Information Technology Reference
In-Depth Information
Somespecialcasesofpartitioningexist.Packetlossisconsideredatemporarypartition-
ing of the system as it applies to the CAP Principle. Another special case is the complete
network outage. Even the most partition-tolerant system is unable to work in that situation.
The CAP Principle says that any one or two of the attributes are achievable in combin-
ation, but not all three. In 2002, Gilbert and Lynch published a formal proof of the original
conjecture, rendering it a theorem. One can think of this as the third attribute being sacri-
ficed to achieve the other two.
TheCAPPrincipleisillustratedbythetrianglein Figure1.9 . Traditionalrelationaldata-
bases like Oracle, MySQL, and PostgreSQL are consistent and available (CA). They use
transactions and other database techniques to assure that updates are atomic; they propag-
atecompletelyornotatall.Thustheyguaranteealluserswillseethesamestateatthesame
time. Newer storage systems such as Hbase, Redis, and Bigtable focus on consistency and
partition tolerance (CP). When partitioned, they become read-only or refuse to respond to
any requests rather than be inconsistent and permit some users to see old data while others
see fresh data. Finally, systems such as Cassandra, Risk, and Dynamo focus on availability
and partition tolerance (AP). They emphasize always being able to serve requests even if it
means some clients receive outdated results. Such systems are often used in globally dis-
tributed networks where each replica talks to the others by less reliable media such as the
Internet.
Search WWH ::




Custom Search