For the week of June 29th, 2012:

• Away Goes Flash(y) My Oracle Support
• How Do You Avoid Losing Big Data and IT (the department)?

“IOUG Podcast 29-JUN-2012: No Flashy Oracle Support / Protecting Big Data and Your IT Dept”

Subscribe to this Podcast (RSS) or iTunes

Away Goes Flash(y) My Oracle Support

As we announced in a podcast last month, Oracle is planning to retire the existing Adobe Flash-based interface for My Oracle Support (http://supporthtml.oracle.com) and continue future development using the OA Framework-based HTML version.  On July 13, 2012, Oracle will be upgrading the HTML interface with new functionality which will include:

• Oracle Configuration Manager
• Patch Recommendations
• On Demand and CRM On Demand (aka Cloud Services)
• Many Usability enhancements

Desupport of Internet Explorer 6 is part of this upgrade.  Only IE7+, Firefox 3+, Safari 3.1.2+, and Chrome 5+ are officially supported with My Oracle Support after July 13th, 2012.

Current user benefits included in the revised UI are:

• A streamlined, three-step process for initiating new Service Requests (SRs)
• A single, consistent workflow for both hardware and software incidents
• Enhanced personalization and filtering within the user interface
• New accessibility features (enabling screen readers, large fonts, and ADA compliance)

You can learn more about of the enhancements to the MOS UI at Doc ID 1385682.1 (https://support.oracle.com/CSP/main/article?cmd=show&type=NOT&id=1385682.1)

How Do You Avoid Losing Big Data and IT?

In several articles published by CIO.com (an online technical journal), Thor Olavsrud, a technical author specializing in IT security, Big Data and open source technology, covers some emerging trends related to how Big Data is evolving in organizations that traditionally have thought only in terms of data and hardware resource management.

He points out that the evolution started with the availability of products outside of the enterprise that simply work better, faster or easier than those provided from within the business. The culture that spawned bring-your-own-device (BYOD) as a potential expense savings to the business, probably didn’t realize that the same principle applies at the larger enterprise level – the traditional dominion of an IT department.

Olavsrud observes that not only employees but whole divisions can turn to the cloud of external applications and avoid IT entirely if they believe they are not getting the tools they need or want from the internal sources. To develop a highly automated, scalable and adaptable IT function requires ‘dematerialization,’ similar to how audio CD’s gave way to digital MP3 files and DVD’s are being replaced by cloud-streamed media content. Leslie Muller, co-founder and CTO of DynamicOps, a Cloud automation and management solutions provider, says, “The future data center is 100 percent virtualized or as close to it as possible. Storage, network, apps, and everything else is 100 percent automated. We don’t have users anymore, we have consumers. It’s a mental shift. It’s not a technology issue. The place to start is to understand who your users—your customers—are and then adapt automation and processes to those users’ demands.”

Along with these increases in flexible, adaptable and scalable solutions, comes the creation of petabytes of data being generated in response by these wonderfully useful customer-focused applications. This has a corresponding demand for trend analysis and decision intelligence based upon that data, which in return, requires more data generation and storage.  If you attempt to place that volume of data into a conventional RAID array, you’ll find that the rebuild time for the inevitable fractional hardware failure now climbs into days and weeks when dealing with latest generation 1- to 3-terabyte drives amassed into petabyte-sized arrays of storage.

Solution providers are returning to an older technology for the answer: the hashing algorithms of error correction code (ECC). Whereas parity algorithm-based solutions (RAID 0) are designed around the detection of missing bits to flag when a bit needs to be restored from a mirror copy (RAID 1), ECC and forward error correction (FEC) are used to reconstruct missing bits based upon the final packaging of bits received.  Originally used for sending data over unreliable channels, such as phone-line modems, FAX transmissions, and deep space probe communication, these erasure code packets, when added to the original data packet, allow reconstruction of the packet upon receipt, even if whole portions of the packet are missing or corrupted.

Chicago, Ill.-based Cleversafe, a Big Data storage technology specialty company, adds location information to the ECC/FEC coding which allows users to store chunks or slices of data in geographically separate places.  While the individual slices are mathematically useless on their own, ensuring privacy and security, this dispersal coding allows a single data repository with minimally added erasure code overhead, to become recoverable without requiring the data duplication needed in standard array architectures. Since erasure coding is software-based, it becomes hardware vendor-independent allowing commodity or cloud data services to be employed without the usual data storage replication requirements.

Oracle’s own engineered system products such as Exadata have packaged the required hardware, with similar storage management software, and incorporated certified versions of its own Database engine, as well as a newly certified NoSQL engine into scalable packages designed to fit a number of different customer capacities. Though these off-shelf solutions come at a not-so-commodity price point, the savings in time and human resources required to deploy these ever-increasingly complex solutions could well justify the investments.

You can read more about Cleversafe’s solutions at cleversafe.com, read the complete library of Olavsrud’s insights at cio.com, or explore Oracle’s own solutions at oracle.com/bigdata