Chapter 12: Physical Storage Systems

Database System Concepts, 7th Ed.

©Silberschatz, Korth and Sudarshan
See www.db-book.com for conditions on re-use
 Storage Hierarchy

Primary Storage

Volatile Storage

Non Volatile Storage

Secondary Storage

Tertiary Storage

Database System Concepts - 7th Edition 12.3 ©Silberschatz, Korth and Sudarshan
 Storage Hierarchy: Access Time

1-5 nanoseconds

60-100 nanoseconds
cache-line-at-a-time

20-100 microseconds
page-at-a-time

5-10 milliseconds
page-at-a-time

100+ milliseconds

10-100 seconds or more

Database System Concepts - 7th Edition 12.4 ©Silberschatz, Korth and Sudarshan
 Storage Interfaces
§ Disk interface standards families
• SATA (Serial ATA)
§ SATA 3 supports data transfer speeds of up to 6 gigabits/sec
• SAS (Serial Attached SCSI)
§ SAS Version 3 supports 12 gigabits/sec
• NVMe (Non-Volatile Memory Express) interface
§ Works with PCIe connectors to support lower latency and
higher transfer rates
§ Supports data transfer rates of up to 24 gigabits/sec
§ Disks usually connected directly to computer system
§ In Storage Area Networks (SAN), a large number of disks are
connected by a high-speed network to a number of servers
§ In Network Attached Storage (NAS) networked storage provides a
file system interface using networked file system protocol, instead of
providing a disk system interface

Database System Concepts - 7th Edition 12.6 ©Silberschatz, Korth and Sudarshan
 Magnetic Hard Disk Mechanism

Schematic diagram of magnetic disk drive Photo of magnetic disk drive

Database System Concepts - 7th Edition 12.7 ©Silberschatz, Korth and Sudarshan
 Magnetic Disks (Cont.)
§ Disk controller – interfaces between the computer system and
the disk drive hardware.
• accepts high-level commands to read or write a sector
• initiates actions such as moving the disk arm to the right track and
actually reading or writing the data
• Computes and attaches checksums to each sector to verify that
data is read back correctly
§ If data is corrupted, with very high probability stored checksum
won’t match recomputed checksum
• Ensures successful writing by reading back sector after writing it
• Performs remapping of bad sectors

Database System Concepts - 7th Edition 12.9 ©Silberschatz, Korth and Sudarshan
 Performance Measures of Disks
§ Access time – the time it takes from when a read or write request
is issued to when data transfer begins. Consists of:
• Seek time – time it takes to reposition the arm over the correct track.
§ Average seek time is 1/2 the worst case seek time.
• Would be 1/3 if all tracks had the same number of sectors, and we
ignore the time to start and stop arm movement
§ 4 to 10 milliseconds on typical disks
• Rotational latency – time it takes for the sector to be accessed to appear
under the head.
§ 4 to 11 milliseconds on typical disks (5400 to 15000 r.p.m.)
§ Average latency is 1/2 of the above latency.
• Overall latency is 5 to 20 msec depending on disk model
§ Data-transfer rate – the rate at which data can be retrieved from
or stored to the disk.
• 25 to 200 MB per second max rate, lower for inner tracks

Database System Concepts - 7th Edition 12.10 ©Silberschatz, Korth and Sudarshan
 Performance Measures (Cont.)
§ Disk block is a logical unit for storage allocation and retrieval
• 4 to 16 kilobytes typically
§ Smaller blocks: more transfers from disk
§ Larger blocks: more space wasted due to partially filled blocks

§ Sequential access pattern

• Successive requests are for successive disk blocks
• Disk seek required only for first block
§ Random access pattern
• Successive requests are for blocks that can be anywhere on disk
• Each access requires a seek
• Transfer rates are low since a lot of time is wasted in seeks
§ I/O operations per second (IOPS)
• Number of random block reads that a disk can support per second
• 50 to 200 IOPS on current generation magnetic disks

Database System Concepts - 7th Edition 12.11 ©Silberschatz, Korth and Sudarshan
 Performance Measures (Cont.)
§ Mean time to failure (MTTF) – the average time the disk is
expected to run continuously without any failure.
• Typically 3 to 5 years
• Probability of failure of new disks is quite low, corresponding to a
“theoretical MTTF” of 500,000 to 1,200,000 hours for a new disk
§ E.g., an MTTF of 1,200,000 hours for a new disk means that
given 1000 relatively new disks, on an average one will fail
every 1200 hours
• MTTF decreases as disk ages
§ Annualized Failure Rate (AFR): =( (365*24) / MTTF)*100%
• MTTF=1,200,000 è AFR = 0.73%
§ Suppose MTTF is 1,200,000 hours for a disk. Then, in a
system with 1000 disks, how often will a disk fail on average?
• Answer: on average one will fail every 1200 hours (50 days)
§ Equivalently, 7.3 disks per year

Database System Concepts - 7th Edition 12.12 ©Silberschatz, Korth and Sudarshan
 Flash Storage
§ NOR flash vs NAND flash
§ NAND flash
• used widely for storage, cheaper than NOR flash
• requires page-at-a-time read (page: 512 bytes to 4 KB)
§ 20 to 100 microseconds for a page read
§ Not much difference between sequential and random
read
• Page can only be written once
§ Must be erased to allow rewrite
§ Solid state disks
• Use standard block-oriented disk interfaces, but store data on
multiple flash storage devices internally
• Transfer rate of up to 500 MB/sec using SATA, and
up to 3 GB/sec using NVMe PCIe

Database System Concepts - 7th Edition 12.13 ©Silberschatz, Korth and Sudarshan
 Flash Storage (Cont.)
§ Erase happens in units of erase block
• Takes 2 to 5millisecs
• Erase block typically 256 KB to 1 MB (128 to 256 pages)
§ Remapping of logical page addresses to physical page addresses
avoids waiting for erase
§ Flash translation table tracks mapping
• also stored in a label field of flash page
• remapping carried out by flash translation layer
Page write
Physical Page Physical Page
Logical Page
Address Structure
Address
Logical address
Valid bit
Page Data

Flash Translation
Table

Database System Concepts - 7th Edition 12.14 ©Silberschatz, Korth and Sudarshan
 Flash Storage (Cont.)
§ SLC After about 1,00,000 erases (SLC Flash) to as low as 10,000 or
1000 erases (TLC/QLC Flash) erase block becomes unreliable and
cannot be used
• wear leveling: store infrequently updated (“cold”) data in blocks that
have been erased many times already

Database System Concepts - 7th Edition 12.15

Source: Kingston.com
©Silberschatz, Korth and Sudarshan
 SSD Performance Metrics
§ Random reads/writes per second
• Typical 4 KB reads: 10,000 reads per second (10,000 IOPS)
• Typical 4KB writes: 40,000 IOPS
• SSDs support parallel reads
§ Typical 4KB reads:
• 100,000 IOPS with 32 requests in parallel (QD-32) on
SATA
• 350,000 IOPS with QD-32 on NVMe PCIe
§ Typical 4KB writes:
• 100,000 IOPS with QD-32, even higher on some models
§ Data transfer rate for sequential reads/writes
• 400 MB/sec for SATA3, 2 to 3 GB/sec using NVMe PCIe
§ Hybrid disks: combine small amount of flash cache with larger
magnetic disk

Database System Concepts - 7th Edition 12.16 ©Silberschatz, Korth and Sudarshan
 Storage Class Memory
§ 3D-XPoint memory technology pioneered by Intel
§ Available as Intel Optane
• SSD interface shipped from 2017
§ Allows lower latency than flash SSDs
• Non-volatile memory interface announced in 2018
§ Supports direct access to words, at speeds comparable to
main-memory speeds

Database System Concepts - 7th Edition 12.17 ©Silberschatz, Korth and Sudarshan
 RAID

§ RAID: Redundant Arrays of Independent Disks

• disk organization techniques that manage a large numbers of disks,
providing a view of a single disk of
§ high capacity and high speed by using multiple disks in parallel,
§ high reliability by storing data redundantly, so that data can be
recovered even if a disk fails
§ The chance that some disk out of a set of N disks will fail is much higher
than the chance that a specific single disk will fail.
• E.g., a system with 100 disks, each with MTTF of 100,000 hours (approx.
11 years), will have a system MTTF of 1000 hours (approx. 41 days)
• Techniques for using redundancy to avoid data loss are critical with large
numbers of disks

Database System Concepts - 7th Edition 12.18 ©Silberschatz, Korth and Sudarshan
 Improvement of Reliability via Redundancy

§ Redundancy – store extra information that can be used to rebuild

information lost in a disk failure
§ E.g., Mirroring (or shadowing)
• Duplicate every disk. Logical disk consists of two physical disks.
• Every write is carried out on both disks
§ Reads can take place from either disk
§ Mean time to data loss depends on mean time to failure,
and mean time to repair
• E.g. MTTF of 100,000 hours, mean time to repair of 10 hours
gives mean time to data loss of 500*106 hours (or 57,000 years)
for a mirrored pair of disks (ignoring dependent failure modes)

Database System Concepts - 7th Edition 12.20 ©Silberschatz, Korth and Sudarshan
 Improvement in Performance via Parallelism

§ Goals of parallelism in a disk system:

1. Load balance multiple small accesses to increase throughput
2. Parallelize large accesses to reduce response time.
3. Improve transfer rate by striping data across multiple disks.
§ Bit-level striping
• Not used in practice
§ Block-level striping – with n disks, block i of a file goes to disk (i
mod n) + 1
• Requests for different blocks can run in parallel if the blocks reside on
different disks
• A request for a long sequence of blocks can utilize all disks in parallel

Database System Concepts - 7th Edition 12.22 ©Silberschatz, Korth and Sudarshan
 RAID Levels
§ RAID Level 0: Block striping; non-redundant.
• Used in high-performance applications where data loss is not critical.
§ RAID Level 1: Mirrored disks with block striping
• Offers best write performance.
• Popular for applications such as storing log files in a database system.

Database System Concepts - 7th Edition 12.23 ©Silberschatz, Korth and Sudarshan
 RAID Levels (Cont.)
§ Parity blocks: Parity block j stores XOR of bits from block j of each
disk
• When writing data to a block j, parity block j must also be computed
and written to disk
§ Can be done by using old parity block, old value of current block
and new value of current block (2 block reads + 2 block writes)
§ Or by recomputing the parity value using the new values of blocks
corresponding to the parity block
• More efficient for writing large amounts of data sequentially
• To recover data for a block, compute XOR of bits from all other blocks
in the set including the parity block

Database System Concepts - 7th Edition 12.24 ©Silberschatz, Korth and Sudarshan
 RAID Levels (Cont.)
§ RAID Level 5: Block-Interleaved Distributed Parity; partitions data
and parity among all N + 1 disks, rather than storing data in N disks
and parity in 1 disk.
• E.g., with 5 disks, parity block for nth set of blocks is stored on disk
(n mod 5) + 1, with the data blocks stored on the other 4 disks.

Database System Concepts - 7th Edition 12.25 ©Silberschatz, Korth and Sudarshan
 RAID Levels (Cont.)
§ RAID Level 5 (Cont.)
• Block writes occur in parallel if the blocks and their parity blocks
are on different disks.
§ RAID Level 6: P+Q Redundancy scheme; similar to Level 5, but
stores two error correction blocks (P, Q) instead of single parity
block to guard against multiple disk failures.
• Better reliability than Level 5 at a higher cost
§ Becoming more important as storage sizes increase

Database System Concepts - 7th Edition 12.26 ©Silberschatz, Korth and Sudarshan
 RAID Levels (Cont.)
§ Other levels (not used in practice):
• RAID Level 2: Memory-Style Error-Correcting-Codes (ECC)
with bit striping.
• RAID Level 3: Bit-Interleaved Parity
• RAID Level 4: Block-Interleaved Parity; uses block-level
striping, and keeps a parity block on a separate parity disk for
corresponding blocks from N other disks.
§ RAID 5 is better than RAID 4, since with RAID 4 with random
writes, parity disk gets much higher write load than other
disks and becomes a bottleneck

Database System Concepts - 7th Edition 12.27 ©Silberschatz, Korth and Sudarshan
 Choice of RAID Level
§ Factors in choosing RAID level
• Monetary cost
• Performance: Number of I/O operations per second, and
bandwidth during normal operation
• Performance during failure
• Performance during rebuild of failed disk
§ Including time taken to rebuild failed disk
§ RAID 0 is used only when data safety is not important
• E.g. data can be recovered quickly from other sources

Database System Concepts - 7th Edition 12.28 ©Silberschatz, Korth and Sudarshan
 Choice of RAID Level (Cont.)

§ Level 1 provides much better write performance than level 5

• Level 5 requires at least 2 block reads and 2 block writes to write
a single block, whereas Level 1 only requires 2 block writes
§ Level 1 had higher storage cost than level 5
§ Level 5 is preferred for applications where writes are sequential
and large (many blocks), and need large amounts of data storage
§ RAID 1 is preferred for applications with many random/small
updates
§ Level 6 gives better data protection than RAID 5 since it can
tolerate two disk (or disk block) failures
• Increasing in importance since latent block failures on one disk,
coupled with a failure of another disk can result in data loss with
RAID 1 and RAID 5.

Database System Concepts - 7th Edition 12.29 ©Silberschatz, Korth and Sudarshan
 Hardware Issues

§ Software RAID: RAID implementations done entirely in

software, with no special hardware support
§ Hardware RAID: RAID implementations with special hardware
• Use non-volatile RAM to record writes that are being executed
• Beware: power failure during write can result in corrupted disk
§ E.g. failure after writing one block but before writing the
second in a mirrored system
§ Such corrupted data must be detected when power is
restored
• Full scan of disk may be required!
• NV-RAM helps to efficiently detected potentially
corrupted blocks

Database System Concepts - 7th Edition 12.30 ©Silberschatz, Korth and Sudarshan
 Hardware Issues (Cont.)
§ Latent sector failures: data successfully written earlier gets damaged
• can result in data loss even if only one disk fails
§ Data scrubbing:
• continually scan for latent failures, and recover from copy/parity
§ Hot swapping: replacement of disk while system is running, without
power down
• Supported by some hardware RAID systems,
• reduces time to recovery, and improves availability greatly
§ Spare disks are kept online, and used as replacements for failed disks
immediately on detection of failure
• Reduces time to recovery greatly
§ To avoid single point of failure
• Redundant power supplies with UPS backup
• Multiple network controllers/network interconnections

Database System Concepts - 7th Edition 12.31 ©Silberschatz, Korth and Sudarshan
 Optimization of Disk-Block Access

§ Buffering: in-memory buffer to cache disk blocks

§ Read-ahead: Read extra blocks from a track in anticipation that
they will be requested soon
§ Disk-arm-scheduling algorithms re-order block requests so that
disk arm movement is minimized
• elevator algorithm
R6 R3 R1 R5 R2 R4

Inner track Outer track

Database System Concepts - 7th Edition 12.32 ©Silberschatz, Korth and Sudarshan
End of Chapter 12

Database System Concepts, 7th Ed.

©Silberschatz, Korth and Sudarshan
See www.db-book.com for conditions on re-use
 Magnetic Tapes

§ Hold large volumes of data and provide high transfer rates

• Few GB for DAT (Digital Audio Tape) format, 10-40 GB with DLT
(Digital Linear Tape) format, 100 GB+ with Ultrium format, and
330 GB with Ampex helical scan format
• Transfer rates from few to 10s of MB/s
§ Tapes are cheap, but cost of drives is very high
§ Very slow access time in comparison to magnetic and optical
disks
• limited to sequential access.
• Some formats (Accelis) provide faster seek (10s of seconds) at
cost of lower capacity
§ Used mainly for backup, for storage of infrequently used
information, and as an off-line medium for transferring information
from one system to another.
§ Tape jukeboxes used for very large capacity storage
• Multiple petabyes (1015 bytes)
Database System Concepts - 7th Edition 12.35 ©Silberschatz, Korth and Sudarshan
 Figure 10.03
(a) RAID 0: nonredundant striping

C C C C

(b) RAID 1: mirrored disks

P P P

(c) RAID 2: memory-style error-correcting codes

(d) RAID 3: bit-interleaved parity

(e) RAID 4: block-interleaved parity

P P P P P

(f) RAID 5: block-interleaved distributed parity

P P P P
P P

(g) RAID 6: P + Q redundancy

Database System Concepts - 7th Edition 12.36 ©Silberschatz, Korth and Sudarshan
Database System Concepts - 7th Edition 12.37 ©Silberschatz, Korth and Sudarshan