IBM PureSystems takes on big data
Subscribe now for $100 (23 issues) and save more than 37% off the cover price!
Get the latest news from Computerworld delivered via email.
Sign up now
IBM has added another trio of models to its PureSystems range of integrated computing systems to deal specifically with “big data” challenges. The Pure Data System, launched earlier this month at the IBM Interconnect conference in Singapore, emphasises high-performance data services for local storage or cloud operations.
The PureSystems range uses expertise derived from IBM’s own operations and those of its partners to tune computing capability ‘out of the box’ for maximum performance in commonly encountered business environments. These templates, IBM claims, facilitate more rapid implementation than conventional development techniques, and are better integrated than systems assembled from separately sourced components.
The three models of Pure Data System are designed for transactions, analytics and operational analytics respectively. Analytics on huge amounts of structured and unstructured data to predict, for example, the buying patterns of customers, is perhaps the best known application of “big data”.
Pure Data for Transactions is applied to environments such as retail and credit-card processing, where individual transactions represent small data volumes, but they occur in large numbers and rapidly.
Pure Data Systems for Operational Analytics deals with analysis of data deriving from the operations of the business in real time - for example detection of credit-card fraud or advice to call-centre operators in dealing with customer enquiries.
The handling of big data may seem less relevant to smaller markets like New Zealand, where businesses are generally much smaller, says IBM big data strategy executive Tim Young, but the other two chief characteristics of big data - its variety, consisting of structured, unstructured and semi-structured data, and the speed it flows through the ICT infrastructure and is refreshed with new data, means the techniques are just as relevant in this market.
Techniques to analyse unstructured data are well within the state of the art, Young concedes, but until recently “you’d have to go to a specialist” and face the challenge of integrating the unstructured-data analysis part with the rest of the system. “Now it’s become mainstream,” he says.
Though New Zealand does not have the volume for many big data implementations in the literal sense, its ICT users are “sophisticated”, Young says, and likely to provide a ready market for analytical techniques if well integrated with the rest of their business ICT.
The IBM approach faces strong competition from Oracle and the VMware-EMC-Cisco combination, as well as smaller contenders such as Teradata, but those do not have as good an integration story to tell, Young says.
Enterprise architects will feel wanted
IBM’s encoding of expert knowledge in its PureSystems range will not only remove some of the drudgery from configuring computer systems, it will help enterprise architects feel wanted, according to IBM .