Oracle unveils faster Exalytics platform
Oracle recently unveiled a faster version of its Exalytics in-memory analytics database. The company claims the X3-4 version delivers up to 25× performance improvements over its predecessor, thanks largely to pure and simple hardware upgrades – i.e. more memory.
Ovum does not doubt the sheer speed of Exalytics, particularly for data analysis where ASAP is simply not quick enough. We see the most immediate applicability for such an ultra-responsive analytics infrastructure in driving more agile business planning and budgeting processes, but there are also a raft of other emerging use cases. Certainly benefits are to be gained from the raw power of Exalytics, but companies must consider carefully their strategies for realizing the benefits in business-critical applications.
Exalytics X3-4 reflects the accelerating pace of doing business today
Nowadays, data analysis is a deft mix of data scale and processing speed: whether it is about providing recommendations to online ecommerce customers; identifying credit card fraud during a point-of-sale transaction; or allowing call center agents to spot cross-/upsell opportunities while on the phone with customers.
It requires BI and analytics to sift through Big Data sets at unprecedented breakneck speeds (something Oracle dubs “speed of thought” analytics). This technical requirement was the genesis of its first generation Exaltyics X2-4 launch in February 2012. It now seems that X2-4 is no longer fast enough, hence the X3-4 release. But it is perhaps a natural evolution of a product that takes advantage of Moore’s law in improving its performance and capacity to accelerate analytic applications using in-memory processing.
The performance math certainly makes for impressive reading
The key to this acceleration is Oracle building more muscle into Exalytics hardware: 2TB of main memory (double that of X2-4), 5.4TB of hard disk (up from 3.6TB), and a new cache of 2.4TB of flash memory, which existing X2-4 users can upgrade to at a cost.
Internal Oracle benchmarks for X3-4 boast dramatic improvements in load and calculation times – both of which are key to analytic agility. Oracle claims the additional memory provides nine times quicker calculation times for multidimensional analytic applications running concurrently on the Oracle Essbase OLAP server.
Oracle has also upgraded certain software components of the Exalytics foundational platform, notably: Oracle Business Intelligence Foundation Suite 220.127.116.11 (over 200 enhancements have been made), and simplified integration with the Oracle Endeca Information Discovery 3.0 software and Oracle BI Mobile HD solution that enables blended analysis of unstructured social and mobile data sources.
Significantly, many of the Oracle Hyperion Enterprise Performance Management 18.104.22.168 suite components have also been optimized and certified for X3-4 (including Planning, Profitability and Cost Management, Spend Classification, and Indirect Spend Planning), as have all the Oracle BI Applications. The tighter integration with these software assets should remove the need for customers to re-code these applications to benefit fully from Exalytics’ memory-driven caching and accelerated processing speed.
Exalytics promises to bring greater agility to financial planning and other EPM tasks
Customer references are rapidly emerging around Exalytics, each typically offering an impressive story about performance gains and hardware consolidation. One immediate opportunity for Exalytics is to enhance business planning, scenario modeling, and what-if analysis.
These are all primarily read-write analytic applications, which reference and change/modify certain data elements in the analytic model. As such, they benefit significantly from greater flash-driven in-memory performance. It’s no coincidence that the “sweet spot” of Applix – one of the earliest in-memory OLAP engines, and now part of the IBM Cognos suite – was financial planning, budgeting, and forecasting solutions.
Ovum believes that advanced EPM applications represent the “low-hanging fruit” for Exalytics right now. Oracle does as well, promising 3–10× faster performance gains on its Hyperion planning and budgeting applications. Areas ripe for benefit are driver-based financial planning and budgeting, and forecasting – through the ability to rapidly (re)build different models and (re)run financial scenarios and simulations. Having the agility to perform these tasks on a more fluid, rolling basis, with constant reviews and revisions, drives more accurate and precise plans, budgets, and forecasts.
In-memory analytics holds considerable potential across an array of business applications and industries: helping retailers to quickly simulate the expected results of changing price levels; performing simulations in the oil & gas industry in finding and completing oil field drilling and production; and insurance companies examining potential costs of catastrophic events to ensure they are setting premiums at the appropriate level to cover the expected risk. This is reflected in a growing number of use cases that are now emerging around Exalytics, especially when paired with other engineered Oracle analytics platforms.
Customers should think carefully about how they can realize the benefits
Clearly technology advancements are making speed-of -thought analytics a reality in many business scenarios. And the benefits on in-memory – quicker decisions, accelerated workload processing, and a reduction in IT costs – are becoming clear to the business. But realizing these benefits is more than just an exercise in deploying a high-performance analytic platform.
It also requires organizations to ensure an equally agile data integration infrastructure, and most importantly to revise and re-engineer their operational and customer-facing business processes to keep pace. The overall aim should be to provide “right-time” (as opposed to simply realtime) analytics that meet real and pressing business problems.
Driving analytics-driven business transformation will not happen overnight, and companies need to carefully formulate implementation strategies that seize business opportunities from ultra-responsive processing platforms such as Exalytics.
- » A guide to computational storage: Boosting performance for SSD storage arrays
- » The top 10 cybersecurity companies to watch in 2020: How AI and ML is a key differentiator
- » Snowflake secures $479m in latest funding round alongside key Salesforce partnership
- » How cloud software companies can cut costs – without impacting on quality
- » Don’t forget to budget for business objectives to gain digital transformation ROI