Cloud Bigtable’s personnel : Google of scale-out big data strategy

Google introduced in 2004, it had a major impact on big data. And now, what Google is trying to bring about what the big data? He was interviewed on Cloud Bigtable personnel.

2004, the US Google that underlie the ingenious white paper that the “MapReduce” was introduced to the world as the initial initiative of big data, it’s programming concept that has been devised in the distant past of 1958. Using this MapReduce, Google tried to deal with problems that prevent the processing of search data. In the essence, MapReduce and the big data is to some extent divided, it is something that is to be processed using the “Hadoop” performed by low-cost commercial hardware.

The company, which provides a search engine, has started an extension of the data processing strategy. Recently, it announced a fully managed scalable NoSQL database service of “Cloud Bigtable”.

apache spark

Internet search, social media, IoT (Internet of things) is, but the IT field, which is facing a rapid increase in the amount of data. In a conventional relational database, required by the true digital society, when they are not able to process the data surging as tsunami experts predict.

In MapReduce of white paper that has been published in more than 10 years ago, Google of Jeffrey Dean and Mr. Sanjay Gemawatto, a single infrastructure is described as not to be able to process by scheduling a job for different types in a single common infrastructure ing. At that time, according to specific environments and architectures, it is necessary to create everything manually.

According to the product manager Mr. Cory O’Connor in charge of Cloud Bigtable, the company now, that is working on the third big data vision be revised from the announcement of this white paper. “2002 to 2004 is the year that Big data has been extensively discussed, but also this time the Company has written a white paper on MapReduce”

“We, such reviewing the technique of building a huge computer to solve the problem from the root. Use the only commercially available computer, assuming that would system fails.”

Extremely high scalability, the idea of ​​using the low-cost general-purpose infrastructure, Big 4 IT supplier is in the almost counter electrode and method to tackle big data. Of course, the Big 4 to are also working on NoSQL, it’s also provide Hadoop in the cloud. But for example, Germany SAP is required for the customer to be spending millions of dollars to the “S / 4 HANA”, the US Oracle is pushing the appliance family that “Exadata” and the company has design, rice IBM is “z13” Mainframe and sell the benefits, the US Microsoft recommend “SQL Server.”

For example, Z13 mainframe can analyze fraud in real-time in financial transactions. According to Mr. O’Connor, that such time data series can be treated naturally in Cloud Bigtable. “The approach to the problem, it is important to understand that there are a variety of ways if the vertical is. Hours, you’ll which method is most efficient or demonstration”

According to Mr. O’Connor, Google that different from other companies is that “it knows how to manage very large data sets, but it has a big data architecture of fully managed”.