Discovering Hard Disk Physical Geometry through Microbenchmarking « Blog

Modern hard drives store an incredible amount of data in a small space, and are still the default choice for high-capacity (though not highest-performance) storage. While hard drives have been around since the 1950s, the current 3.5″ form factor (actually 4″ wide) appeared in the early 1980s. Since then, the capacity of a 3.5″ drive has increased by about 106 times (from 10 MB to about 10 TB), sequential throughput by about 103 times, and access times by about 101 times. Although the basic concept of spinning magnetic disks accessed by a movable stack of disk heads has not changed, hard drives have become much more complex to enable the increased density and performance. Early drives had tens of thousands of sectors arranged in hundreds of concentric tracks (sparse enough for a stepper motor to accurately position the disk head), while current drives have billions of sectors (with thousands of defects), packed into hundreds of thousands of tracks spaced tens of nanometers apart.Beyond just the high-level performance (throughput and seek time) measurements, which drive characteristics can be characterized using microbenchmarks? I had initially set out to detect the number of platters in a disk without opening up a disk, but in modern disks this simple-sounding task requires measuring several other properties before inferring the count of recording surfaces. Characterizing disk drive geometry has been done in the past [1, 2], and the algorithms I used aren’t very different. However, older algorithms often make assumptions that are no longer true on modern drives. For example, the Skippy [2] algorithm (a fast algorithm to measure the number of surfaces, cylinder switch times, and head switch times) no longer works on modern drives because the algorithm assumes one particular ordering of tracks onto multiple platters that is no longer used on modern disks (that several head switches occur before a seek to the next cylinder).Hard disk drives store data on a stack of one or more rotating magnetic disks. Data is written in concentric tracks. A stack of read/write heads move (radially) across the disks to position the head over the desired track. There are typically two heads per platter (one for each side), and the entire stack of heads move together as a single unit.

Reading data occurs by moving the disk head to the desired track (a seek), waiting until the beginning of the desired data passes under the disk head, and then continuing to read sequentially until either all of the requested data is read, or the end of the track, when the head needs to be moved to the next track. A hard drive’s “geometry” describes how data is arranged into platters, tracks, and sectors. Historically, this was described using three numbers: Cylinders (number of concentric rings from outside to inside), Heads (number of recording surfaces, or the number of tracks per cylinder), and Sectors per track, leading to the well-known acronym CHS. The capacity of a hard drive in sectors is simply C×H×S. Today, C and S are variable and only H is still constant. The number of tracks is not necessarily the same on each recording surface, and the number of sectors per track varies across the disk (more sectors in the longer outer tracks than the inner tracks).This article describes several microbenchmarks that try to extract the physical geometry of hard disk drives, and a few other related measurements. These measurements include rotation period, the physical location (angle and radius) of each sector, track boundaries, skew, seek time, and some observations of defective sectors. I use these microbenchmarks to characterize a variety of hard drives from 45 MB (1989) to 5 TB (2015). There is no attempt to characterize other important performance aspects such as caching. The remainder of this article begins with a background on hard drive geometry. It then describes the collection of microbenchmarks, starting with a basic read access time measurement and building towards increasingly complex algorithms. The second part of the article presents microbenchmark measurements for each of the 17 drives that were tested.Summary Background: Hard drives consist of spinning disks, and a stack of heads. Data is arranged on recording surfaces (2 sides per platter), tracks, and sectors. “Cylinders” no longer exist in drives newer than around 2000. What can be measured: RPM, angular position of every sector, and seek times, by timing specific sequences of sector reads. These basic measurement methods can then be used to find track boundaries, how tracks are arranged on a surface, and the number of surfaces.

Access time: Out of the drives I tested, it takes 1.3 to 3.6 revolutions to do a full-stroke seek. Heads accelerate slowly: Very few tracks are accessible in the first revolution. Short stroking offers limited reduction in seek times because even short seeks take a relatively long time. Seek time is non-trivial to measure. A seek time plot can be used to observe acoustic management (AAM). AAM slows down long-distance seeks to reduce noise, but not short-distance seeks. Track boundaries can be found by searching the disk for track skew. Newer disks use different densities (track size) on each surface. Track density and bit density can be estimated by knowing track count and size. In the newest drive I have, average track pitch is 80 nm and an average bit is 17 nm in length. Combining seek profile and track size together usually reveals the track layout. There is a large diversity of track layouts. Old drives had “cylinders” (several head switches occurs before a seek to the next track), but new drives use groups of adjacent tracks before changing heads. Track skew can be measured using track boundary information. There is more than one type of skew. A cylinder, serpentine, or zone change usually uses a bigger skew than for adjacent tracks. Track skew is usually constant from beginning to end of the drive, but not on the Maxtor 7405AV. Track skew is usually the same on every recording surface, but not on the Seagate ST1. Combining the above tools can find and visualize defective sectors. Most disks have holes of defective sectors, while some skip over entire tracks. Microbenchmarking is hard. Despite much effort, my algorithms do not work flawlessly. Measurement results for 17 hard drive models from 45 MB to 5 TB on Page 2. Background: Hard drive geometrySectors, tracks, and cylinders. A cylinder is the collection of all tracks at the same radius (6 tracks per cylinder shown here, on two sides and three platters).As seen by software, a hard drive looks like a big block of sectors, traditionally 512 bytes each (now 4,096 bytes), with little knowledge of the physical location of the sectors.

For example, a 300 GB drive might have 585,937,500 512-byte sectors, numbered 0 through 585,937,499. Some early hard drive interfaces required the drive controller on the host computer to know the physical layout of the disk (because the controller sent commands to move the disk head). IDE hard drives (Integrated Drive Electronics, mid 1980s) finally integrated the disk controller into the drive. The integrated disk controller translates logical (as seen by software) sector numbers into physical locations, which presents a simple block-of-sectors interface to the host computer while allowing much more complex physical layouts. Unfortunately, for software compatibility reasons, the sector number was still encoded as a CHS triplet (three numbers, but unrelated to the true number of cylinders, heads, or sectors/track of the drive) for many more years until LBA (logical block addressing, encoding a sector number as one number) became popular.While logically just a big block of sectors, sectors, tracks, and heads (or recording surfaces) still physically exist. There is just no easy way for software to know about them. This section gives some basic definitions for these physical features. Many readers may already be familiar with these.SectorData is stored in blocks of equal size. A sector is the smallest unit of data that can be read or written to a disk. 512-byte sectors have been standard since the 1980s, while new drives (around 2011) use 4096-byte sectors (branded as Advanced Format). Sectors have additional metadata that are also written to the disk surface (such as its sector number and error correction codes). On drives using embedded servo (all non-ancient drives), there are also servo patterns on the disk which are used to position the disk head. All of this occupies space on the disk surface but is invisible to the host.TracksA track is a circle of consecutive sectors placed on one disk surface along one revolution of the disk. Reading sectors within a track is done by having the head follow the track while the disk rotates. Crossing a track boundary requires moving the disk head to the next track (a track-to-track seek) or switching to a different head to read a track from a different disk surface (a head switch).

Prior to zone bit recording, every track on a disk was the same size (number of sectors). Zone bit recording packs more sectors into physically-longer outside tracks and fewer in the shorter inner tracks. Because one track (regardless of length) is read per revolution, hard drives have higher throughput near the beginning of the drive. Track size can also vary between recording surfaces, because the recording surface and head quality varies even within a single drive.Hard drives (like floppy disks) use concentric circular tracks, unlike CDs and DVDs which use a single spiral track.CylindersA cylinder is a collection of tracks on multiple surfaces that are located at the same radius (If a track is a circle, then a stack of circles of the same diameter forms a cylinder). On older drives, tracks on different surfaces were aligned so that accessing tracks within the same cylinder only required switching heads (a faster electrical operation) but not moving the heads (a slower mechanical operation). Cylinders are no longer meaningful on modern drives. With increased track density, tracks on different recording surfaces aren’t aligned well enough to form cylinders and a head switch requires a larger head movement than moving to an adjacent track on the same surface, which makes a head switch slower than moving to an adjacent track.ZonesDisk throughput benchmark. About 20 zones are visible. Outer tracks have higher throughput than inner tracks, and throughput is constant within a zone.Because the outer tracks are physically longer, more sectors can be packed into outer tracks than the inner tracks. For simplicity, adjacent tracks are grouped into zones where every track within a zone has the same track size (number of sectors per track). Zone bit recording gives the familiar throughput curve where outer tracks have higher throughput than inner tracks, with throughput decreasing in discrete steps.Traditionally, zones were thought of as groups of cylinders, where all tracks on all surfaces in the zone had the same size. Because track size on modern drives can differ between recording surfaces, it only makes sense to think of a zone as a group of adjacent tracks on the same recording surface on modern drives.Track SkewLeft: No skew. Every track starts at the same angular position. There is one wasted revolution after every track because it takes non-zero time to move to the next track. Right: Skew of 72° (1/5 revolution).