Database Reference
In-Depth Information
Database systems are difficult to configure and maintain
. Administrative costs
can easily account for a significant fraction of the total cost of ownership of a
database system. Furthermore, it is extremely difficult for untrained professionals
to get good performance out of most commercial systems
Diversification in available systems complicates its selection
. The rise of spe-
cialized database systems for specific markets (e.g. main memory systems for
OLTP or column-stores for OLAP) complicates system selection, especially for
customers whose workloads do not neatly fall into one category.
Peak provisioning leads to unnecessary costs
. Web scale workloads are often
bursty in nature, and thus, provisioning for the peak often results in excess of
resources during off-peak phases, and thus unnecessary costs.
Recently, the new wave of NoSQL systems have started to gain some mindshares
as an alternative model for database management. In principle, some of the main
advantages of NoSQL systems can be summarized as follows:
Elastic Scaling
: For years, database administrators have relied on the
scale
up
approach rather than the
scale out
approach. However, with the current
increase in the transaction rates and high availability requirements, the economic
advantages of the scaling out approach on commodity hardware has become very
attractive. RDBMS might not scale out easily on commodity clusters but NoSQL
systems are initially designed with the ability to expand transparently in order to
take advantage of the addition of any new nodes.
Less Administration
: Despite the many manageability improvements introduced
by RDBMS vendors over the years, high-end RDBMS systems cannot be
maintained without the assistance of expensive, highly trained DBAs. DBAs
are intimately involved in the design, installation, and ongoing tuning of high-
end RDBMS systems. On the contrary, NoSQL databases are generally designed
from the ground up to require less management. For example, automatic repair
and the simpler data model features should lead to lower administration and
tuning requirements.
Better Economics
: While RDBMS tends to rely on expensive proprietary servers
and storage systems, NoSQL databases typically use clusters of cheap commod-
ity servers to manage the exploding data and transaction volumes. Therefore,
the cost per gigabyte or transactions per second for NoSQL can be many times
less than the cost for RDBMS which allows a NoSQL setup to store and process
more data at a much lower price. Moreover, when an application uses data that
is distributed across hundreds or even thousands of servers, simple economics
points to the benefit of using no-cost server software as opposed to that of paying
per-processor license fees. Once freed from license fees, an application can safely
scale horizontally with complete avoidance of the capital expenses.
Flexible Data Models
: Even minor changes to the data model of a large
production RDBMS have to be carefully managed and may necessitate downtime
or reduced service levels. NoSQL databases have more relaxed (if any) data
model restrictions. Therefore, application changes and database schema changes
can be changed more softly.
