Databases Reference
In-Depth Information
memory buffers. Data can be read from the memory buffers for data extraction
or processing as needed. In a database environment, you must move large vol-
umes of data through the memory buffers. The data remains in memory until
the space is needed and it is overwritten by new data coming into the buffers.
Because of this, when a request is made for data, the memory buffers are searched
first to see whether the requested data is already there. If so, then the data can
be accessed from memory and used much faster than when being read from the
hard disk.
Computer systems use I/O (input/output) mechanisms to move data into and
out of the computer's main memory from
secondary storage,
a term used to
refer to
nonvolatile storage
media that does not lose the data it contains if the
computer is turned off, such as hard disks. Modern PCs improve transfer per-
formance through
direct memory access (DMA)
channels that let devices such
hard disk drives, floppy drives, video adapters, and network adapters transfer
data directly to and from memory without going through the processor cache.
Another I/O concern is communication between computers. In earlier data-
base environments, most of the processing was done locally around a central-
ized database. Today's database environments are different. Distributed databases
have become common. Data communications components move high volumes
of data between distributed databases at different locations. Communication links
with sufficient capacity and speed to handle the data movement are critical.
These communications links include both local communications through a local
network, and remote communications through a wide-area or enterprise net-
work. The Internet is used extensively as the communication path of choice for
remote communications.
Data storage devices such as disk, tape, and CD or DVD drives are critical.
Without nonvolatile data storage on secondary storage devices, there is no data-
base. How you use secondary storage and how you are able to retrieve data from
it affects the performance of the database environment. Often, storage device per-
formance is the most limiting factor when optimizing database server performance.
Two primary concerns are capacity and speed. You were introduced earlier to
the data repository as containing all of the database data. Your secondary storage
must have sufficient capacity to hold all of this data along with program files,
operating system files, and any other data stored locally on the computer. It must
also provide for high-speed data access so that data requests are processed quickly
and efficiently. Secondly, data storage should be fault tolerant. Data storage mech-
anisms should ensure that the database operations continue even when some mal-
function affects parts of the storage. RAID technology devices are a common solu-
tion and allow you to store data so that even if one disk fails, the database
operations continue.
RAID,
which stands for redundant array of inexpensive (or
independent, depending on who you ask) disks, provides high-performance disk
storage. Most RAID configurations also provide fault-tolerant storage.
