- Fact:
- A block is identified
by a block address
- A record is identified by a record address
(A record address is an extension of a block address, so in this discussion, I will often use: block/record address and handle both topics at the same time)
- A block is identified
by a block address
- Fact:
- A block/record can be read into the computer memory
- Identifying a block/record
when it is in memory:
- Use a virtual memory address (paging)
- There are 2 types of
address to
identify a
block/record in use:
- Database Address:
used to
identifies data (block or record) stored on
disk
There are 2 kinds of database addresses:
- Physical address
- Logical address
- Virtual memory address: used to identify data (block or record) stored in (virtual) memory
Graphically:
- Database Address:
used to
identifies data (block or record) stored on
disk
- Database Address:
- Database Address = an address used to identify data (block/record) stored on disk
- The are 2 common techniques to
identify a
block/record stored on
disk:
- Physical address
- Logical address
- Physical Address of a block:
- Physical Address of a block =
address used to
identify block
on a disk
- Components of a
block physical address:
- Host ID that contains the disk
- Disk ID the identifies the disk on the host
- Block (cylinder/track/section) number on the disk
- Physical Address of a block =
address used to
identify block
on a disk
- Physical Address of a
record:
- Physical Address of a record =
address used to
identify record
on a disk
- Components of a
record physical address:
- Physical address of a block
- Offset of the record in the block
- Physical Address of a record =
address used to
identify record
on a disk
- Logical block address:
- Each block/address is
assigned a unique
logical address
- Logical address =
an arbitrary string of
fixed length bits
(Can be generated automatically using some sequence generator or keep adding 1 to a counter)
- Each block/address is
assigned a unique
logical address
- DBMS uses a map table
to translate:
- Map table:
Logical Address Physical Address -------------------+------------------------------------ 101001.......10 | (HostID, DiskID, ..., BlockID) 010101.......01 | (HostID, DiskID, ..., BlockID) ...
- To speed up access,
the Map Table is organized
as:
- a hash table.
- The map table is stored on disk (in a well-known location)
- Map table:
- How to use the map table:
- How a
(physical/logical) database address
is used:
- When some record x need to
refer to
another record y,
we can store:
- the (physical/logical) database address of y inside x
- When some record x need to
refer to
another record y,
we can store:
- Intro:
Techniques to
referrence (point)
to a record:
- Primary key/foreign key:
- Each record x contains a
unique primary key
- Another record y that
references the
record x
will store the
primary key of
x
(The primary key of x stored in record y is called a foreign key)
- Each record x contains a
unique primary key
- The direct approach:
- The record y stores the physical/logical address of the record x
- Primary key/foreign key:
- Problem with
referencing another record
using a physical address:
- Example: 2 records references the
record Y
- Problem:
If record Y is
moved to a
different part of
the disk:
- We must update many addresses !!!
- Example: 2 records references the
record Y
- Solution using a
logical address as
reference:
- Example: 2 records references the
record Y
- If we move the
record Y, we
only need to:
- Update the physical address in the map table !!!
- Example: 2 records references the
record Y
- Fact:
- A block/record is stored on disk (this is not the problem)
- Caveat/Problem:
- A block/record must be stored in main memory in order to perform read/write operations on data in the block
- Main memory:
- Each data item stored in
(main) memory is
identified by its
virtual memory address
- While a
block/record is
in memory, the
DBMS must use
the
virtual address to
identify the
block/record:
- Each data item stored in
(main) memory is
identified by its
virtual memory address
- Consider the records
stored in a block:
Clearly:
- Address of a record = address of the block that contains the record + some offset information
- There are 2 common ways
to identify
a record:
- (DatabaseAddress(Block X)
+ the
(absolute) offset in block)
Example:
- record 1: (DatabaseAddress(Block X), 234 (offset in bytes))
- record 2: (DatabaseAddress(Block X), 456 (offset in bytes))
- (DatabaseAddress(Block X) +
the index
of the offset table
(in the block header)
Example:
- record 1: (DatabaseAddress(Block X), 0)
- record 2: (DatabaseAddress(Block X), 1)
- (DatabaseAddress(Block X)
+ the
(absolute) offset in block)
- Illustrative example:
2 block with records referening a
record Y:
- When the record Y is
deleted
:
the physical addresses will reference an incorrect record !!!
- Using logical addresses is
easy:
- Before deleting record Y:
After deleting record Y:
(The logical address used by record Y needs to be store in the map table !!!)
- Before deleting record Y:
- Deleting a record using
physical address:
use a tomb stone record
- Tomb stone record:
a (very small)
special purpose record
used to indicate a deleted record
- When a record is delete, it is replaved by the tomb stone record
Example:
- Before deleting record Y:
- After deleting record Y:
- Tomb stone record:
a (very small)
special purpose record
used to indicate a deleted record
- Insertion:
-
When you insert a
new record:
- You cannot use the space of a tomb stone record !!!
Because:
- Existing
record references
to the deleted record will
then
references to the
newly inserted trecord:
-
When you insert a
new record: