Home & Office

Oracle takes aim at big data rivals

IT vendor unveils new product lines at OpenWorld conference that CEO Larry Ellison says are capable of running unstructured data analytics faster. Market analyst notes Oracle challenged in convincing customers to invest.
Written by Edwin Yapp, Contributor

SAN FRANCISCO--Oracle has signaled its intention to take the fight to its competitors in the nascent but burgeoning sector, big data processing, through a new line of appliances.

In his opening keynote at Oracle OpenWorld held here this week, CEO Larry Ellison announced several new products which he said would enable the IT vendor to take on rivals IBM and SAP. These included Oracle's Exalytics business intelligence (BI) and big data appliances.

Both products are designed to address the rising volume of unstructured data that enterprises today need to support, with the growing adoption of social media, blogs and sensors.

Speaking to an audience of 10,000 delegates, partners and reporters, Ellison said Exalytics handles relational and multidimensional data, and also analyzes unstructured data "at the speed of thought". "Nothing is faster. There is no response time," he added.

Underpinning this processing ability is the company's latest "in-memory" technology, which the CEO touted as an industry's first. The Exalytics system comprises hardware and software components engineered to run analytics faster, provide real-time visual analysis, and enable new types of analytic applications, he added.

To complement its foray into in-memory BI, Oracle's big data appliance offers customers a way to deal with the growing volume of unstructured data. Thomas Kurian, the vendor's Oracle's executive vice president of products, said in his presentation that the new offering would give customers the ability to acquire, organize, analyze and maximize the value of big data within their enterprise.

In a statement released earlier, Oracle said the appliance will ship with open source database management framework, Apache Hadoop, Oracle NoSQL Database, Oracle Data Integrator Application Adapter for Hadoop, Oracle Loader for Hadoop, and an open source distribution of R.

Ellison also made reference to the company's recently launched T4-based Sparc SuperCluster and proclaimed that Oracle's goal was not only to run faster than IBM's fastest chip, but to do so at the lowest and best cost-performance.

"We're determine to deliver the best-of-breed in every aspect of our computing architecture, both hardware and software," he said. [We're aiming] to not just be better in database and middleware [but] we want to take on IBM in its strongest suite, which is its microprocessor, and we will try to beat them in arithmetic."

The key to achieving this is Oracle's "parallel everything" mantra, he noted. Acknowledging that IBM was still the leader in computational integer arithmetic, Ellison said having raw processing power meant nothing to enterprises which demanded more than that.

He explained that computational performance should help customers access both structured and unstructured data, and provide the ability to process this data faster than before. To do this, IT vendors must offer better ways to squeeze more power from the entire computing infrastructure, rather than just the microprocessor, he said, noting that there was only so much power one could squeeze out of current chips.

Adding that the industry was essentially already using the same commodity devices, from disk to flash drives, Ellison said: "So how do we go faster? Not by having a single large infrastructure but by making everything run in parallel--from storage, to database, to middleware.

"It's not the only way to do it, but it's certainly one of the ways."

Oracle's challenges remain
According to Tim Jennings, Ovum's research fellow and chief technology analyst, Oracle's big data appliance was the strongest announcement to emerge from OpenWorld.

Jennings told ZDNet Asia that the IT vendor's stack approach made sense as it removed some of the complexity from big data processing, underlining the potential benefits of engineered systems. Oracle described such systems to have tightly integrated hardware and software, which are built from ground up and optimized to work in union with each other.

Jennings said: "These systems mainly play at the top end of the market. So whilst Oracle is building a strong pipeline for the products, it is not a game-changer for enterprise IT as a whole."

He added that the tech vendor must convince sufficient numbers of customers to invest in these integrated systems. With the majority of these purchases driven by consolidation play--comprising database, server or application consolidation--this would inevitably be a gradual process, he said.

"I think we are likely to see further mid-market launches to increase the reach of these systems," the Ovum analyst said.

Oracle also faced the challenge of persuading customers to take all factors in assessing the total cost of ownership (TCO) of these systems, including hardware, software, deployment costs and ongoing management and maintenance.

"If these are considered separately [as is often the case], then it makes direct comparison with Oracle's integrated solutions difficult," Jennings explained. "Oracle should be offering customers TCO calculators to assist this process."

Edwin Yapp, a freelance IT writer based in Malaysia, reported from Oracle's OpenWorld in San Francisco, USA.

Editorial standards