisation planning digital lessons needs to know how many pupils may download materials at once, how large the files are and which devices or connections are involved. A theoretical eight-second single-transfer benchmark cannot be multiplied into a guaranteed classroom schedule without considering shared capacity and the real system.
The school should calculate rough capacity needs with transparent units and then measure actual performance. If it compresses files or provides offline copies, it can compare those changes using useful outcomes such as successful access and reduced wait time.
Statements to parents should avoid promising ‘instant’ downloads merely because a plan number looks large. An accurate estimate can still be encouraging: teachers can say that the link’s advertised rate corresponds to up to 12.5 MB/s before accounting for conditions and overhead.
The Gold Standard connects correct technical quantities to practical accessibility and instruction. Efficient resources matter only if students can receive and use them when learning requires them.
11. The final decision is to correct the unit, not blame the system
The imaginary class should withdraw its one-second promise, preserve the measured file size and link specification, and teach the conversion of eight bits per byte. Its idealised calculation should be recorded as eight seconds for a 100 MB file at 100 Mbps of sustained useful throughput.
If the observed transfer takes twelve seconds, a separate measurement note can report approximately 66.7 Mbps of average payload throughput during that test and state what it does not establish about the service provider.
The next quality gate checks unit conventions, calculation steps, measurement conditions and whether any public claim is supported by real performance data. A correct formula with the wrong units remains a wrong operational decision.
The Gold Standard shows students that precision begins with definitions. The difference between Mb and MB is one letter on the page and an eightfold change in the quantity.
Unit and network-performance checklist
- Write b for bit and B for byte consistently.
- Specify decimal MB versus binary MiB when numerical precision matters.
- Convert quantities to compatible units before dividing.
- Label an eight-second calculation as idealised, not guaranteed.
- Measure actual file transfer in context and distinguish throughput from advertised service rate.
- Keep source files and testing data appropriate for privacy and accessibility.
Frequently asked questions
How long should 100 MB take at 100 Mbps?
In an idealised model with constant 100 Mbps of useful data, 100 MB contains 800 Mb and takes eight seconds. Real downloads can take longer.
Is Mbps the same as MB/s?
No. Mbps is megabits per second; MB/s is megabytes per second. Divide the numerical Mb/s rate by eight to express an equivalent decimal MB/s rate.
What is the difference between MB and MiB?
MB means one million bytes; MiB means 1,048,576 bytes under the binary-prefix standard.
Does a twelve-second download prove the provider underdelivered?
Not alone. It gives a useful file-transfer average under specified conditions, but server, protocol, Wi-Fi and other factors can affect observed time.
Why is the ideal result not one second?
The calculation that yields one second ignores the eight bits in each byte.
Primary technical references and connected cases
- NIST CSRC · Byte definition
- NIST · Binary Prefixes and SI Units
- The Gold Standard of Computing
- The Gold Standard of Information Technology
- Gold Standard Case 215 · Backup job versus recoverable data
- Gold Standard Case 202 · Containers versus TEU
- Gold Standard Case 219 · 70% probability versus time share
- Gold Standard · Core Requirements and Evidence
- Gold Standard · Practical Casebook
The NIST references establish unit definitions, not the performance of a particular telecom plan. The figures in the teaching scenario are stipulated and do not constitute a real service benchmark.