Transcript Storage - CS-People by full name

Storage and Disks
Now Something Different
1st part of the course: Application Oriented
2nd part of the course: Systems Oriented
What is “Systems”?
A: Not Programming
Not programming big things..
Systems = Efficient and safe use of limited resources (e.g., disks)
Efficient: resources should be shared, utilized as much as possible
Safe:
Database System Concepts
sharing should not corrupt work of individual jobs
11.2
General Overview
 Relational model - SQL
 Formal & commercial query languages
 Functional Dependencies
 Normalization
 Txn Processing, & CC
 Physical Design
 Indexing
Application
Oriented
 Query evaluation
 Query optimization
 ….
Database System Concepts
Systems Oriented
11.3
The systems side of Databases
What will we talk about?
1. Data Organization: physical storage strategies to support
efficient updates, retrieval (Ch. 11)
2. Data retrieval: auxiliary data structures to enable efficient retrieval
(Ch. 12). Techniques for processing queries to ensure efficient
retrieval (Ch. 13-14)
3. Data Integrity: techniques for implementing Xactions, to ensure
safe concurrent access to data (Ch. 16). Ensuring data is safe in
the presence of system crashes (Ch. 17)
Database System Concepts
11.4
Data Organization
Key points
1. Storage Media
 “Memory hierarchy”
 Efficient/reliable transfer of data between disks and main
memory
 Hardware techniques (RAID disks)
 Software techniques (Buffer mgmt)
2. Storage strategies for relations-file organization
 Representation of tuples on disks
 Storage of tuples in pages, clustering.
Database System Concepts
11.5
CPU
...
M
Typical
Computer
C
...
Secondary
Storage
Database System Concepts
11.6
Storage Media: Players
 Cache – fastest and most costly form of storage; volatile;
managed by the computer system hardware.
 Main memory:
 fast access (10s to 100s of nanoseconds; 1 nanosecond =
10–9 seconds)
 generally too small (or too expensive) to store the entire
database
 Volatile — contents of main memory are usually lost if a
power failure or system crash occurs.
 But… CPU operates only on data in main memory
Database System Concepts
11.7
Storage Media: Players
 Disk
 Primary medium for the long-term storage of data; typically
stores entire database.
 random-access – possible to read data on disk in any order,
unlike magnetic tape
 Non-volatile: data survive a power failure or a system crash,
disk failure less likely than them
Database System Concepts
11.8
Storage Media: Players
 Optical storage
 non-volatile, data is read optically from a spinning disk using a laser
 CD-ROM (640 MB) and DVD (4.7 to 17 GB) most popular forms
 Write-one, read-many (WORM) optical disks used for archival
storage (CD-R and DVD-R)
 Multiple write versions also available (CD-RW, DVD-RW, and DVDRAM)
 Reads and writes are slower than with magnetic disk
 Tapes
 Sequential access (very slow)
 Cheap, high capacity
Database System Concepts
11.9
Memory Hierarchy
cache
Main memory
V
NV
disk
Optical storage
Tapes
Traveling the hierarchy:
1. speed ( higher=faster)
2. cost (lower=cheaper)
3. volatility (between MM and Disk)
4. Data transfer (Main memory the “hub”)
5. Storage classes (P=primary, S=secondary,
T=tertiary)
Database System Concepts
11.10
Memory Hierarchy
 Data transfers
 cache – mm : OS/hardware controlled
 mm – disk : <- reads, -> writes controlled by DBMS
 disk – CD-Rom or DVD
 disk – Tapes
Backups (off-line)
Database System Concepts
11.11
Main memory  Disk Data Xfers
Concerns:
1. Efficiency (speed)
can be improved by...
a. improving raw data transfer speed
b. avoiding untimely data transfer
c. avoiding unnecessary data transfer
2. Safety (reliability, availability)
can be improved by...
a. storing data redundantly
Database System Concepts
11.12
Main memory  Disk Data Xfers
Achieving efficiency:
1. Improve Raw data Xfer speed
1. Faster Disks
2. Parallelization (RAID)
2. Avoiding untimely data xfers
1. Disk scheduling
2. Batching
3. Avoiding unnecessary data xfers
1. Buffer Management
2. Good file organization
Database System Concepts
11.13
Hard Disk Mechanism
Database System Concepts
11.14

Read-write head
 Positioned very close to the platter
surface (almost touching it)


Surface of platter divided into circular
tracks
Each track is divided into sectors.
 A sector is the smallest unit of data that
can be read or written.

To read/write a sector
 disk arm swings to position head on
right track
 platter spins continually; data is
read/written as sector passes under
head


Block: a sequence of sectors
Cylinder i consists of ith track of all the
platters
Database System Concepts
11.15
“Typical” Values
Diameter:
Cylinders:
Surfaces:
1 inch  15 inches
100  2000
1 or 2 per platter
(Tracks/cyl)
2 (floppies)  30
Sector Size: 512B  50K
Capacity:
360 KB (old floppy)
 300 GB
Database System Concepts
11.16
Performance Measures of Disks
Measuring Disk Speed
 Access time – consists of:
 Seek time – time it takes to reposition the arm over the correct track.
 (Rotational) latency time – time it takes for the sector to be accessed
to appear under the head.
 Data-transfer rate – the rate at which data can be retrieved from or
stored to the disk.
Analogy to taking a bus:
1. Seek time: time to get to bus stop
2. Latency time; time spent waiting at bus stop
3. Data transfer time: time spent riding the bus
Database System Concepts
11.17
Example
ST3120022A : Barracuda 7200.7
Capacity:120 GB
Interface: Ultra ATA/100
RPM: 7200 RPM
Seek time: 8.5 ms avg
Latency time?:
7200/60 = 120 rotations/sec
1 rotation in 8.3 ms => So, Av. Latency = 4.16 ms
Database System Concepts
11.18
Random vs sequential i/o
 Ex: 1 KB Block
 Random I/O:
 15 ms.
 Sequential I/O:  1 ms.
Rule of
Thumb
Database System Concepts
Random I/O: Expensive
Sequential I/O: Much less ~10-20 times
11.19
Performance Measures (Cont.)
 Mean time to failure (MTTF) – the average time the disk is
expected to run continuously without any failure.
 Typically 5 to 10 years
 Probability of failure of new disks is quite low, corresponding to a
“theoretical MTTF” of 30,000 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
Database System Concepts
11.20
RAID
RAID: Redundant Arrays of Independent (Inexpensive) Disks
 disk organization techniques that manage a large numbers of disks,
providing a view of a single disk
 Idea: cheaper to have many small disks, than few big disks
 bonus: also advantageous for:
1. speed (efficiency)
2. reliability (safety)
Database System Concepts
11.21
Improvement in Performance via Parallelism
Choices:
D1
D2
D3
....
1. Distribute files (f1  D1, f2  D2, ....)
or
2. Distribute parts of files (“striping”)
 block striping
 sector striping
......
 bit striping
Database System Concepts
11.22
Dn
Parallelization
File distribution
+: Availability: Many files still available if a disk goes down
recovery requires fewer disks
- : but still sequential read for each file
Striping
+: improved ||’ism (speed)
( - : but a single disk failure catastrophic!)
Database System Concepts
11.23
Improving Reliability
Reliability:
• Measure: MTTF
•Striping reduces reliability: why?
Solution = Redundancy
Redundancy: store data on more than 1 disk
E.g. “mirroring” (duplicate disks) (1 disk stored on 2)
logical disk
Then, MTTF for both disks: 57,000 yrs! assuming MTTF for
each disk is 11 yrs.
Database System Concepts
11.24
RAID Levels
 Schemes to provide redundancy at lower cost by using disk
striping combined with parity bits
 Different RAID organizations, or RAID levels, have differing cost,
performance and reliability characteristics
RAID Level 0: Block striping; non-redundant.
 Used in high-performance applications where data lost 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
11.25
RAID Levels (Cont.)
 RAID Level 2: Memory-Style Error-Correcting-Codes (ECC) with bit
striping.
 RAID Level 3: Bit-Interleaved Parity

a single parity bit is enough for error correction, not just detection, since
we know which disk has failed
 When writing data, corresponding parity bits must also be computed and
written to a parity bit disk
 To recover data in a damaged disk, compute XOR of bits from other disks
(including parity bit disk)
Database System Concepts
11.26
RAID Levels (Cont.)
 RAID Level 3 (Cont.)
 Faster data transfer than with a single disk, but fewer I/Os per
second since every disk has to participate in every I/O.
 Subsumes Level 2 (provides all its benefits, at lower cost).
 RAID Level 4: Block-Interleaved Parity; uses block-level
striping, and keeps a parity block on a separate disk for
corresponding blocks from N other disks.
 When writing data block, corresponding block of parity bits must
also be computed and written to parity disk
 To find value of a damaged block, compute XOR of bits from
corresponding blocks (including parity block) from other disks.
Database System Concepts
11.27
RAID Levels (Cont.)
 RAID Level 4 (Cont.)
 Provides higher I/O rates for independent block reads than Level 3
 Provides high transfer rates for reads of multiple blocks than no-striping
 Before writing a block, parity data must be computed
 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)
 Parity block becomes a bottleneck for independent block writes since
every block write also writes to parity disk
Database System Concepts
11.28
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
11.29
RAID Levels (Cont.)
 RAID Level 5 (Cont.)
 Higher I/O rates than Level 4.
 Block writes occur in parallel if the blocks and their parity blocks are on
different disks.
 Subsumes Level 4: provides same benefits, but avoids bottleneck of
parity disk.
 RAID Level 6: P+Q Redundancy scheme; similar to Level 5, but
stores extra redundant information to guard against multiple disk
failures.

Better reliability than Level 5 at a higher cost; not used as widely.
Database System Concepts
11.30
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




Level 2 and 4 never used since they are subsumed by 3 and 5
Level 3 is not used anymore since bit-striping forces single block reads
to access all disks, wasting disk arm movement, which block striping
(level 5) avoids
Level 6 is rarely used since levels 1 and 5 offer adequate safety for
almost all applications
So competition is between 1 and 5 only
 Full mirroring (1) and block-interleaved distributed parity (5)
Database System Concepts
11.31