Since HBuilder code is running natively in Hadoop and Spark it can execute any business logic, including iterative algorithms. And since it does that in-memory it is also lightning-fast. This eliminates the need to write many MapReduce jobs and wire them together in workflow sequences – this means less code, less complexity, and faster execution.
Writing HBuilder apps is highly agile – you can process flat data records as well as hierarchical nested data structures. You can choose if the data must strictly adhere to a schema definition, a partial schema or no schema at all. And if you want, you can store the full lineage of the processed records giving you a complete view of all the data that went into the computation of an output record.
When processing master data, the detection of changes is always an important and slow step – HBuilder comes with super-fast delta detection: in a dataset with millions of rows, it will determine the Inserts/Deletes and Updates in seconds.
HBuilder batch apps can process data that is already stored in the Hadoop cluster or data that comes from external sources. At the same time, very often the outputs of a HBuilder application run need to be fed back to consuming system. HBuilder is designed to be open and can read/write from and to a variety of systems.
Flat Files, RDBMS and WebServices are standard data sources and targets to which HBuilder can read and write its data. To make it even easier to leverage HBuilder, it comes with its own adapter which connect it directly to SAP’s ERP/CRM or to BW. HBuilder can take SAP IDocs/BDocs as input or it can directly read from SAP application tables. It can connect to BW as a Data Source or (as of BW 7.4) it can bypass the PSA altogether and directly write to an InfoObject or an InfoCube.
When processing large amounts of data, transparency is particularly important: users need to be sure what has been processed. HBuilder comes with a flexible statistics module that allows the easy definition which attributes should be counted and how: attribute values can be counted individually or grouped according to a hierarchical structure. Numeric attributes can be assigned to freely defined number ranges.
HBuilder apps are targeted at large data sets: millions and even billions of records. Hbuilder is explicitly designed for performance: such apps will normally execute in seconds and minutes. HBuilder live stats give you a live update of the state of your application run. You can see how many records have been processed so far, how they are distributed and what quality problems have been detected – every second, in real-time.
Big Data applications depend on - obviously - big data sets. To get data to use for development and testing is very often a time-consuming activity for engineers. Source systems might not be ready to provide data, sometimes they might even not exist yet! Other times the data has to be cleansed and transformed and anonymized to comply with legal regulations.
HBuilder makes that step easy for you: it comes with a flexible and easy to use, yet powerful data generator. Just specify the target data structure, create a statistical profile and choose from several underlying raw data sources. Within minutes you have high-fidelity test data sets with hundreds of millions of records available for your development activities, both for static and for streaming use cases.
HBuilder contains a state-of-the-art graphical user interface (GUI) which makes working with Hadoop easy. No arcane command-line interface, instead an intuitive and colourful UI that removes the complexity and lets you focus on the work that needs to be done.
The UI layer is completely web-based and 100% HTML5. All modern browsers can be used to access and display the HBuilder WebUI.
Enterprise applications not only need a robust and fast execution layer. They also have common requirements for higher-level services: it does not make sense to re-invent the wheel and build all these services from scratch. HBuilder does that for you. It comes with a library of AppServices that handle many routine tasks for you: whether you have to extract delta information from two large data sets, execute quality checks and validate the integrity of a data set or to provide in-flight statistics on large batch jobs.
With HBuilder AppServices you don't have to build these functions yourself. You just add them via drag and drop and configure their runtime behaviour via a graphical wizards. This makes your development team so much more productive. It saves you time and money and helps you become really agile and fast-moving.