Big data start-ups Wibidata Hadoop "packaged"

Large data start-ups wibidata is ready to encapsulate Hadoop into a specific software collection packet, which may open up a new Hadoop journey. The co-founder of the company also created the "famous" Cloudera, and the company won Google chairman Schmidt's two favor.

Wibidata the start-up may be "obscure", but the company's co-founder Christophe Bisciglia and Aaron Kimball also created Cloudera, a company founded in 2008 that can be counted as " Famous ". In 2012, Cloudera also received a 65 million-dollar e-round financing, which was also the No.1 of the 2012-year large data-sector financing.

In fact, Wibidata at the beginning of the inception, there is no grand strategic planning, just want to be the first to provide a business application based on Hadoop software provider (if not first, it should be one). In May of this year, Wibidata announced a 15 million-dollar round of financing, funded by Canaan Partners, as well as the original investor NEA and Google chairman Eric Schmidt, who helped Wibidata accomplish this ambitious goal.

Now, the market for Hadoop is also "hot", and many companies are ambitious enough to do it, but expensive consultancy fees, as well as the need to develop different applications for specific goals, have blocked the footsteps of most innovative companies. Even the result is that many businesses are not limited to moving from unstructured data to structured data (often called ETL) or performing some backend analytics. In fact, Wibidata has spent 18 months doing it. Bisciglia said the three giants of the Hadoop industry (Cloudera, Hortonworks and mapr have largely monopolized all market share.

Home Energy management start-up Opower is a good user case. The company was originally a Cloudera customer, but "when Opower wanted to use SaaS tools to go beyond the bulk analysis and ETL workloads," Bisciglia said Opower had to seek wibidata help. Originally Opower Services can only focus on the analysis of users at night energy use data, now can be dynamically recommended to users, in the user and the application to build a new way of contact.

Wibidata System Architecture

Wibidata has built its core technology architecture, linking the back-end Hadoop environment with customer-facing predictive applications, including hbase datasets, data-formatting tools, and machine learning algorithms. Wibidata also created an Open-source project called Kiji.

Kiji Open Source Project

Wibidata Create open source projects Kiji to provide a framework for building large data applications. The core module of Kiji is Kijischema, which provides a simple Java API for storing and managing typed data in HBase by Avro storage serialization. Like a layer of wrappers built on HBase, the goal is to make it easier for application writers to manage structured data with hbase instead of using it as a flat table.

The basic operations of Kiji include the creation and modification of kijitable and the reading and writing of entity data. The process steps of the operation are similar to those of hbase, and there are many corresponding conceptual objects such as configuration/admin/table.

Generally speaking, Kiji is basically a encapsulation of the hbase application pattern, using Avro to host the object data to facilitate the evolution of schema. Strengthen the concept of object-oriented programming from a data perspective (relative to HBase Table). It's about users who want to be able to use HBase to store data and develop applications quickly. (Information from CSDN blog)

Bisciglia said Wibidata will focus on developments in two sectors such as finance and retailing in the near future. He expects that future-shaped products will include predefined patterns of data formatting and prefabricated predictive models, both of which are intended primarily for industry customers in broad terms, rather than for individual users. Wibidata will also provide different kinds of interfaces for different types of users, such as data scientists, system engineers, and business users, and they can interact with the data in their own unique way.

Time tells us everything, and if wibidata can really achieve its goals, it can be a new milestone to encapsulate Hadoop in a specific software collection packet. Even industry giants like Cloudera see these requirements, but their full business is integrating Hadoop into the existing environment for early use and operation. Cloudera CEO Mike Olson has said, "If anyone has the ambition to solve the problem of convergence between Hadoop and apps, call me and I'll give you the money right away!"