Intel sees Hadoop opportunity for APAC govt, telcos
SINGAPORE--Organizations in Asia-Pacific are looking to take advantage of big data and Intel believes its Hadoop distribution will allow them to mine big data cost-efficiently.
At a media briefing here Wednesday to launch the company's distribution for Apache Hadoop, RK Hiremane, Intel's Asia-Pacific regional director for data center and connected system product marketing, told ZDNet Asia organizations in the region were keen to take advantage of data they had amassed to improve their service quality.
"They all want to start with some small proof of concept," said Hiremane. With the Intel distribution for Apache Hadoop, organizations do not need to incur huge cost of purchasing vertical application stack or spend time modifying open source software to suit their business, he said.
In Asia-Pacific, he highlighted telcos and governments as two segments which will benefit from big data and Hadoop, and added that Intel had worked with organizations from these two industries when it developed its distribution.
For telcos, the velocity of data growth and increasing subscribers mean traditional data analytics software will take months to process information which is needed in real-time, he said. In India, mobile advertising firm Flytxt used Intel's Hadoop distribution to help its telco customers deliver relevant real-time information to subscribers. For example, a subscriber who has tried buying a product when they do not have enough prepaid mobile credit will be reminded to make the purchase after they reload credits into their accounts, he explained.
With 22 of the world's 37 megacities expected to be in Asia-Pacific by 2025, Hiremane said regional governments are investing in building smart cities.
The government of the Chinese province of Zhejiang, for example, is tapping Intel's Hadoop to generate about 2.5 petabytes of data each month in the form of video streams from CCTV cameras. With Hadoop, the Zhejiang government was able to solve the big data problem of storing, monitoring, searching ,and analyzing the data in real-time, he said.
He added that in healthcare research, which is a focus of the Singapore government, big data and Hadoop can shorten the time needed to mine data for medical research from months and years, to days and hours.
China among Hadoop testbeds
Intel had been working on Hadoop since 2009 and officially offered the product in China in July 2012, making the country one of the earliest to experience Intel's distribution, Hiremane said.
He explained that China was chosen because of the opportunities and the vendor's relationship with companies that want to try open source.
China is more open to open source than proprietary software because of capabilities that can be gained from an engineering perspective, he said, adding that China Mobile and China Unicom were early users of the distribution.
For Intel, key lessons learned from its Hadoop implementation include the need to simplify the management of the software, he noted. That is why Intel included a manager for its software to simplify the deployment, configuration, and monitoring of cluster for system administrators.
Hiremane added that customers had asked for more security with Hadoop, leading to Intel building security into its Hadoop distribution. Its distribution includes silicon-based encryption supported by Intel AES New Instructions (Intel AES-NI) in the Xeon processor.
Its distribution for Apache Hadoop will be sold through Intel partners comprising systems integrators (SI), independent software vendors (ISV), original equipment manufacturers (OEM), and training partners.