Business Analytics: Outdated Software Slows Down New Hardware

Business Analytics: Hardware manufacturers have long complained that their further developments do not significantly accelerate, the applications because the programs available for this do not appropriately use the new possibilities. But improvement is in sight. The need for robust data analysis tools is booming. The reason for this is the rapidly increasing flood of data – both for structured and for the many new unstructured data, such as those generated by the popular social media channels Facebook, Twitter and Google+. Keyword: big data.

IDC market researchers predict that this market will swell by 8.9% to just under $ 34 billion this year. According to the report, Oracle is the market leader with a 20% share, followed by SAP, IBM, SAS Institute and Microsoft.

Business Analytics: Classic Software Is Powerless To Face Big Data

But despite these impressive growth rates, the business analytics offering is in a crisis. “Classic analysis

software is relatively powerless to face the new challenges posed by big data”. Although the hardware computing power is getting better and better, most traditional programs can hardly use the many new hardware features such as multi-core processing and column-oriented database organization in the main memory since they are strictly sequential processing.

To change that, the providers are currently pursuing two strategies: bundling software with hardware in an appliance. EMC follows this path with Greenplum, and IBM also offers such a solution with the Netezza appliance. The second way is to recode the software to take full advantage of all modern hardware features. But this is not only a very costly and time-consuming process – it is also a precarious approach. “With parallel processing, you always have to consider what you can process independently of each other, and when does a program step have to wait for the end of another process – this is particularly tricky with time series,” about one of the most significant Recent SAS projects because his company has completely rewritten the analytics software.

Hadoop Is Growing In Popularity

In addition to these two paths, i.e. application-optimized hardware or rewritten software. The use of the open-source software Hadoop is becoming more and more popular. Apache sells a data processing platform, and many large hardware providers now offer that.

Oracle has been offering Hadoop for a long time. IBM also uses this platform. Microsoft even recently stopped developing its own big data analytics platform, Dryad, favoring Hadoop. EMC, Teradata and Hewlett-Packard also offer their hardware in a package with Hadoop.

The fact that so many established providers rely on Hadoop software is due to its remarkable architecture, which on the one hand, fully exploits all the possibilities of parallel processing, but on the other hand, does not require any special hardware.

The reason for this is the rapidly increasing flood of data – both for structured and for the many new unstructured data, such as those generated by the popular social media channels Facebook, Twitter and Google+. Keyword: big data. The second way is to recode the software to take full advantage of all modern hardware features. But this is not only a very costly and time-consuming process – it is also a precarious approach. Oracle has been offering Hadoop for a long time. IBM also uses this platform. Microsoft even recently stopped developing its own big data analytics platform, Dryad, favoring Hadoop. EMC, Teradata and Hewlett-Packard also offer their hardware in a package with Hadoop.

ALSO READ: “Big Data Comes Faster Than Many Think!”