Fact-checked by Grok 2 weeks ago

Data-rate units

Data-rate units are standardized measures used to quantify the speed of data transmission, , or in systems and , primarily expressed as bits per second (bit/s) or bytes per second (B/s), where a byte consists of eight bits. These units employ (SI) prefixes—such as kilo- (k, 10³), mega- (M, 10⁶), giga- (G, 10⁹), and tera- (T, 10¹²)—to denote decimal multiples, resulting in common designations like kilobit per second (kbit/s = 1,000 bit/s) and gigabyte per second (GB/s = 1,000,000,000 B/s). In , data-rate units distinguish between bit rates, which measure the flow of binary digits, and byte rates, which aggregate eight bits into larger data chunks for practical assessment in and storage interfaces. For instance, is typically quoted in megabits per second (Mbps) using decimal prefixes, as specified in IEEE Ethernet standards where 1 Mbps equals exactly 1,000,000 bit/s. Similarly, the National Institute of Standards and Technology (NIST) endorses decimal prefixes for bit and byte rates to maintain consistency with physical signaling rates in . A notable aspect is the historical overlap between decimal SI prefixes and binary powers of two (2¹⁰ = 1,024) inherited from early computing memory addressing, which caused confusion in data quantities but less so in rates. To resolve this, the International Electrotechnical Commission (IEC) introduced binary prefixes in 1999 through Amendment 2 to IEC 60027-2, such as kibi- (Ki, 2¹⁰), mebi- (Mi, 2²⁰), and gibi- (Gi, 2³⁰), specifically for byte multiples in storage (e.g., 1 KiB = 1,024 B) while recommending SI decimal prefixes for transfer rates to align with transmission standards, which were later incorporated into IEC 80000-13:2008. Despite this, decimal prefixes remain dominant for data rates in practice, as evidenced by IEEE 802 standards for wireless and wired networks, where rates like 100 Mbps or 1 Gbps use 10-based scaling.

Fundamentals of data rates

Definition and basic concepts

A data rate, also known as a data transfer rate, refers to the amount of transmitted over a or network in a given unit of time, typically measured in bits per second. This metric quantifies the speed at which information is moved from one point to another in and systems, serving as a fundamental measure of performance in data transmission. The base unit for data rates is the bit per second (bit/s or bps), where a bit represents the smallest unit of digital , capable of holding one of two values: 0 or 1. In data communications, bps is widely used to express the speed of modems, transmission carriers, and network interfaces, reflecting the flow of essential for encoding and decoding . The concept of data rates in bits per second originated in early , particularly with the development of modems in the mid-20th century, where transmission speeds were quantified in bits to match the binary nature of signals. It is important to distinguish this from baud rate, which measures symbols (signal changes) per second rather than actual bits transmitted, as multiple bits can be encoded per symbol depending on schemes. Mathematically, the data rate R is calculated as R = \frac{n}{t}, where n is the total number of bits transmitted and t is the time duration in seconds. This formula provides the average rate over a period, establishing the groundwork for scaling measurements in higher-speed contexts like modern networks, where understanding the bit as the ensures consistent evaluation of before applying prefixes or standards.

Bits versus bytes

A bit, or binary digit, is the smallest unit of digital information, representing a single value of either 0 or 1. In contrast, a byte is a unit of digital information consisting of exactly eight bits, providing a more practical grouping for representing larger amounts of data, such as characters in text encoding. This fixed relationship—1 byte = 8 bits—arises from historical conventions in computing, where eight bits were sufficient to encode a single character in early systems like ASCII. In the context of data rates, bit rates measure the transmission of individual binary digits per unit time, typically expressed in bits per second (bps), and are commonly used to quantify bandwidth and signaling efficiency. Byte rates, denoted in bytes per second (Bps), are employed for assessing data throughput in and contexts, where is handled in octet-sized chunks. The distinction is critical because byte rates are inherently one-eighth of equivalent bit rates due to the 8:1 ratio, meaning a 100 Mbps connection delivers at approximately 12.5 MBps under ideal conditions. Standard notation differentiates the units with lowercase "b" for bits (e.g., kbps, Mbps) and uppercase "B" for bytes (e.g., KBps, MBps), while the "per second" indicator is abbreviated as "/s" or "ps" to specify the rate. This , aligned with standards for clarity in , helps avoid errors in interpreting specifications. However, common confusions arise in and contexts, such as when capacities like hard drive sizes are advertised in bytes (e.g., 1 TB = 10^12 bytes) but perceived or labeled in ways that imply equivalents, exacerbating misunderstandings between bit-based transmission speeds and byte-based metrics. Such ambiguities often lead users to overestimate effective transfer rates by neglecting the bit-to-byte factor.

Standards for notation and prefixes

SI decimal prefixes and unit symbols

The (SI) utilizes decimal prefixes to form multiples and submultiples of units, including those applied to data-rate expressions in and . These prefixes are based on powers of 10^3, ensuring a standardized, metric-aligned scaling that differs from binary systems based on powers of 2^10. The Bureau International des Poids et Mesures (BIPM) defines these prefixes to promote uniformity in scientific and technical measurements. Relevant SI decimal prefixes for data rates include kilo- (symbol k, factor 10^3), mega- (M, 10^6), giga- (G, 10^9), tera- (T, 10^12), peta- (P, 10^15), and exa- (E, 10^18). These are affixed to the base units for data rates: bit per second (symbol bit/s or bps, where bit denotes a binary digit) and byte per second (B/s or Bps, where B represents eight bits). The prefixes multiply the base unit by the specified , such as 1 kbit/s equating to 10^3 bit/s. Guidelines from the BIPM and the (IEC) specify that SI unit symbols incorporate prefixes directly without spaces or hyphens, and the (/) indicates "per second" for rates. Examples include kbit/s for kilobit per second and MB/s for per second. This notation adheres to rules in IEC 60027-2 for letter symbols in electrical technology, particularly and , emphasizing decimal scaling for data rates to avoid ambiguity with conventions. Decimal prefixes strictly represent powers of 1000 (10^3), distinguishing them from binary multiples of (2^10) used in some contexts; this separation is critical for precise data-rate reporting in fields like networking. The 9th edition of the SI Brochure, published in 2019 by the BIPM, reaffirms these and notation rules for data rates, with a minor update in 3.02 (August 2025) primarily affecting prefixes and other non-decimal elements, remaining authoritative as of November 2025.

IEC binary prefixes

The (IEC) introduced binary prefixes to standardize the representation of powers of two in and data-related measurements, addressing the need for precision in -based systems. These prefixes are defined for multiples such as 2^{10}, 2^{20}, and higher, with distinct names and symbols to differentiate them from SI decimal prefixes. The primary binary prefixes are kibi (Ki), mebi (Mi), gibi (Gi), tebi (Ti), pebi (Pi), and exbi (Ei), corresponding to 1024, 1,048,576, 1,073,741,824, 1,099,511,627,776, 1,125,899,906,842,624, and 1,152,921,504,606,846,976, respectively. The system was extended in 2005 to include zebi (Zi, 2^{70}) and yobi (Yi, 2^{80}). In 2025, IEC 80000-13 was updated to include additional binary prefixes such as robi (Ri, 2^{90}) and quebi (Qi, 2^{100}), aligning with new SI prefixes for high-scale data applications. These prefixes were formally defined in Amendment 2 to IEC 60027-2 in , with further amendments in 2005 and incorporation into the ISO/IEC 80000-13 standard in 2008 to promote consistent usage in and . By November 2025, these standards reflect ongoing efforts to support emerging data technologies. In data-rate contexts, binary prefixes are applied to units like kibibit per second (Kibps) or mebibyte per second (MiB/s), where they denote exact binary multiples, such as 1 KiB/s equaling precisely 1024 bytes per second. This approach is particularly valuable in , memory allocation, and file system reporting, where binary addressing is inherent, ensuring unambiguous calculations. Their advantage lies in eliminating confusion with decimal-based interpretations, which are more common in general scientific metrics. Adoption of IEC binary prefixes has been gradual but increasing, especially in and precise engineering applications by 2025. However, they remain rare in networking protocols, where prevail for consistency with international standards, while becoming a in file systems for accurate capacity reporting.
PrefixSymbolValue (as power of 2)Numerical Equivalent
kibiKi2^{10}1,024
mebiMi2^{20}1,048,576
gibiGi2^{30}1,073,741,824
tebiTi2^{40}1,099,511,627,776
pebiPi2^{50}1,125,899,906,842,624
exbiEi2^{60}1,152,921,504,606,846,976

Common variations and ambiguities

In data-rate units, common variations arise from the interchangeable use of abbreviations like '' to denote both the decimal kilo- (10^3 or 1000) and the IEC kibi- (2^10 or ), leading to inconsistent interpretations across contexts. Similarly, '' is frequently applied to represent either 10^6 bytes in decimal notation or 2^20 bytes in binary notation, exacerbating confusion in specifications and measurements. These ambiguities are particularly pronounced between industries: standards typically employ , where Mbps denotes 10^6 bits per second, while and contexts favor prefixes, such as MB/s equating to 2^20 bytes per second. This divergence results in discrepancies of 7-10% when comparing equivalent nominal values—for instance, a 1 transfer rate interpreted as (approximately 1.074 × 10^9 bytes) exceeds the decimal equivalent (10^9 bytes) by about 7.37%. Prior to the formal introduction of binary prefixes in 1999 by the International Electrotechnical Commission (IEC) through Amendment 2 to IEC 60027-2, the absence of standardized distinctions fueled significant legal disputes in the early 2000s, including class-action lawsuits against hard drive manufacturers like Western Digital and Seagate for allegedly overstating capacities by using decimal prefixes while operating systems applied binary interpretations. To mitigate such issues, best practices emphasize explicitly specifying whether or notation is intended in technical documentation and interfaces. The ISO/IEC 80000-13:2008 standard, which consolidates and IEC prefixes for , recommends using unambiguous symbols and context-clear descriptions to prevent misinterpretation, such as distinguishing "" (, ) from "" (mebibyte, ). The standard was updated in 2025 to include additional binary prefixes. As of November 2025, prefixes see increased adoption in and for storage and allocation, reflecting the nature of , though continue to dominate specifications in , including standards, with research for expected to follow decimal conventions.

Decimal multiples for bits per second

Kilobit per second

The (kbps) is a of transfer rate defined as 1,000 bits per second, following the decimal system where "" denotes a factor of 10^3. This decimal scaling distinguishes it from binary prefixes and aligns with standards for metrics. The standard symbols for this are kbit/s or kb/s, though uppercase Kbps is also commonly used in technical documentation. To express a rate in base bits per second (bps), the conversion equation is: \text{Rate (bps)} = \text{kbps} \times 1000 This linear relationship facilitates straightforward calculations for and throughput. In practical applications, kbps serves as a measure for low-speed data connections, such as legacy dial-up modems that operated at up to 56 kbps over standard lines, enabling basic in the late and early . It also underpins basic audio streaming, where rates around 64–128 kbps support acceptable quality for voice transmission or compressed music playback without excessive demands. One kilobit per second equates to 125 bytes per second in decimal notation, since there are 8 bits in a byte, providing a useful cross-reference for contexts.

Megabit per second

The (Mbps) is a unit of data transfer rate defined as exactly 1,000,000 bits per second (bps). This equates to one million bits transmitted every second, employing the SI "mega-" to denote the factor of 10^6. The standard symbols for this unit are Mbit/s or Mb/s, which are widely used in technical specifications and marketing materials for . In contemporary networking, Mbps is the primary metric for measuring speeds in consumer services, including DSL and internet connections that commonly offer plans at 100 Mbps for residential use. These Mbps-scale speeds enable reliable support for bandwidth-intensive activities, such as streaming, where services like recommend a minimum of 5 Mbps to deliver content without interruptions. For example, a typical 100 Mbps home plan allows multiple simultaneous streams across devices, accommodating household demands for entertainment and . The relationship between Mbps and the fundamental bit per second unit is expressed by the equation: \text{bps} = \text{Mbps} \times 10^6 This conversion underscores the unit's role in scaling data rates for practical analysis in network engineering. As of the third quarter of 2025, the global average fixed download speed reached approximately 107 Mbps, reflecting the widespread adoption of Mbps-level in and suburban areas worldwide.

Gigabit per second

The gigabit per second (Gbps) is a unit of data transfer rate defined as exactly 1,000,000,000 bits per second, utilizing the SI decimal prefix "giga-" to denote a factor of 10^9. This decimal-based definition aligns with international telecommunications standards, distinguishing it from binary prefixes. The common symbols for this unit are Gbit/s or Gb/s, reflecting its application in measuring bandwidth for high-speed networks. To express a data rate in base bits per second (bps), the conversion equation is: bps = Gbps × 10^9. This straightforward scaling facilitates calculations in network engineering, where Gbps quantifies throughput for applications requiring substantial . In practical deployments, 1 Gbps serves as a standard for in local area networks (LANs), enabling efficient in and home environments via the IEEE 802.3ab specification for 1000BASE-T over twisted-pair cabling. Similarly, 5G mobile networks achieve peak data rates in the Gbps range, supporting ultra-high-definition streaming and with theoretical downlink speeds up to 20 Gbps under optimal conditions like millimeter-wave . For data centers, the IEEE ratified the 802.3df standard in , specifying 400 Gbps and 800 Gbps Ethernet operations to meet escalating demands for and workloads.

Terabit per second

The terabit per second (Tbps) is a unit of transfer rate equal to one (1012) bits per second, representing a decimal multiple in the (SI) for measuring high-speed digital communications. This scale is essential for quantifying capacities in next-generation networks where gigabit per second (Gbps) rates fall short for emerging demands like ultra-high-definition streaming and large-scale analytics. The symbol for terabit per second is typically Tbit/s or Tb/s, with the prefix "tera-" denoting 1012 as per SI conventions. To express a rate in bits per second (bps), the formula is: \text{bps} = \text{Tbps} \times 10^{12} This unit facilitates precise scaling in designs, such as aggregating multiple Gbps channels to reach Tbps thresholds. In research, Tbps rates enable breakthrough transmissions, with laboratory demonstrations achieving 402 Tbps over standard single-mode fibers using expanded wavelength bands, sufficient to transmit over 16 million movies simultaneously. For prototypes, wireless systems have demonstrated 1 Tbps air interface user rates at frequencies (330–500 GHz), leveraging photonics-assisted techniques to support future mobile networks with minimal and massive .

Petabit per second

The petabit per second (Pbps) is a decimal unit of data rate defined as exactly $10^{15} bits per second, representing one quadrillion bits transmitted in one second. This scale follows the (SI) prefix "peta-" for multiples of 10, applied to the base bit per second (bps). The standard symbols for the are Pbit/s or Pb/s, though Pbps is commonly used in technical literature for brevity. To express a data rate in base bits per second, the conversion equation is: \text{Rate (bps)} = \text{Pbps} \times 10^{15} This unit quantifies ultra-high-speed data transmission primarily in experimental contexts, where it serves as a benchmark for pushing the theoretical limits of communications. In fiber optics research, Pbps rates explore boundaries using advanced techniques like space-division with multi-core fibers, enabling transmission equivalent to millions of simultaneous streams. For instance, such rates approach the limit for information in standard single-mode fibers, informed by seminal work on nonlinear optical effects and . Applications of Pbps focus on frontier research in long-haul networks, where lab demonstrations highlight potential for future backbone infrastructure supporting exascale data demands from and . A notable milestone is the 2023 achievement of 22.9 Pbps over a short using a single 4-core with 38 spatial and modes, doubling prior records and equivalent to over 3 million video channels. By 2025, while capacity records remain sparse beyond this, advancements include 1.02 Pbps sustained over 1,808 with a 19-core , emphasizing practical long-distance viability through cladding diameters. These developments build on terabit per second precedents but scale dramatically for theoretical limits.

Decimal multiples for bytes per second

Kilobyte per second

The per second (kB/s) is a of data transfer rate defined as exactly 1,000 bytes per second, employing the SI decimal prefix "kilo-" for 10^3. This is commonly applied in contexts involving byte-oriented data flows, such as file transfers and storage operations. The symbol kB/s explicitly indicates to avoid confusion with kilobits per second (kbps or kbit/s). Since one byte equals eight bits, 1 kB/s corresponds to 8,000 bits per second (bps). The bit rate equivalent can be calculated using the equation: \text{Bit rate (bps)} = \text{kB/s} \times 8000 This relationship facilitates conversions between byte- and bit-based rates, with kB/s serving as the byte-focused counterpart to kbps. Prior to the , kB/s was prevalent in storage device specifications for its representation of modest transfer capabilities; for instance, the Seagate ST-506 hard drive, introduced in 1980, achieved a transfer rate of 625 using encoding. It also characterized slow file download speeds in early networking, such as dial-up modems limited to approximately 7 on a 56 kbps due to overhead. Additionally, early USB implementations like the low-speed of USB 1.1 operated at 1.5 Mbps, equivalent to 187.5 .

Megabyte per second

The per second (MB/s) is a unit of data-rate that represents a transfer speed of one million bytes per second, where one equals 10^6 bytes as defined by the (SI). This unit employs the SI prefix "mega-" for scaling, distinguishing it from binary prefixes like mebi-. The symbol for this unit is conventionally MB/s, and it is widely adopted in specifications for and peripheral interfaces to quantify bulk throughput in bytes rather than bits. In practical applications, MB/s is prevalent for describing sequential read and write performance in consumer storage devices, such as solid-state drives (SSDs) and hard disk drives (HDDs). For instance, interfaces, common in entry-level and mid-range systems, limit SSD speeds to approximately 500 /s due to the protocol's 6 Gbit/s ceiling, after accounting for overhead. Similarly, (also known as SuperSpeed USB) offers a theoretical maximum of 5 Gbit/s, equivalent to 625 /s in byte terms, enabling faster transfers compared to prior USB generations. These rates highlight MB/s as a for everyday tasks like file copying and application loading. To relate MB/s to bit-based rates, the conversion is: \text{Bit rate (bps)} = \text{MB/s} \times 8 \times 10^6 This arises from the encoding of one byte as 8 bits and the decimal scaling of the . Thus, a operating at 1 MB/s transfers data at 8 megabits per second (Mbps). Advancements in have pushed MB/s capabilities higher; by 2025, NVMe SSDs over PCIe 4.0 lanes commonly achieve sequential speeds up to 7000 MB/s, supporting demanding workloads in , , and data analytics. This evolution underscores the unit's role in scaling with improvements while maintaining decimal consistency for .

Gigabyte per second

The per second (GB/s) is a of data transfer rate that measures the amount of data processed or transmitted in one second, defined as exactly 10^9 bytes per second using the prefix. This convention aligns with international standards for and capacities, distinguishing it from prefixes. The symbol for this is GB/s. To relate GB/s to bit-based rates, the conversion equation is given by: \text{Bit rate (bps)} = \text{GB/s} \times 8 \times 10^9 This follows from the standard definition of one byte equaling eight bits, yielding 8 gigabits per second (Gbps) for each GB/s. In professional storage and computing environments, GB/s is commonly applied to high-performance interfaces such as (PCIe) 4.0 and 5.0, where configurations like an x4 PCIe 5.0 slot achieve up to 16 GB/s throughput. It also supports rapid data transfers in data centers, enabling efficient handling of large-scale arrays and interconnects for tasks like AI training and analytics. By 2025, enterprise arrays utilizing advanced NVMe SSDs and PCIe 6.0 have demonstrated aggregate speeds exceeding 100 GB/s, marking a significant in scalable performance.

Terabyte per second

The terabyte per second (TB/s) is a decimal unit of data transfer rate defined as exactly 10^{12} bytes per second. This equates to 8 \times 10^{12} bits per second, since one byte consists of eight bits. The symbol for this unit is TB/s. To convert a data rate from TB/s to bits per second, the following equation is used: \text{Bit rate (bps)} = \text{TB/s} \times 8 \times 10^{12} This conversion is essential for comparing byte-based rates with bit-based network specifications. TB/s rates are critical in supercomputer interconnects, enabling the rapid exchange of massive datasets across thousands of processors. For instance, NVIDIA's NVLink interconnect delivers up to 1.8 TB/s of bandwidth per GPU, supporting exascale computing tasks that require synchronized data movement at this scale. Similarly, the LUMI supercomputer achieves a maximum I/O bandwidth of 2 TB/s, facilitating high-throughput simulations in climate modeling and scientific research. In cloud environments, TB/s supports large-scale data handling for AI training clusters and backups, where petabyte-scale datasets must be processed or archived efficiently. By 2025, systems like Sycomp's parallel for AI platforms deliver over 1.2 TB/s of aggregate throughput, optimizing storage for workloads. Google's (TPU) pods exemplify this, with v4 configurations providing 24 TB/s of internal per pod as of 2024, enabling accelerated training of large language models through high-speed inter-chip communication.

Binary multiples for bits and bytes per second

Although standardized, prefixes for data rates are less commonly used than prefixes in and networking, where units prevail, but they are recommended for contexts involving -aligned data.

Kibibit and kibibyte per second

The kibibit per second (symbol: Kibit/s or Kibps) is a unit of data rate defined as exactly 1024 bits per second (bps), introduced to provide unambiguous measurements in contexts where powers of two are standard. This contrasts with the kilobit per second (kbps), which equals 1000 bps, ensuring precision for -aligned systems like memory and storage addressing. Similarly, the kibibyte per second (symbol: KiB/s) represents 1024 bytes per second, or equivalently 8192 bps since each byte consists of 8 bits. These units stem from the (IEC) standard IEC 60027-2, amended in 1999 to formalize binary prefixes and avoid confusion between decimal and interpretations of "kilo." In practice, conversion from these rates to base bits per second is straightforward: for kibibits, multiply by (i.e., \text{bps} = \text{Kibit/s} \times 1024); for kibibytes, multiply by 8192 (i.e., \text{bps} = \text{KiB/s} \times 8192). In applications, kibibit/s and kibibyte/s are favored for software metrics where binary precision matters, such as reporting or file transfer speeds in tools that align with hardware realities. For instance, utilities like iostat report disk rates using binary multiples labeled as kilobytes per second (kB/s), where 1 kB equals 1024 bytes (a kibibyte), to reflect exact block sizes, a practice adopted widely since the early 2000s following IEC guidelines. They also apply to calculations, where systems like DDR4 modules achieve rates on the order of tens of gibibytes per second per but can be scaled down to kibibyte/s for per-pin or low-level analyses in .

Mebibit and mebibyte per second

The mebibit per second (symbol: Mibit/s) is a unit of defined as exactly 1,048,576 bits per second (bps), derived from the mebi- prefix representing 2^{20}. This prefix was standardized by the (IEC) in 1999 to distinguish binary multiples from decimal ones in contexts. The conversion from mebibits per second to bits per second is given by the equation: \text{bps} = \text{Mibit/s} \times 2^{20} where $2^{20} = 1,048,576. The mebibyte per second (symbol: MiB/s) extends this to byte-based rates, equaling 1,048,576 bytes per second, or 8,388,608 bps since each byte comprises 8 bits. It is commonly used in software environments where data transfers align with binary addressing schemes. The conversion equation is: \text{bps} = \text{MiB/s} \times 2^{20} \times 8 This unit helps avoid ambiguity in systems that traditionally use powers of 2 for memory and storage calculations. In practical applications, Mibit/s and MiB/s appear in mid-range file transfer scenarios within software, such as torrent clients where download progress is displayed in MiB/s to reflect binary-aligned throughput—for instance, qBittorrent reports speeds using these units for user clarity in peer-to-peer transfers. Similarly, GPU memory bandwidth testing tools, like NVIDIA's bandwidthTest utility, often output results in MiB/s to match binary conventions in hardware specifications, enabling precise evaluation of data movement rates in graphics processing tasks. These units provide conceptual insight into scalable binary data flows without requiring exhaustive decimal adjustments.

Gibibit and gibibyte per second

The gibibit per second (symbol: Gibit/s or Gib/s) is a unit of data rate defined as exactly 2^{30} bits per second, equivalent to 1,073,741,824 bits per second or approximately 1.074 \times 10^9 bits per second. This unit employs the (Gi), which denotes multiplication by 2^{30}, to provide clarity in contexts where powers of two are standard. The gibibit per second is part of the binary multiple system standardized for applications to distinguish from decimal prefixes like gigabit (10^9 bits). The gibibyte per second (symbol: GiB/s) extends this to byte-based rates, defined as 2^{30} bytes per second, or 1,073,741,824 bytes per second, which equates to approximately 8.589 \times 10^9 bits per second when accounting for 8 bits per byte. To convert these units to base bits per second (bps), the relations are: $1~\text{Gibit/s} = 2^{30}~\text{bps} $1~\text{GiB/s} = 2^{30} \times 8~\text{bps} These definitions were formalized in the IEEE 1541-2021 standard to promote unambiguous usage of binary prefixes in and , revising earlier versions for consistency with international conventions. In practical applications, gibibit and gibibyte per second rates are relevant to high-end computing scenarios requiring substantial throughput. For instance, server RAM bandwidth in multi-channel DDR5 configurations can reach or exceed 100 GiB/s, enabling efficient data handling for virtualization and database operations in enterprise environments. Similarly, high-resolution video encoding, such as for 8K content, demands GiB/s-scale rates to process uncompressed streams without bottlenecks, as seen in professional workflows using hardware accelerators.

Conversions and reference tables

Bit-to-byte rate conversions

Converting between bit rates and byte rates is fundamental in data communications, as bits represent the smallest unit of while bytes group eight bits together to form addressable units for and . The core factor stems from the that one byte equals eight bits, allowing straightforward interchanges between the two. Thus, to derive the byte rate from a , divide the bit rate by 8; conversely, multiply the byte rate by 8 to obtain the bit rate. This relationship holds across all prefixes, such as megabits per second (Mbps) to megabytes per second (MB/s), provided consistent or scaling is applied. For instance, a speed of 100 Mbps equates to 12.5 /s, illustrating how the division by 8 scales down the rate when shifting from bits to bytes. Similarly, a transfer rate of 10 /s corresponds to 80 Mbps. These conversions are essential for aligning specifications between networking equipment, which often quotes in bits, and file systems, which use bytes. In practice, protocol overhead—such as headers in / or Ethernet frames—reduces the effective payload byte rate, typically by 10-20% depending on the protocol and packet size. For example, Ethernet framing and / headers can consume additional bits not available for data, lowering efficiency. The effective byte rate can be calculated as: \text{Effective B/s} = \frac{\text{bps} \times \text{[efficiency](/page/Efficiency)}}{8} where is a factor between 0 and 1 for overhead (e.g., 0.8 for 20% loss). This adjustment is crucial when evaluating real-world throughput. Such bit-to-byte mismatches frequently arise in scenarios, where users confuse bit rates with byte capacities, leading to unexpected discrepancies.

Prefix multiplier comparisons

In data-rate units, prefix multipliers distinguish between decimal-based systems, defined by the International System of Units (SI) as powers of 10, and binary-based systems, standardized by the International Electrotechnical Commission (IEC) as powers of 2 for computing contexts. Decimal prefixes like kilo (k) denote 10^3 = 1000, while the corresponding binary prefix kibi (Ki) denotes 2^10 = 1024, resulting in a ratio of 1000/1024 ≈ 0.9766, meaning one decimal kilounit is approximately 97.66% of its binary counterpart. Similarly, the mega prefix (M) is 10^6 = 1,000,000, compared to mebi (Mi) at 2^20 = 1,048,576, with a ratio of ≈0.9537. These comparisons extend to higher levels, such as giga (G) at 10^9 versus gibi (Gi) at 2^30 = 1,073,741,824 (ratio ≈0.9313), tera (T) at 10^12 versus tebi (Ti) at 2^40 = 1,099,511,627,776 (ratio ≈0.9095), and peta (P) at 10^15 versus pebi (Pi) at 2^50 = 1,125,899,906,842,624 (ratio ≈0.8882). The table below summarizes these prefix multipliers up to the peta level, highlighting the decimal value, binary value, and the decimal-to-binary ratio:
PrefixDecimal Multiplier (10^n)Binary Multiplier (2^{10n})Ratio (Decimal / Binary)
Kilo/Kibi (n=1)1,0001,0240.9766
Mega/Mebi (n=2)1,000,0001,048,5760.9537
Giga/Gibi (n=3)1,000,000,0001,073,741,8240.9313
Tera/Tebi (n=4)1,000,000,000,0001,099,511,627,7760.9095
Peta/Pebi (n=5)1,000,000,000,000,0001,125,899,906,842,6240.8882
These ratios illustrate how the discrepancy grows with higher powers: at the level, the difference is about 4.6%, increasing to roughly 7% at and over 11% at . For instance, 1 (decimal) equals 1,000 TB, whereas 1 PiB (binary) equals 1,024 , emphasizing the scaling variance in large data-rate contexts. are conventionally applied in and standards for measurements, aligning with conventions, while prefixes are preferred for and storage capacities to reflect base-2 addressing. The underlying relationship can be expressed mathematically: the is 1024^n, the is 1000^n, and their ratio is (1000/1024)^n, where n is the . As of 2025, software tools like Python's humanize facilitate context-aware conversions, automatically applying decimal or prefixes based on user specification to mitigate scaling errors in data-rate representations.

Comprehensive unit conversion table

The following table provides equivalents for common data-rate units in bits per second (bps), encompassing decimal (SI) prefixes for bits and bytes, as well as binary (IEC) prefixes up to the exa scale. These conversions adhere to the SI Brochure for decimal multiples and IEC 80000-13 for binary multiples, with no protocol overhead or encoding factors included.
Unit SymbolFull NameEquivalent in bpsNotes
bpsbit per second1SI base unit for raw data transmission rates.
B/sbyte per second8Assumes 1 byte = 8 bits; common in storage contexts.
kbpskilobit per second1,000SI decimal; kilo = 10³; used in .
kB/skilobyte per second8,000Decimal byte rate; 1 kB = 10³ bytes.
Kibpskibibit per second1,024IEC binary; kibi = 2¹⁰; preferred for .
KiB/skibibyte per second8,192Binary byte rate; 1 KiB = 2¹⁰ bytes.
Mbpsmegabit per second1,000,000SI decimal; mega = 10⁶; standard in networking (e.g., 1 Mbps = 1.25 × 10⁵ B/s).
MB/smegabyte per second8,000,000Decimal; 1 MB = 10⁶ bytes.
Mibpsmebibit per second1,048,576IEC binary; mebi = 2²⁰; ≈ 0.9537 Mbps.
MiB/smebibyte per second8,388,608Binary; 1 MiB = 2²⁰ bytes.
Gbpsgigabit per second1,000,000,000SI decimal; giga = 10⁹.
GB/sgigabyte per second8,000,000,000Decimal; 1 GB = 10⁹ bytes.
Gibpsgibibit per second1,073,741,824IEC binary; gibi = 2³⁰.
GiB/sgibibyte per second8,589,934,592Binary; 1 GiB = 2³⁰ bytes.
Tbpsterabit per second1 × 10¹²SI decimal; tera = 10¹².
TB/sterabyte per second8 × 10¹²Decimal; 1 TB = 10¹² bytes.
Tibpstebibit per second1.0995 × 10¹²IEC binary; tebi = 2⁴⁰.
TiB/stebibyte per second8.796 × 10¹²Binary; 1 TiB = 2⁴⁰ bytes.
Pbpspetabit per second1 × 10¹⁵SI decimal; peta = 10¹⁵.
PB/spetabyte per second8 × 10¹⁵Decimal; 1 PB = 10¹⁵ bytes.
Pibpspebibit per second1.1259 × 10¹⁵IEC binary; pebi = 2⁵⁰.
PiB/spebibyte per second9.007 × 10¹⁵Binary; 1 PiB = 2⁵⁰ bytes.
Ebpsexabit per second1 × 10¹⁸SI decimal; exa = 10¹⁸; emerging for 2030s high-speed networks.
EB/sexabyte per second8 × 10¹⁸Decimal; 1 EB = 10¹⁸ bytes.
Eibpsexbibit per second1.1529 × 10¹⁸IEC binary; exbi = 2⁶⁰.
EiB/sexbibyte per second9.223 × 10¹⁸Binary; 1 EiB = 2⁶⁰ bytes.

Practical examples and applications

Network and internet speeds

Network and internet speeds represent practical applications of data-rate units in everyday , where theoretical maxima often differ from real-world performance due to factors like protocol overhead, signal interference, and . For instance, traditional dial-up connections using V.90 modems achieved a maximum downstream speed of 56 kilobits per second (kbps), enabling basic web browsing but limiting file downloads to mere kilobytes per second under optimal conditions. In contrast, modern wireless local area networks (WLANs) leveraging (IEEE 802.11ax) deliver typical real-world speeds of around 1 gigabit per second (Gbps) in home or office environments, supporting high-definition streaming and multiple devices simultaneously, though this is well below the standard's theoretical peak of 9.6 Gbps due to channel contention and distance. Mobile and wide-area networks have scaled dramatically with technology, which boasts a peak downlink speed of 20 Gbps as defined in M.2410 requirements for enhanced mobile broadband, facilitating ultra-high-definition video calls and applications in low-mobility scenarios. Fixed broadband via fiber optics, such as XGS-PON systems under G.9807.1, provides symmetric 10 Gbps upload and download rates, ideal for cloud backups and professional content creation where balanced throughput is essential. However, effective speeds in these networks often reach 70-90% of theoretical values owing to overhead from error correction, routing latency, and , ensuring reliability but capping instantaneous performance. As of late 2025, satellite services like deliver median download speeds of around 100-200 Mbps in the , bridging connectivity gaps in remote areas with low-Earth orbit constellations that minimize propagation delays compared to geostationary alternatives. Globally, the average download speed stands at approximately 60 Mbps as of early 2025 according to Ookla's Speedtest Global Index data, reflecting widespread / adoption but highlighting disparities in infrastructure across regions. Looking ahead, research trials target peak rates of 1 terabit per second (Tbps), as outlined in Ericsson's spectrum studies, promising transformative applications in holographic communications and real-time processing.

Storage and media transfer rates

Storage and media transfer rates refer to the speed at which data is read from or written to devices and during playback, typically measured in bytes per second (e.g., /s or GB/s) to reflect the volume of data handled by these systems. These rates are crucial for tasks like file copying, , and content rendering, where byte-centric units provide a direct measure of throughput relevant to storage operations. Unlike bit-based metrics, storage rates emphasize practical data movement within local devices, often limited by standards, media type, and hardware capabilities. A common example is the USB 3.2 Gen 1 interface, which supports a theoretical maximum transfer rate of 625 MB/s, enabling rapid data exchange between external drives and computers for large media files. For media playback, video decoding from storage typically requires a sustained rate of around 3 MB/s to handle bitrates of 25 Mbps, ensuring smooth rendering without buffering on local devices. Optical media like Blu-ray discs demonstrate varying read speeds; a 12x drive achieves up to 54 MB/s, sufficient for high-definition content extraction. In , solid-state drives (SSDs) using PCIe 5.0 interfaces represent the pinnacle of storage performance, with sequential read speeds reaching 14 GB/s on consumer models like the 9100 PRO, ideal for demanding applications such as video editing workflows. By contrast, traditional hard disk drives (HDDs) top out at sustained sequential rates of about 250 MB/s for modern 7200 RPM models, making them suitable for bulk archival but slower for interactive media tasks. These rates often distinguish between burst performance—short peaks enabled by onboard s—and sustained throughput, where cache exhaustion can drop speeds significantly during prolonged transfers, such as copying gigabyte-sized video files. To enhance performance, configurations like 0 stripe data across multiple drives, effectively doubling sequential transfer rates; for instance, two 500 MB/s SSDs in 0 can achieve approximately 1 GB/s for large file operations. This setup amplifies media transfer efficiency but requires careful consideration of drive matching to avoid bottlenecks. Overall, these rates underscore the evolution from mechanical HDDs to flash-based SSDs, prioritizing reliability and speed in storage ecosystems.

Emerging high-speed technologies

Emerging high-speed technologies are pushing data rates into the petabit per second (Pbps) range and beyond through innovations in optical fibers, communications, and space-based systems, addressing the growing demands of , networks, and deep-space exploration. In December 2024, NTT demonstrated a rate of 455 terabits per second (Tbps) over 1,000 km using standard single-mode optical fibers, leveraging advanced multi-core fiber and techniques as part of their Innovative Optical and (IOWN) initiative. This achievement highlights the potential for scaling beyond current commercial limits, with the system designed to support 125 times higher capacity while reducing power consumption by a factor of 100 compared to traditional copper-based interconnects. In (THz) communications, experimental demonstrations have reached multi-gigabit speeds over practical distances, paving the way for ultra-short-range, high-capacity links in and beyond. A study reported an 84 gigabits per second (Gbps) air interface rate at 0.22 THz over 1.26 km, utilizing advanced modulation and to mitigate atmospheric absorption, demonstrating feasibility for backhaul in dense urban environments. Theoretical models for THz systems suggest capacities up to several Tbps in controlled short-range scenarios, driven by the vast available above 100 GHz, though practical implementations remain constrained by losses. For space applications, NASA's (DSOC) experiment on the mission achieved a record 267 Mbps from over 140 million miles away in , using laser-based to beam and test data back to Earth. Future iterations aim for 10 to 100 times higher rates, potentially reaching several Gbps, to support from Mars missions and beyond. Prototypes in integrated are also advancing internal data rates for systems. Intel's 2024 optical I/O prototype enables 4 terabytes per second (TB/s) bidirectional transfer—equivalent to 32 Tbps per direction—via co-packaged , targeting data centers to reduce and power draw in high-performance interconnects. In , emerging brain-inspired architectures promise massive , with theoretical internal bandwidths scaling to exabyte per second (EB/s) levels in large-scale systems, though current prototypes focus on efficiency rather than peak rates. Challenges in these technologies include approaching the nonlinear limit of approximately 100 Tbps per fiber pair in standard single-mode fibers, beyond which nonlinear optical effects degrade signals. Additionally, power consumption and thermal management remain critical hurdles; high-speed transceivers generate significant heat, necessitating advanced cooling solutions to prevent performance throttling and hardware failure in dense deployments. Quantum communication links offer theoretical potentials up to 10 Pbps in entanglement-based networks, but practical demos prioritize secure over raw throughput.

References

  1. [1]
    Definitions of the SI units: The binary prefixes
    Faced with this reality, the IEEE Standards Board decided that IEEE standards will use the conventional, internationally adopted, definitions of the SI prefixes ...Missing: rate | Show results with:rate
  2. [2]
    When bandwidth and storage size matters: Bits vs. bytes - Red Hat
    Sep 3, 2020 · When data is transmitted over a network medium, it is typically written in bits/s, kilobits/s(kbps), megabits/s (Mbps), or gigabits/s (Gbps).
  3. [3]
    IEEE 802.3-2022 - IEEE SA
    Jul 29, 2022 · This project will specify additions to and appropriate modifications of IEEE Std 802.3 to add 100 Mb/s Physical Layer specifications and ...
  4. [4]
    About bits and bytes: prefixes for binary multiples - IEC
    In the IEC 80000-13:2008 Standard, all branches of the IT industry, and in fact our society at large, have a tool with which to iron out this inconsistency.
  5. [5]
    IEC 80000-13:2008 - Quantities and units — Part 13 - ISO
    IEC 80000-13:2008 provides names, symbols, and definitions for quantities and units in information science and technology, and is now withdrawn.
  6. [6]
    What is a Data Transfer Rate (DTR)? | Definition from TechTarget
    Dec 29, 2022 · The data transfer rate (DTR) is the amount of digital data that's moved from one place to another in a given time.
  7. [7]
    Data Transmission Rate - an overview | ScienceDirect Topics
    Data transmission rate refers to the speed at which data is transmitted over a communication system, typically measured in bits per second (bps), ...
  8. [8]
    What are bits per second (bps or bit/sec)? - TechTarget
    Jun 1, 2021 · In data communications, bits per second (bps or bit/sec) is a common measure of data speed for computer Modems and transmission carriers.
  9. [9]
    What is bit (binary digit) in computing? | Definition from TechTarget
    Jun 6, 2025 · A bit (binary digit) is the smallest unit of data that a computer can process and store. It can have only one of two values: 0 or 1.
  10. [10]
    A Brief History of Data Communications | Incognito Insights
    Jul 11, 2017 · 1958: American Telephone and Telegraph Bring the Digital Subset ... Speed was boosted by over 4 times, sending a new record of 110 bits per second ...
  11. [11]
    What's The Difference Between Bit Rate And Baud Rate?
    Serial-data speed is usually stated in terms of bit rate. However, another oft-quoted measure of speed is baud rate. Though the two aren't the same, ...
  12. [12]
    Data Rate vs Bandwidth: What's the Difference? - Altium Resources
    Oct 22, 2020 · Data rate is exactly what it sounds like: the number of bits transmitted through a channel or by a component per unit of time.
  13. [13]
    Byte - Glossary | CSRC - NIST Computer Security Resource Center
    A byte is a sequence of eight bits, treated as a single entity or an array of eight individual bits.
  14. [14]
    Why Do We Use Bits to Measure Internet Speed but Bytes to ... - NCTA
    A byte on the other hand is a unit of memory that usually contains 8 bits. This is because historically, 8 bits are needed to encode a single character of text.
  15. [15]
    Bits, Bytes and Bandwidth Reference Guide - SpeedGuide
    Bits, Bytes and Bandwidth Reference Guide · 1 bit (b) = 0 or 1 = one binary digit · 1 kilobit ( kb) = 10^3 bits = 1,000 bits · 1 Megabit (Mb) = 10^6 bits = ...
  16. [16]
    SI prefixes - BIPM
    Decimal multiples and submultiples of SI units ; centi. c. 10 ; milli. m. 10 ; micro. µ. 10 ; nano. n. 10 ...
  17. [17]
    NIST Guide to the SI, Appendix D: Bibliography
    Jan 28, 2016 · Letter symbols to be used in electrical technology, Part 2: Telecommunications and electronics, IEC 60027-2 (2005). Note: As pointed out in Sec.Missing: rates | Show results with:rates
  18. [18]
    SI and Binary Prefixes: Clearing the Confusion
    Aug 1, 2023 · SI prefixes are misused for binary multiples, like KB for 1024 bytes instead of 1000. Correct binary prefixes are kibi, mebi, gibi, and tebi.Missing: rate bit/ s<|control11|><|separator|>
  19. [19]
    MiB vs. MB: Understanding the Units for Data Rates and Bandwidth
    kilo, Mega, Giga or Tera are prefixes meant for and based on the decimal metric system. They stand for multiples of 103 (1000) and are standardized in the ...Missing: transfer | Show results with:transfer
  20. [20]
    Data storage values - IBM
    Table 1 compares the names (prefixes), symbols, and values of decimal and binary units. Table 2 shows the increasing percentage of difference between decimal ...
  21. [21]
    Western Digital Settles Hard-Drive Capacity Lawsuit - CRN
    Jun 28, 2006 · The lawsuit claimed Western Digital overstated the capacities of its hard drives by about 7 percent. For example, the settlement documents said, ...Missing: 2000s | Show results with:2000s
  22. [22]
    Seagate settles suit over capacity of hard drives - Los Angeles Times
    Nov 3, 2007 · Seagate Technology has agreed to reimburse potentially millions of customers and pay as much as $1.79 million in plaintiffs' attorney fees ...Missing: 2000s sources
  23. [23]
    5G service requirements - 3GPP
    Feb 27, 2017 · User experienced data rates vary, from 1 Gbps downlink and 500 Mpbs uplink for indoor hotspot environments to 50 Mbps downlink and 25 Mbps ...Missing: decimal | Show results with:decimal
  24. [24]
    What are Binary Prefixs? Why Memory/Storage Capacity Less than ...
    Jul 23, 2025 · Binary prefixes are units of measurement of storage capacity of Hard disk or SSD. The difference between the advertised storage capacity and actual capacity in ...Missing: percentage | Show results with:percentage
  25. [25]
    What is Kilo, Mega, Giga, Tera, Peta, Exa, Zetta and All That?
    Nov 28, 2022 · For example, 1 kilobit per second (kbps) is equal to 103, or 1,000 bits per second (bps); 1 megabit per second (Mbps) is equal to 106, or ...
  26. [26]
    Convert kilobit/second to bps - Conversion of Measurement Units
    Unlike many other computer-related units, 1 kbps is traditionally defined as 1000 bps, not 1024 bps. Metric conversions and more. ConvertUnits.com provides an ...
  27. [27]
    Kilobits Per Second (Kbps) - Data Rate Conversions (Decimal)
    A Kilobit per second is a unit used to measure data transfer rates and is based on Decimal multiples of bits. The symbol for Kilobit per second is Kbps or ...
  28. [28]
    Convert bps to Kbps - Data Rate Conversions (Decimal)
    There are 1,000 bits per second in a Kilobit per second. What is a Kilobit per second (Kbps)?. A Kilobit per second is a unit used to measure data transfer ...
  29. [29]
    Dial-up modems - Modem speeds - USR
    For dial-up modems, due to the technology used to deliver normal phone service, the fastest rate available is 56 kilobits (one thousand bits) per second.
  30. [30]
    A Complete Guide on Audio Bitrate - Gumlet
    Aug 5, 2024 · Audio bitrate is the measure of bits processed over time, defining audio quality, and is represented in kilobits per second (kbps).
  31. [31]
    Convert Kbps to Bps - Data Rate Conversions (Decimal)
    Kilobits per second to Bytes per second Conversion Chart ; 1 Kbps, 125 Bps ; 2 Kbps, 250 Bps ; 3 Kbps, 375 Bps ; 4 Kbps, 500 Bps.
  32. [32]
    Convert Mbps to bps - Data Rate Conversions (Decimal)
    The symbol for bit per second is bps or b/s or bit/s. There are 1,000,000 bits per second in a Megabit per second.
  33. [33]
    What does Mbps (megabits per second) mean? - IONOS
    Jun 27, 2023 · Megabits per second is a unit used to measure the speed of a network and, along with the ping command, can be used to determine the quality of a connection.
  34. [34]
    What are Megabits per second(Mbps)? - GeeksforGeeks
    Jul 23, 2025 · One megabit per second (Mbps) can transmit one million bits, or roughly one small picture, per second. It can also be expressed as Mbit/s or Mb/s.
  35. [35]
    DSL vs. Cable Internet: Speeds & Plans | HighSpeedInternet.com
    Mar 28, 2025 · Faster DSL plans can offer connections over 100 Mbps, which is plenty for most online activities, including video streaming. . For context ...
  36. [36]
    Speedtest Global Index – Internet Speed around the world
    The Speedtest Global Index now includes internet speed rankings for some of the world's largest cities. Use the toggle below to explore median download speeds.Missing: ITU | Show results with:ITU
  37. [37]
    Gigabits Per Second (Gbps) - Data Rate Conversions (Decimal)
    A Gigabit per second is a unit used to measure data transfer rates and is based on Decimal multiples of bits. The symbol for Gigabit per second is Gbps or ...
  38. [38]
  39. [39]
  40. [40]
    What Are 5G Speeds? - Cisco
    5G speeds are between 10 and 20 Gbps, with download speeds 1.4 to 14 times faster than 4G, and upload speeds up to 30% faster.What is a 5G network? · What are 5G speeds for... · What is mmWave?
  41. [41]
    IEEE P802.3df Task Force - IEEE 802
    Feb 20, 2024 · The work of the IEEE P802. 3df 400 Gb/s and 800 Gb/s Ethernet Task Force completed with the approval of IEEE Std 802.3df-2024 by the IEEE-SA ...Missing: date | Show results with:date
  42. [42]
    Petabits Per Second (Pbps) - Data Rate Conversions (Decimal)
    A Petabit per second is a unit used to measure data transfer rates and is based on Decimal multiples of bits. The symbol for Petabit per second is Pbps or ...
  43. [43]
    Petabits Per Second to Gigabits Per Second - Kyle's Converter
    Unit Descriptions. 1 Petabit per Second: 1 Petabit per second is equal to 1,000,000,000,000,000 bits per second. A terabit contains 1000^5 bits.Missing: definition | Show results with:definition
  44. [44]
    World Record Optical Fiber Transmission Capacity Doubles to 22.9 ...
    Nov 30, 2023 · NICT demonstrated the possibility of fiber-optic data communication at 22.9 petabits per second, which is more than double the previous record of 10.66 ...
  45. [45]
    World record optical fiber transmission capacity doubles to 22.9 ...
    Nov 30, 2023 · A record-breaking data rate of 22.9 petabits per second using only a single optical fiber, which was more than double the previous world record of 10.66 ...
  46. [46]
    World record optical fiber transmission capacity doubles to 22.9 ...
    Nov 30, 2023 · A record-breaking transmission capacity of 22.9 petabits per second in a single optical fiber was demonstrated.
  47. [47]
    With 19-core Optical Fiber with Standard Cladding Diameter ... - NICT
    May 29, 2025 · ... a transmission rate of 1.02 petabits per second over a distance of 1,808 km. Achieved using a newly developed standard 19-core optical fiber ...
  48. [48]
    World record achieved in transmission capacity and distance
    May 29, 2025 · ... a transmission rate of 1.02 petabits per second over a distance of 1,808 km. - Achieved using a newly developed standard 19-core optical ...
  49. [49]
    Kilobytes Per Second (KBps) - Data Rate Conversions (Decimal)
    A Kilobyte per second is a unit used to measure data transfer rates and is based on Decimal multiples of bits. The symbol for Kilobyte per second is KBps or ...
  50. [50]
    What is a megabyte (MB)? | Definition from TechTarget
    Dec 20, 2023 · A megabyte is a unit of data capacity that is equal to 1,000,000 bytes in decimal notation (base 10) or 1,048,576 bytes in binary notation ...
  51. [51]
    SATA SSD Performance & Benchmark Tests 2025 - SSD-Tester
    In practice, a speed of up to approx. 550 MB/s is possible for reading and writing. SATA SSDs are slower compared to PCIe NVMe SSDs, which makes them cheaper.
  52. [52]
    What is the actual data transfer rate for USB 3.0? - AKiTiO
    What is the actual data transfer rate for USB 3.0? · USB 3.0 up to 5Gbps (625MB/s) · USB 2.0 up to 480Mbps (60MB/s).
  53. [53]
    What is a gigabyte (GB) and how is it measured? - TechTarget
    Oct 21, 2021 · The base 10 definition of gigabyte uses the decimal system to show that 1 GB is equal to 109 bytes, or 1 billion bytes. Today, most storage ...
  54. [54]
    How much is 1 GB per second bandwidth? Is it 2^30 bps or 10^9 bps?
    Jan 17, 2015 · A gigabit per second (Gbps) is equal to 1,000,000,000 bits per second. A gigabyte per second (GBps) is equal to 8,000,000,000 bits per second.When is a gigabyte not less than 10^12 bytes? - QuoraHow many bytes are in a gigabyte (GB) and how does it compare to ...More results from www.quora.com<|control11|><|separator|>
  55. [55]
    Silicon Motion: PCIe 5.0 SSD Controller to Debut Next Year
    Feb 6, 2021 · The PCIe 5.0 interface will increase data transfer speeds to 32 GT/s per ... PCIe x4 slot can transfer up to ~16 GB/s. Increased transfer ...
  56. [56]
    Breaking the 100 GB/s Barrier: Astera Labs and Micron Demonstrate ...
    Breaking the 100 GB/s Barrier: Astera Labs and Micron Demonstrate Production-Ready PCIe 6 Storage Performance at FMS 2025 · The AI Storage Performance Crisis.
  57. [57]
    Terabytes Per Second (TBps) - Data Rate Conversions (Decimal)
    A Terabyte per second is a unit used to measure data transfer rates and is based on "Decimal multiples of bits".Missing: definition | Show results with:definition
  58. [58]
    NVLink & NVSwitch: Fastest HPC Data Center Platform | NVIDIA
    A single NVIDIA Blackwell GPU supports up to 18 NVLink 100 gigabyte-per-second (GB/s) connections for a total bandwidth of 1.8 terabytes per second (TB/s )—2X ...
  59. [59]
    Our Supercomputers - EuroHPC JU - European Union
    In total, LUMI has a storage of 117 petabytes and a maximum I/O bandwidth of 2 terabytes per second. Applications: AI, especially deep learning, and ...
  60. [60]
    Sycomp Delivers Cloud-Based AI Storage at 1.2 Terabytes per ...
    Jun 6, 2025 · Sycomp has delivered a parallel file system, an optimized AI platform, capable of achieving greater than one terabyte per second (1.2 TB/s) of throughput.
  61. [61]
    units(7) - Linux manual page - man7.org
    The names are formed by taking the first syllable of the names of the decimal prefix with roughly the same size, followed by "bi" for "binary".
  62. [62]
    iostat(1) - Linux manual page - man7.org
    ... (kibibytes, mebibytes) written to the device per second. dsec/s (dkB/s, dMB/s) The number of sectors (kibibytes, mebibytes) discarded for the device per second.
  63. [63]
    Make option to see Mbps or MiB/s · Issue #19425 - GitHub
    Aug 8, 2023 · Make Mbps standard, and a option to switch to Gbps, MBps, GBps or MiBps. Use case: See more relevant speed when downloading or uploading.
  64. [64]
    BandwidthTest reports in MB or MiB? - NVIDIA Developer Forums
    Jul 28, 2013 · Is the MB/sec reported in the bandwidthTest utility a 10^9 SI megabyte or a 2^20 mebibyte? The other values reported for GPU specs are in the 2 ...
  65. [65]
    IEEE 1541-2021 - IEEE SA
    Feb 18, 2022 · IEEE 1541-2021 defines prefixes for binary multiples of 2^10n, where n=1-8, primarily used in information technology.
  66. [66]
    Theoretical Maximum Memory Bandwidth for Intel® Core™ X-Series...
    For example: For DDR4 2933 the memory supported in some core-x -series is (1466.67 X 2) X 8 (# of bytes of width) X 4 (# of channels) = 93,866.88 MB/s ...Missing: GiB/ | Show results with:GiB/
  67. [67]
    1541-2021 - IEEE Standard for Prefixes for Binary Multiples
    Feb 18, 2022 · This standard defines prefixes and letter symbols that denote multiplication of a unit by the binary multiplier 1024 (equal to 210) and integral powers thereof.
  68. [68]
    Convert bps to Bps - Data Rate Conversions (Decimal)
    The formula to convert from bits per second to Bytes per second is: Bytes per second = bits per second ÷ 8.
  69. [69]
    Convert Mbps to MB/s - Conversion of Measurement Units
    75 Mbps to MB/s = 9.375 MB/s. 100 Mbps to MB/s = 12.5 MB/s. Want other units? You can do the reverse unit conversion from MB/s to Mbps, or enter any two units ...
  70. [70]
    Open Source Fast File Transfers | FileCatalyst - GoAnywhere
    Jun 18, 2012 · 20% protocol overhead, example: (100 Mbps link will only be able to send at 80 Mbps); no jumbo packet support; no multi-threading, meaning that ...
  71. [71]
    Metric (SI) Prefixes
    ### Summary of Data Rate Units and Bit per Second from NIST Metric (SI) Prefixes
  72. [72]
    IEC or SI units (binary prefixes) used for network measurement?
    Oct 17, 2013 · The IEC recommends the standard use of SI decimal prefixes. They have created additional, unambiguous, binary prefixes kibi-, mebi-, gibi- etc.
  73. [73]
    Decimal and Binary Prefixes - Blocks and Files
    Apr 23, 2022 · Decimal and Binary Prefixes – The SI system, a decimal system, measures units in powers of 10. The binary numeric system measure units in ...<|control11|><|separator|>
  74. [74]
    humanize - PyPI
    This modest package contains various common humanization utilities, like turning a number into a fuzzy human-readable duration ("3 minutes ago") or into a human ...
  75. [75]
    How fast is Wi-Fi 6? Here are our latest speed test results - CNET
    Aug 2, 2019 · The first Wi-Fi 6 router we tested was the Netgear Nighthawk AX12, which promises speeds of up to 1.2 Gbps on the 2.4GHz band and up to 4.8 Gbps on the 5GHz ...
  76. [76]
    [PDF] Report ITU-R M.2410-0
    The minimum requirements for peak data rate are as follows: –. Downlink peak data rate is 20 Gbit/s. –. Uplink peak data rate is 10 Gbit/s. 4.2. Peak spectral ...Missing: Gbps | Show results with:Gbps
  77. [77]
    Best practices for right-sizing your Apache Kafka clusters to optimize ...
    Mar 17, 2022 · For production clusters, it's a best practice to target the actual throughput at 80% of its theoretical sustained throughput limit. Consider ...
  78. [78]
    SpaceX: Starlink Speeds Have Increased 50% This Year - PCMag
    Oct 27, 2025 · The graph shows that median download rates for Starlink were between 140Mbps and 160Mbps in January, but have since risen to 210 to 220Mbps—or ...
  79. [79]
    6G spectrum - enabling the future mobile life beyond 2030 - Ericsson
    This white paper focuses on the role of spectrum to unleash the full potential of 6G, the importance of existing spectrum as well as additional spectrum.
  80. [80]
    What's the Difference between USB 3.0, 3.2 Gen 1 and 3.2 Gen 2?
    The only difference between USB 3.2 Gen 1 and USB 3.2 Gen 2 is speed. ... Up to 2,000MB/s Read, 2,000MB/s Write. SHOW MORE + SHOW LESS -. Learn more Buy.
  81. [81]
    USB 3.2 Specification
    USB 3.2 identifies three transfer rates – 20Gbps, 10Gbps, and 5Gbps. Key characteristics of the USB 3.2 specification include: Defines multi-lane operation for ...
  82. [82]
    Choose live encoder settings, bitrates, and resolutions - YouTube Help
    Audio sample rate: 44.1 KHz for stereo audio, 48 KHz for 5.1 surround sound. Audio bitrate: 128-Kbps for stereo or 384 Kbps for 5.1 surround sound. Color ...
  83. [83]
    How fast can you read/write data on a Blu-ray disc - Blue-Cloner
    According to the Blu-ray Disc specification, 1x speed is defined as 36Mbps. However, as BD-ROM movies will require a 54Mbps data transfer rate the minimum ...
  84. [84]
    Amazon.com: Samsung SSD 9100 PRO 2TB, PCIe 5.0x4 M.2 2280 ...
    BREAKTHROUGH PCIe 5.0 PERFORMANCE: Supercharge your workflow and gaming with PCIe 5.0, boasting up to 14,800/13,400 MB/s sequential read/write speeds.<|control11|><|separator|>
  85. [85]
    Fastest hard drives of 2025 - TechRadar
    Jan 30, 2025 · It boasts exceptional read and write speeds; we recorded speeds of 286.8MB/s and 279.3MBps respectively in the 20TB iteration, based on testing ...
  86. [86]
    SSD Cache Explained: Burst vs. Sustained Performance
    Aug 15, 2025 · In SSDs, burst capacity allows the storage to run faster than baseline speed during short periods, improving responsiveness during workloads.
  87. [87]
    Understanding RAID Performance at Various Levels | Arcserve
    Feb 14, 2024 · RAID 0 is always the highest-performing level. An example would be an eight-spindle RAID 0 array. If an individual drive in the array delivers ...
  88. [88]
    HDD Benchmarks Hierarchy 2025: Here's all the hard disks we've ...
    May 28, 2025 · Our HDD benchmarks hierarchy shows all of the high-capacity hard drives that we've tested over the years, ranked in order of overall sequential throughput.
  89. [89]
  90. [90]
    High-speed 0.22 THz communication system with 84 Gbps for real ...
    Sep 13, 2024 · In this work, we report a 0.22 THz communication system, which can provide 84 Gbps air interface rate over 1.26 kilometers distance.
  91. [91]
    A Survey on Advancements in THz Technology for 6G - arXiv
    Jul 2, 2024 · In this paper, we survey recent advancements in THz technology and its role in future 6G wireless networks, with a particular emphasis on the 200–400 GHz ...Missing: Tbps | Show results with:Tbps
  92. [92]
    NASA's Optical Comms Demo Transmits Data Over 140 Million Miles
    Apr 25, 2024 · NASA's optical communications demonstration has shown that it can transmit test data at a maximum rate of 267 megabits per second (Mbps) from ...
  93. [93]
    Laser communication: the future of communicating in space
    Dec 15, 2023 · NASA assumes a 10 to 100-fold increase in data volume with new laser communication systems. This, in turn, will help with future human and ...
  94. [94]
    Intel Demonstrates First Fully Integrated Optical I/O Chiplet
    Jun 26, 2024 · Intel's OCI chiplet enables co-packaged optical input/output in emerging AI infrastructure for data centers and high performance computing applications.Missing: exabit | Show results with:exabit
  95. [95]
    Approaching the Non-Linear Shannon Limit | Request PDF
    Aug 5, 2025 · However, the maximum capacity of commonly deployed single-mode fibers (SMFs) is limited to approximately 100 Tbps. This limitation arises from ...
  96. [96]
    Thermal challenges and Mitigation Strategies of High speed designs
    Aug 8, 2023 · 2. Hot Spots: High-speed devices can experience localized areas of intense heat, known as hot spots, which can degrade performance and ...
  97. [97]
    Towards metropolitan free-space quantum networks - Nature
    Sep 27, 2023 · We achieved up to 5.7 kbps secure key rates over two different links and under various link conditions, with over 2.5 kbps in full daylight, ...