Database Reference
In-Depth Information
that play well with Cassandra. These include Helenus (the Node.js driver), Priam
(a Cassandra automation tool), and Hector (the Java driver), just to name a few.
What Is NoSQL?
There is no single definition for NoSQL. To some it stands for “Not Only SQL”;
to others it means “No SQL.” Either way, it refers to the departure from the tradi-
tional relational database technologies that have dominated the development land-
scape for the past few decades.
What is likely the largest driver of the NoSQL movement is a commonly held
belief that relational databases are not well suited to large amounts of data and
scale. Whether or not this is true, the emergence of the key/value, graph, docu-
ment, and “big table” data storage engines shows that a new generation of data-
base technologies is taking center stage.
There is no single database technology that is synonymous with the NoSQL
movement. Branding and marketing seem to be mostly what determine how relev-
ant a technology is to the terminology.
There's No Such Thing as “Web Scale”
Another marketing term that gets thrown around quite frequently is “Web
scale.” It is used quite often when discussing how to determine whether a database
system is suitable for a particular Web application's needs and whether it will hold
up as the application grows. This is a very subjective term as everyone's needs
are different. A simple SQL setup will achieve most scalability needs. Depending
on the read/write patterns of an application, one may need a specialized database,
such as Kyoto Cabinet (previously named Tokyo Cabinet) for key/value or Mon-
goDB as a document store. In a system that needs high write throughput and linear
scalability, Cassandra is a great fit and will hold up under some very heavy work-
loads.
The key point to remember when discussing the idea of Web scale technologies
is that nearly everything out there will scale with enough money, hardware, and
headaches. The trick is to figure out which piece of software is best suited for your
usage patterns and workloads and will scale out in a way suitable for your applic-
ation and your organization.
