pebibyte (PiB)

A pebibyte (PiB) is a unit of measure used to describe data capacity. The prefix pebi was created as part of the binary system for measuring computing and storage capacity, which is based on powers of two. One pebibyte equals 250 or 1,125,899,906,842,624 bytes.

The term pebibyte and the other binary prefixes -- kibi, mebi, gibi, tebi, exbi, zebi and yobi – were created by the International Electrotechnical Commission (IEC) in 1998. Prior to developing the binary prefixes, computing capacity was measured using the metric prefixes of the International System of Units (SI). These older SI prefixes -- including kilo, mega, giga, tera, peta, exa, zetta and yotta -- are also referred to as the decimal system of prefixes.

Before 1998, the SI prefixes were used to refer to both the power-of-10 multipliers in the decimal system of measurement and power-of-two multipliers used in the binary system. During the decades when the SI prefixes were used to refer to both the decimal and binary multiplier systems, context was the only way to determine which multiplier was being referred to when, say, the term petabyte (PB) was used. That term could refer to either 1015 (1,000,000,000,000,000) bytes or 250 (1,125,899,906,842,624) bytes.

Since the IEC created the binary prefixes in 1998, the system people use often depends on what discipline or industry they work in. The decimal system's power-of-10 metric prefixes are more likely to be used by people in communications, electronics and physics. The power-of-two binary prefixes are more likely to be used in the IT and data storage industries.

What pebibytes are used for

In the early days of computing, using the same prefixes to refer to both the power-of-10 and the power-of-two multipliers wasn't a significant problem. Because data storage capacity was much smaller than what it is today, the difference between the two systems was relatively small.

For instance, 1 kilobyte (KB) of data capacity equals 1,000 bytes using the power-of-10 system. Using the power-of-two system, 1 KB of data equals 1,024 bytes. The 24-byte difference between the two systems was small enough to be considered negligible. However, as computing and data capacity have grown over the years, the difference between the numbers generated by the two measurement systems has been magnified.

A petabyte of data using the power-of-10 decimal system equals 1,000,000,000,000,000 bytes, while a petabyte of data using the power-of-two binary system equals 1,125,899,906,842,624 bytes. That's a difference of more than 125 trillion bytes. At higher levels of capacity, the differences in the numbers generated by the two measurement systems began to cause problems and could no longer be ignored.

This video from Manx Computing and ICT
explains units of memory.

One problem was particularly bothersome to users of hard disk drives (HDDs) and solid-state drives (SSDs). Disk drive manufacturers used the decimal system when they labeled the capacity of HDDs and SSDs. But operating system (OS) vendors often used the binary system to report computer memory and data storage capacity. When a customer bought a new hard drive that the manufacturer labeled using the decimal system as having 100 gigabytes (GB) of capacity, the computer OS, using the binary system of measurement, would report that it had only 93.13 GB of capacity. Customers weren't happy to lose what looked like nearly 7 GB of capacity before they had even used the HDD.

The development of the prefixes associated with the IEC standard for the binary system of measurement was supposed to eliminate this type of confusion. In the example above, the OS would have used the binary system of prefixes instead of the decimal system, and would have reported 93.13 gibibytes (GiB) instead of 93.13 GB of capacity. The use of the new prefixes was supposed to let consumers know that the OS vendor was using a different system of measurement from the HDD manufacturer.

Binary and decimal data measurement systems

However, the new prefixes haven't been widely used, and the creation of the binary prefixes didn't solve the capacity measurement problem. The binary prefixes are mostly used by academics, technical writers and open system developers. Commercial vendors and IT professionals generally aren't using the binary prefixes.

How large is a pebibyte?

A pebibyte is at the larger end of the binary measurement scale. It's larger than the following binary data capacity measures:

  • A PiB is equal to 1,125,899,906,842,624 bytes.
  • A kibibyte (KiB) -- a PiB is equal to 1,099,511,627,776 KiB.
  • A mebibyte (MiB) – a PiB is equal to 1,073,741,824 MiB.
  • A gibibyte -- a PiB is equal to 1,048,576 GiB.
  • A tebibyte – a PiB is equal to 1,024 TiB.

Binary data capacity measures that are larger than a pebibyte include an exbibyte (EiB), which is 1,024 times the size of a PiB. A zebibyte (ZiB) is 1, 048,576 times the size of a PiB. And a yobibyte (YiB) is 1,073,741,824 times the size of a PiB.

Visualizing the size of a pebibyte

With the advent of big data and data analytics, an increasing number of organizations have data stores in the pebibyte range. In 2017, the European nuclear research center CERN announced that it has 200 PB, the equivalent of 178 PiB, archived in its tape library. Facebook, Google and YouTube combined accounted for approximately 31,000 PiB (31 EiB) of data generated in 2017.

Here are some more concrete ways to visualize petabytes of data in the real world:

  • A PiB can store more than a trillion pages of plain text type.
  • 239,553 movies on standard, single-layer DVDs is equivalent to 1 PiB.
  • All U.S. research libraries combined hold 1.78 PiB of data.
  • A pebibyte of songs, averaging four minutes of playing time each, would last more than 2,250 years playing continuously.

Pebibyte vs. petabyte and tebibyte

As previously mentioned, one pebibyte is equal to 250 or 1,125,899,906,842,624 bytes, and one petabyte is equal to 1015 or 1,000,000,000,000,000 bytes. A pebibyte is more than 12.5% larger than a petabyte. When assessing storage capacity, that's a significant difference.

When working with the lower capacity ranges, such as a kibibyte vs. a kilobyte or a mebibyte vs. a megabyte, the differences between binary and decimal units of measure is negligible. However, with pebibytes vs. petabytes, the difference between the two systems matters and must be taken into account.

A pebibyte equals 1,024 TiB. To do the conversion, divide the 1,125,899,906,842,624 bytes that make up a pebibyte by the number of bytes in a tebibyte, or 1,099,511,627,776. That gives you 1,024 TiB. 

binary system of measurement

Each consecutively higher prefix in the binary system of capacity measurement has 210, or 1,024, times the number of bytes of the next lower prefix. A pebibyte is 1,024 times larger than a tebibyte in terms of bytes. An exbibyte is 1,024 times larger than a pebibyte. And a zebibyte is 1,024 times larger than an exbibyte and so on.

How much does a pebibyte cost?

A pebibyte of storage will vary in cost depending on the type of storage and device used. Using 512 GB HDDs, ranging in price from $44 to $65, 1 PiB of storage would cost from $97,000 to $143,000 as of mid-2018.

Specific products that scale in the pebibyte range include Dell EMC Isilon Scale-Out NAS Storage with a single file system that spans the nodes within a cluster can scale to more than 17 PB per cluster. It has a starting list price of $42,552. Alternatively, there's cloud storage. A pebibyte of Google Coldline storage costs about $7,880 a month. A pebibyte of Amazon Glacier storage costs slightly more than $4,500 a month.

This was last updated in September 2018

Continue Reading About pebibyte (PiB)

Dig Deeper on Storage management and analytics

Disaster Recovery
Data Backup
Data Center