IBM’s big data eggs start to hatch

IBM’s big data eggs start to hatch

IBM’s big data eggs start to hatch

0 comments 📅04 July 2015, 02:45

IBM’s big data eggs start to hatch

IBM now develops a gamut of innovative software and hardware technologies that can help companies ingest and make sense of Big Data. This was the key message Ovum came away with from a recent analyst conference at IBM’s historic Almaden Labs in San Jose, California, an apt location given the lab’s reputation for ground-breaking innovation in IT, notably the hard disk drive and the SQL database.

The IBM announcements centered on providing greater speed, easier consumption, and quicker deployment of Big Data analytics solutions that map to core IBM technologies such as DB2, Big Data Platform, and PureData systems.

IBM certainly has enough eggs to create a mean Big Data omelette, but the question is whether it has a basket big enough to hold them all securely and present a coherent strategy to customers.

Big Data continues to creep up the corporate agenda

IBM believes that Big Data now represents a boardroom-level conversation, discussed seriously by non-technical business users as a competitive game-changer. It is therefore no surprise that IBM is taking an interest. Importantly, IBM’s broad definition of Big Data makes the V-debate increasingly moot.

Transactional, social, M2M, and unstructured data all contain business insights, and now factor into IBM’s Big Data analytics strategy, and the debate is now shifting to business-driven analytics use cases, homing in on use cases in three distinct areas: customer management, risk and fraud prevention, and augmenting traditional enterprise data warehouse-driven business analysis, that can all be applied across a range of vertical industries.

At the same time, IBM believes that for customer adoption, it is critical to link its Big Data management and analytics strategy to a clear enabling IT “platform”. The platform that IBM is putting in place to leverage all types of data and new types of analytics enabled by Big Data is now starting to take shape and should help companies to link boardroom debates to actionable reality.

The IBM brand certainly commands respect and recognition in corporate boardrooms. For many CIOs, IBM infrastructure will be a natural and safe first port of call for Big Data needs. However, the risk for IBM is presenting to customers an embarrassment of overlapping and confusing technological riches.

A Big Data solution cannot be delivered on a plate because it is a strategy that’s unique to every enterprise and industry vertical, and to succeed it requires investment in both technology and expertise.

BLU acceleration stands out from the latest IBM package of Big Data announcements

A recent Ovum research note, “IBM introduces new analytic Architecture”, about IBM’s Big Data platform centers on the PureSystems line of appliances. It highlights a broad package of announcements across a number of so-called Big Data platforms, including IBM’s first Hadoop appliance; incremental functional enhancements to the BigInsights Hadoop offering, notably direct support for SQL; and introduction of the new BLU architectural design pattern for DB2 that includes columnar table support.

Of the announcements, the BLU architecture is the most significant for analytics. Individually, the features of BLU are not unique, but together can provide a holistic approach that could be used for data acceleration across a number of IBM’s data platforms. With this, IBM is enabling “speed of thought” analytic insights derived from Big Data through its BLU acceleration initiative.

The goal is to accelerate query workloads, reduce storage savings, and increase simplicity by removing the need for indexes, aggregates, and tuning of SQL schema changes, and replaced by a notion of load-and-go Big Data analysis.

This might sound too good to be true because it risks underestimation of some stiff data-quality and governance challenges. Nevertheless, IBM is adamant that it has the technological know-how to make it happen, thanks in no small part to the confluence of several hardware and database trends converging favorably and economically.

A peek under BLU’s hood certainly highlights advances in the way that IBM is intelligently leveraging in-memory, vector-based parallel processing, “actionable” compression, and clever data-skipping techniques, all of which combine to drive optimal analytics systems design, storage, and performance configurations.

BLU acceleration represents the start of a journey by IBM. The technology is being released first in DB2 10.5 and also in its Informix TimeSeries platform, with plans to extend it to more workload classes and new products.

The Big Data analytics market is still up for grabs

IBM is swimming in a vast a sea of vendors vying to help organizations make business sense of their Big Data. Its rivals are the usual suspects in the BI and analytics world, and competition is certainly heating up. Ovum believes that Big Data could be the next big battleground in the war between the large IT vendors, notably IBM, Oracle, and SAP. All three companies have now assembled impressive analytics solutions that are itching to be pointed at Big Data.

Arguably, IBM has the biggest set of technological assets, data science eggheads, and professional services to help companies manage and make sense of their Big Data.

However, Ovum believes that IBM should pay particular attention to the issues of Big Data “ownership” and “understanding”, which we believe should replace the now hackneyed debates on the obvious and ever-expanding Vs of Big Data.

In this respect, IBM has considerable technological and intellectual clout, but sadly, we did not hear too much about that in Almaden. However, IBM’s data management group certainly has enough resources and technology, particularly with its InfoSphere data integration, data quality, and master data management (MDM) products, as well as its Optim and Guardium suites, to make sure that Big Data behaves itself.

No Comments

No Comments Yet!

You can be first one to write a comment

Leave a comment