(and What Makes It So Awesome?)
You can learn about why Kite is awesome by watching this Kite Overview video.
Kite is an open source set of libraries, references, tutorials, and code samples for building data-oriented systems and applications.
CDH and Cloudera Manager provide a data platform and data management platform, but they aren’t enough. You need higher-level tools, APIs, documentation, and examples to get working quickly.
Generally, everyone agrees that Hadoop is an incredibly powerful platform. It provides the infrastructure needed to process and store data at scale.
Hadoop is also incredibly flexible. The ecosystem includes tools for ETL, machine learning, analysis, data science, and all the services that comprise an enterprise data hub. Other platforms are not as flexible.
However, what makes Hadoop flexible is its very basic APIs, operating at a very low level. Hadoop requires you to build a lot of plumbing and infrastructure to create the most basic system. You often have to repeat the process for every system you create, with an understanding of some of the quirks and challenges of integrating the individual components to meet your specific business needs.
Hadoop is daunting from the perspective of a developer who just needs to get some work done. You should be able to ramp up quickly on a few key technologies or concepts, rather than first learn everything about the 13-14 different technologies required to use Hadoop effectively
Things should just work together. Hadoop forces you to spend more time thinking about infrastructure than your business goals.
Hadoop is not difficult to use. The complexity comes from the many parts that comprise a very large animal. Each piece, in isolation, is straightforward and easy to understand.
This is where Kite comes in. Kite provides additional support for this infrastructure one level up in the stack so that they are codified in APIs that make sense to developers.
Kite is immediately accessible to developers who speak Java. Other languages will follow.
Kite provides higher-level APIs atop existing components of CDH. It doesn’t replace, reimplement, or compete with Hive, Oozie, HDFS or HBase. All of these technologies work and play well with Kite. Kite APIs are the next level up the stack.
The ultimate goal is to provide an API that is closer to how developers think about their data model; a higher-level API that provides Hadoop’s flexibility and performance, but lets developers focus on their business problems.
Kite supports piecemeal adoption via loosely coupled modules. You can pick and choose the parts that make sense for you and your solution, without requiring you to use parts that don’t apply, or forcing you to follow a particular programming paradigm.