Get fresh updates from Hortonworks by email

Once a month, receive latest insights, trends, analytics, offering information and knowledge of the Big Data.

cta

Get Started

cloud

Ready to Get Started?

Download sandbox

How can we help you?

closeClose button
Hortonworks Customer

SoftBank Corp

Key Highlights

icon6.png

Stable enterprise-grade operation of their Hadoop system

icon6.png

Improved efficiency of operation processes through Apache Ambari

icon6.png

Distributed processing of large volumes of data at low cost

A Revamped Platform to Analyze Ever-Increasing Data for a Leading Telecommunication Carrier

SoftBank Corp. (SoftBank), a leading telecommunication carrier, has adopted Hortonworks Data Platform (HDP) as its data analysis technology platform. HDP is a 100% open source package software platform built on the distributed processing software Apache™ Hadoop® and includes other open source software such as Apache Spark. For SoftBank, HDP provides high-speed storage and analysis of ever-increasing telecommunication data at a significantly low cost.

Business Challenges:
• Handle ever-increasing volumes of telecommunications related data
• Improve internal technical skills and increase knowledge base
• Optimise and improve analytics system efficiency

Deployment Results:
• Stable enterprise-grade operation of their Hadoop system
• Improved efficiency of operation processes through Apache Ambari
• Distributed processing of large volumes of data at low cost

Building a Data Storage and Analysis Platform for the Future

Large volumes of telecommunication data are generated on a daily basis at SoftBank. Explaining the handling of this data, Yasuyuki Kato, Senior Director of System Infrastructure Division, says, “SoftBank is a telecommunication carrier that generates large volumes of data in confidential communication records. These require extremely careful handling and as a result, most of it could not be used in the past.”

The advanced security and governance features of the Hortonworks Data Platform enable SoftBank to more safely and securely take advantage of these large volumes of data.

Shin Matsuura, Acting Manager of System Infrastructure Development Section, feels that the adoption of the technology will yield rich dividends in the future. “We are aware of the potential of data utilization as a platform to leverage on SoftBank’s strengths in the future. As the data storage platform is the way to the future, we started preparation from an early stage.”

SoftBank began Hadoop-distributed data analysis testing in 2010. “Before ‘big data’ became a popular term, we were already considering how to leverage our own data for business growth. Fusing the prowess of third party enterprise software with our in-house capability, we decided to deploy Hadoop ourselves. That enabled us to improve our technical skills, as well as reduce costs,” says Matsuura.

As one of the objectives was to improve employees’ skills, the scope of the deployment of Hadoop was not confined to any particular focus area. SoftBank first started from a test environment and tried analyzing data from its existing data warehouse (DWH) for testing.

Using Expert Solutions To Achieve Optimised Data Analysis Platform Environment

According to Matsuura, “We set up the proof of concept, loaded various data on Hadoop, and verified operations and results to gradually build up our knowledge.”

After several years of knowledge building, SoftBank was convinced of the potential of Hadoop and at the same time, opportunities for big data utilization in business also increased. Thus, in 2015, a Hadoop environment of a larger scale was built in SoftBank as a company-wide project. Hortonworks’ HDP was selected as the Hadoop distribution.

According to Matsuura, “Our technical skills have improved over the years in the development and operation of the Hadoop environment. On the other hand, with limited understanding of all the source codes, we were unable to judge whether the environment we created was optimized or not. That was why we decided to engage the expert services of an experienced distribution partner to validate and value-add to our implementation.”

Kato also expressed his satisfaction with Hortonworks’ components and support. “Hortonworks has a lot of experience as a pioneer in this area, and we really appreciate their support. When using open source software, it is quite common to encounter operation problems such as poor performance or absent data, but with Hortonworks’ support, our operation is very stable. I’m very satisfied with the quality they have delivered. Their engineers are very skilled and experienced.”

High Appraisal of Apache Ambari Operation Management

SoftBank implemented the distributed file system HDFS and components such as the Apache Hive for processing SQL. Apache Spark is used for high-speed in-memory processing. Ambari is also used to improve operations and management efficiency across the platform.

“We find Ambari most useful. When Ambari was first installed, we thought it was a typical feature but now, we feel that it provides the capability that we want. The centralized log consolidation feature is great. Without it, we would have had to collect these logs ourselves and store them somewhere, but Ambari automatically consolidates the logs and presents them in an easily viewable format. We hope that more such features will be added in the future,” said Matsuura. Kato feels that the development staff at Hortonworks pay great emphasis on usability.

“The trend is toward more advanced and multiple features, and we are impressed that Hortonworks understands what features we need onsite.”, he said.

Using Data for Various Business Purposes

SoftBank is considering storing and analyzing telecommunication-related information on its data analysis platform, and how to best utilize it for one-to-one marketing in the future.

According to Kato, “We cannot be as aggressive as web service providers because we are handling data as a telecommunication carrier. However, I am confident that by doing Hadoop-based analysis on the data, we will be able to put it to meaningful use. In addition, since we also store large volumes of a variety of data such as IoT data, cost will always be a concern. The open source Hadoop-based HDP is very effective as it stores more data at a lower cost and offers high speed data analysis through distributed processing.”

Talking about the role and mission of the system department in the future, Kato said, “Big data has unlimited potential. It is important that the information systems department can prepare the data and environment in advance to support user analytical needs. SoftBank will continue to its efforts to advance revolutionizing information through state-of-the-art technologies and superior business models in the future.”

Overview of System Platform

Overview of System Platform

About SoftBank Corp.

The SoftBank Group strives to make lives better by revolutionizing information. It aims to be “the corporate group needed most by people around the world.” To achieve its vision, SoftBank will continue to step up its effort to revolutionize information through state-of-the-art technologies and superior business models.