3 Comments

Summary:

For years, Oracle has wowed Wall Street with fat software margins: Large companies depending on Oracle relational databases pay what it takes to keep them up and running. It’s unclear whether Oracle can carry that dominance over into the Big Data era, however.

2058017896_44a24866c1_z

For years, Oracle wowed Wall Street with its fat margins on software licenses and maintenance. The fact was, many large companies depended on Oracle relational databases and paid what it took to keep them running. They still do.

Oracle’s problem is that relational doesn’t go far enough in the era of big data, which requires the handling of unstructured, semi-structured and all manner of information from Twitter, Facebook, server logs, you name it. And, for once, Oracle can’t jam all that stuff into the relational mold. Instead, it launched a Cloudera-based Big Data appliance, in January.

While the pricing was low compared to what analysts expected, it’s still a big “scale-up” sort of appliance. List price for the Big Data Appliance is $450,000 plus another $54,000 for hardware maintenance and $36,000 for operating systems maintenance. That’s a lot less than the Oracle Exadata appliances, but it’s still a significant outlay. You can buy a ton of standard X86 boxes — the sort of gear typically used for big data crunching –for that kind of money.

Oracle will undoubtedly sell some of these machines into existing accounts, but it’s far from clear whether the software giant can carry its dominance from the relational to the big data world. The companies that pioneered big data — Yahoo, Google, et al. — champion the use of low-end, low-cost commodity hardware running open source Hadoop and MapReduce software. For those of you new to Oracle — low-cost software and hardware are not really in the company’s wheelhouse.

The Wall Street Journal’s Heard on the Street columnist Rolfe Winkler laid out Oracle’s dilemma on Monday:

… it turns out that much of this information is best processed not by an Oracle database but by cheap hardware and open-source software. One problem is scale. As data pile up, companies must expand systems quickly to handle this. Another problem is complexity. Much of the data can’t be easily categorized by traditional databases organized in rows and columns: Tweets don’t really fit a spreadsheet.

This big data issue, paired with Oracle’ s difficulty in wringing software-like margins out of its hardware franchise (from the Sun Microsystems acquisition) spells big problems for the enterprise software powerhouse.

For its most-recently closed second quarter, hardware revenue fell 11 percent from the year ago quarter to $1.47 billion, something that prompted Oracle CEO Larry Ellison and co-president Safra Catz to pledge a hardware sales turnaround in the new fiscal year. That news came on the heels of an even worse second quarter for hardware.

Oracle’s hardware issues along with the cloud computing challenge which GigaOM Pro analyst George Gilbert examined extensively, are two big hurdles to Oracle’s continued dominance. But its big data problem may be the capper.

Photo courtesy of Flickr userPeter Kaminski

  1. Maretha Britz Tuesday, April 10, 2012

    Mhh….

    Share
  2. Reblogged this on BriefingsDirect.

    Share

Comments have been disabled for this post