kentoh - Fotolia

APAC banks warming up to big data

Financial institutions in the APAC region fear they may be left behind if they don’t embrace big data

This article can also be found in the Premium Editorial Download: CW ASEAN: CW ASEAN: Rise of the machines

While global banks in Asia-Pacific (APAC) have been adopting big data in droves, more local banks in the region are now doing so to gain a “second-mover advantage”, according to a senior executive from Cloudera.

Without naming specific banks, Cloudera’s financial services industry lead Steve Totman said the smaller banks in APAC now know that they need to embrace big data to detect fraud and gain deeper consumer insights.

The drivers for big data adoption, however, were different eight years ago, according to Totman.

Then, many companies had implemented big data to cut costs, by offloading data warehouse ETL (extraction, transformation and loading) to Hadoop, an open source framework for storing and processing big data sets. That was largely because the storage costs of Hadoop are substantially lower than that of high-end data warehouses.

After realising the cost benefits, Totman said those companies might pick another case such as fraud detection, followed by another one such as mapping a customer’s journey six months later. “But in the past 18 months, we see them starting with as many as 15 cases in 13 months,” he added.

The growing interest in big data among smaller banks in APAC stems from the realisation that they may be left behind if they do not embrace the technology, Totman said.

“If you started with Hadoop eight years ago, you did it because it gave you competitive differentiation: you could analyse data that no one else could, and at a lower price point. But now, it’s about survival,” he said.

Big data an APAC priority

Big data is now a top agenda among senior business leaders in the region. According to an IDC study sponsored by Cloudera, 40% of respondents in APAC said they are deploying or planning to deploy Hadoop in the next 12 months.

Totman said while discussions around big data were confined to technical teams in the past, he now engages top executives and chief information officers (CIOs) on the business benefits of big data, rather than the underlying technology. “Some banks in Singapore are incredibly visionary,” he said.

Take Oversea-Chinese Banking Corporation (OCBC), an early adopter of big data, for example. With the help of its retail partners, OCBC has been analysing customer and transaction data to not only better understand consumer spending behaviour, but also develop products such as Frank, a suite of financial services targeted at young adults.

DBS Bank, another big data trailblazer, has also partnered Cloudera and the Institute for Infocomm Research at Singapore’s Agency for Agency for Science, Technology and Research to detect abnormal transaction activities in trade finance using big data – the first of its kind in Singapore.

Daniel Ng, senior director of marketing at Cloudera in APAC, added that DBS is also looking at using behavioural analytics and machine learning to answer customer questions before they are posed to the bank through its website.

Ng noted that for DBS, such efforts to leverage big data are aimed at building customer loyalty. This dovetails with the results of the IDC study, which found that the top reasons for deploying Hadoop are retaining and capturing customers, as well as growing revenues from existing products and services.

Big data in cyber security

In the same study involving 636 IT decision makers and influencers for big data from large organisations in the APAC region, cyber security was also cited as a driver for implementing big data initiatives.

“Cyber security is an incredibly interesting area in big data,” Totman said, adding that Hadoop can be used to make sense of huge volumes of data not just from security information and event management (SIEM) systems, but also from other sources.

To that, Totman said Cloudera and Intel spearheaded Apache Spot, an open source framework that uses machine learning to make sense of network traffic and identify malicious data packets from benign ones. It also uses an open data model that brings together all security-related data across an organisation.

Such data can be analysed through machine learning techniques and artificial intelligence (AI) algorithms to prevent breaches such as the recent incident involving I-net, a Singapore Ministry of Defence (Mindef) system that provides internet access for servicemen and employees.

According to Mindef, the identity card numbers, telephone numbers and dates of birth of around 850 servicepeople and employees were stolen from the system. No classified information that resides in separate air-gapped systems was compromised.

Mindef said the attack on its system appeared to be targeted and carefully planned. “The real purpose may have been to gain access to official secrets, but this was prevented by the physical separation of I-net from our internal systems,” it added.

Sanjay Aurora, managing director at Darktrace Asia-Pacific, said this latest high-profile breach would further ramp up the pressure for Singapore organisations to adopt AI and machine learning to automatically detect and respond to these inevitable threats, before data can be stolen or changed.

“Although it appears that Mindef has responded swiftly to this incident, the reality is that no human can keep up in this rapidly evolving threat landscape,” Aurora said.

“It is a cyber arms race and AI technology that self-learns what is ‘normal’ for a network and automatically identifies and takes action against abnormal behaviour and genuine threats will be instrumental in safeguarding critical information and infrastructure,” he added.

Read more about data analytics in APAC

Automated security controls are a key driver of the adoption of big data, according to a global survey of more than 330 respondents from 50 countries by security analyst firm KuppingerCole and the Business Application Research Centre.

The survey report noted that by offering a significantly higher degree of automation than previous generation SIEM products, real-time intelligence systems that leverage big data can help even non-technical staff make informed choices from actionable security alerts and initiate an automated incident response workflow.

Such workflows do not only “involve technical steps like blocking an infected workstation on the corporate firewall or disabling a rogue user account, but also a number of organisational, legal and even PR actions to contain every aspect of damage a breach may have caused,” according to the report.

Read more on Big data analytics