Databases Reference
In-Depth Information
mirroring:
The duplication of data on multiple disks; a form of redundancy used to
speed up queries. It has the side effect of making updates slower because of the mul-
tiple copies.
monotonicity:
Functions are monotonic (or monotone) if the order of data after transfor-
mation has the same order as the data prior to transformation. A function
f
is
monotonic if whenever
x
<
y
, then
f
(
x
)
<
f
(
y
).
multicore CPU:
Combines two or more independent processors into an integrated cir-
cuit allowing the combined chip to provide the computing power of several CPUs.
Multicore CPUs offer thread-level parallelism (TLP). Each core in the chip has its
own level 1 memory cache, and often the higher level caches and the bus interface are
shared between the cores.
multidimensional clustering (MDC):
A technique by which a table can be created by
having key attributes in a related base table designated as dimensions, so data can be
clustered by dimensions in the database. Each dimension consists of one or more col-
umns (attributes), and the clusters are formed to take advantage of known and antic-
ipated workloads.
multidimensional OLAP (MOLAP):
An OLAP system that
uses multidimensional
arrays to store multidimensional data efficiently. An alternative to relational OLAP
systems that use fact tables to store data.
NAS:
Network-attached storage, a disk array connected to a controller that provides file
access services over a network.
nested-loop join:
An algorithm that contains a pair of nested for loops to search the
contents of two tables to be joined, matching one key value in a record in one table
with all the key values in records of the other table, then repeating the step for each
record in the first table. It has a complexity of n1 × n2 record accesses, where n1 and
n2 are the number of records (rows) in the two tables.
nonunique index (or secondary index):
An index having each entry consist of a key
value—pointer pair—but the key values are not unique and may have several entries,
one for each record that contains that value; an inverted file.
normalization:
The process of breaking up a table into smaller tables to eliminate prob-
lems with unwanted loss of data (the egregious side effects of losing data integrity)
from the deletion of records, and inefficiencies associated with multiple data updates.
OLAP:
Online analytical processing, a query service that overlays a data warehouse by
creating and maintaining a set of summary views (materialized query tables) to allow
for quick access to summary data. Related architectures include relational online ana-
lytical processing (ROLAP) and multidimensional online processing (MOLAP).