Petabyte to Word

1 petabyte (PB) = 1,000 terabytes (TB) in the SI decimal system. In the binary IEC system, 1 pebibyte (PiB) = 1,024 tebibytes (TiB) = 1,125,899,906,842,624 bytes. The distinction matters for enterprise storage procurement: a petabyte of raw disk capacity appears as about 909 TiB in an OS reporting binary units.

Petabyte-scale storage is common at: social media platforms (Facebook/Meta stores over 100 PB of photos alone), streaming services (Netflix's content library is estimated at 100+ PB), government agencies (US NSA, CERN particle physics data), genomic research institutions, and large financial exchanges storing tick-level trading data. Major cloud providers (AWS, Azure, GCP) collectively store zettabytes.

In 2024, cloud storage costs roughly $20–25 per TB per month (S3 standard tier), making 1 PB approximately $20,000–$25,000/month. Raw enterprise disk hardware for 1 PB runs about $20,000–$50,000 upfront (at $20–50 per TB for high-density drives), plus ongoing power, cooling, and management overhead. Tape-based archival storage is considerably cheaper at $2–5 per TB.

YouTube users upload approximately 500 hours of video per minute, or 720,000 hours per day. At an average compressed size of 1–2 GB per hour of HD video, that equates to roughly 720–1,440 TB (0.7–1.4 PB) of new video data per day — before YouTube re-encodes into multiple formats and quality levels, which multiplies storage requirements several-fold.

The SI prefix hierarchy above petabyte: exabyte (EB, 10¹⁸ bytes), zettabyte (ZB, 10²¹ bytes), yottabyte (YB, 10²⁴ bytes), ronnabyte (RB, 10²⁷ bytes), and quettabyte (QB, 10³⁰ bytes) — the last two added by the BIPM in 2022. Current global data storage is estimated in the hundreds of exabytes; no single organisation approaches yottabyte scale.

A word's size depends on the CPU architecture. In x86/x64 (Intel/AMD) documentation: word = 16 bits, DWORD = 32 bits, QWORD = 64 bits. In ARM 32-bit: word = 32 bits. In most modern 64-bit systems (excluding x86 documentation): word = 64 bits. When reading technical documentation, always check the architecture's definition, as "word" is not a universal fixed size.

In Windows API documentation and x86 architecture, a DWORD (Double Word) = 32 bits = 4 bytes, capable of holding values 0–4,294,967,295 (unsigned) or -2,147,483,648 to 2,147,483,647 (signed). DWORD is the most common fixed-width integer type in the Windows API, used for flags, handles, and return codes. The equivalent in modern C/C++ is uint32_t (unsigned) or int32_t (signed).

A CPU's word size determines: (1) the maximum addressable memory — a 32-bit CPU addresses up to 4 GiB (2³² bytes); a 64-bit CPU addresses up to 16 EiB (2⁶⁴ bytes); (2) the precision of integer arithmetic — a 64-bit word handles numbers up to ~18.4 × 10¹⁸ in a single instruction; (3) performance — operations on data smaller than the word size may require extra sign-extension instructions on some architectures.

Modern x86-64 CPUs (Intel Core, AMD Ryzen) have 64-bit general-purpose registers, so their native word size is 64 bits for most operations. However, x86 documentation maintains the legacy definition: "word" = 16 bits, DWORD = 32 bits, QWORD = 64 bits. This creates a confusing terminology mismatch between the architectural naming convention and the physical register size.

Memory alignment means storing data at addresses that are multiples of the data's size. A 32-bit word should be stored at an address divisible by 4 (bytes); a 64-bit word at an address divisible by 8. Misaligned access is either forbidden (causes a CPU fault) or penalised (requires two memory reads instead of one). Compilers automatically align variables; manual struct packing can create misalignment that causes subtle performance issues or crashes on strict architectures.