By: Helena Schwenk, Principal Analyst, MWD Advisors
Published: 4th October 2011
This work is licensed under a Creative Commons License
Oracle started the Openworld conference this week by announcing the launch of Oracle Exalytics BI machine, an in-memory analytics appliance. Exalytics marks the next permutation in a line of appliances that includes Exadata for data warehousing and Exalogic for cloud computing. But it also marks Oracle’ s biggest foray into in-memory analytics designed to significantly improve the performance, scale and speed of the company’s reporting, analysis and planning applications.
Exalytics is not only about in-memory but visual data discovery too
Exalytics is a pre-tested, pre-tuned and pre-wired hardware and software configuration that packages together an optimised version of Oracle BI Foundation, comprising OBIEE and Essbase, with in-memory processing capabilities courtesy of the TimesTen in-memory database and an adaptive in-memory cache. It manages its high performance and scale through a parallel configuration of servers, networks and storage. Behind the scenes it also runs parallelised versions of the TimesTen in-memory and Essbase databases that run in Exalytics’ one terabyte of main memory, although when Exalytics compression capabilities kick in this can increase between 5 or 10 terabytes of data.
Put together, these capabilities result in a faster and more responsive BI, analysis and planning environment. In its own tests Oracle confirmed that OBIEE on an Oracle database runs around 18 times faster (compared with a standard installation) this in turn increases to a 23 times improvement when used in conjunction with Exadata, the company’s data warehousing system; an optimised version of Essbase running on Exalytics results in a 16 times improvements in response times with the company’s Enterprise Performance Management planning application showing a five times improvement.
However, Exalytics doesn’t just provide in-memory versions of its existing BI and Essbase products; it has—rather wisely—included new analysis capabilities that take advantage of the different, faster and more exploratory nature of in-memory analysis. It has, for example, extended and optimised the existing UI to support faster interactivity and new visualisation capabilities, including a range of interactive chart types, maps, and tabular formats for spotting trends and outliers in large, complex data sets. Likewise it has a presentation suggestion engine (PSE) that provides recommendations on type of visualisations to use to best represent a data set. This is clearly Oracle’s attempt to muscle in on QlikTech and Tableau’s territory whose visualisation capabilities are a firm favourite with line-of-business users looking for a more intuitive and interactive analysis experience. That said, given that Exalytics represents a more heavyweight, industrial strength (and possibly higher cost) approach to visualisation means it’s more likely to appeal to large enterprises, whereas QlikTech and Tableau also have a presence in the mid market.
Exalytics is targeted at three main segments of install base
Given its speed, performance and scalability, the company envisages a number of different usage scenarios for Exalytics which all point to its compatibility with existing Oracle BI environments, which the company claims can run without change. For example, Exalytics can be used as a lightning fast engine for Oracle’s BI reporting and analytic suite with these benefits amplifying when its pre-integration with Exadata is put into effect. There are currently around 1,000 Exadata machines on customer sites today so cross-selling within this customer base remains an obvious target for Oracle. Together both are designed to enable a faster and finer grain of analysis for key BI tasks including operational and financial reporting, multidimensional modeling, data warehouse analysis and so on.
Likewise Exalytics can act as a high performance platform for financial and operational planning for those applications utilising Essbase. The idea here is to reduce planning cycles and increase planning accuracy by allowing a deeper level of operational detail to be included within the plan and by enabling users to perform more what-if scenario modelling iterations in the same budgeting window. This seems like a smart move for Oracle, especially since streamlining the close process is still one of the key challenges for CFOs, but a potential sweetspot for Essbase.
Equally, Exalytics is also seen as a high performance platform for running pre-packaged operational reports and analytics for Oracle’s CRM, ERP and Fusion apps. The company believes the breadth of functionality and the level of performance provided by Exalytics can enable a new or improved type of analytic application such as pricing optimisation and yield management, although this is more of a future aim for Exalytics.
Exalytics looks set to ignite the competition but needs more customer validation
The decreasing cost of memory coupled with advances in hardware are making in-memory a more realistic proposition for many, as witnessed by the success of players such as QlikTech and Tableau. But in truth we believe Exalytics represents a more competitive move against SAP HANA, the company’s in-memory appliance launched at the end of last year. Both companies compete in the BI, analytics, performance management and business applications market and see in-memory as an enabling technology for boosting both the performance and sales of these products. One key difference however between Oracle and SAP’s in-memory offering is in their approach to hardware, with Oracle selling its Exa* machines on its own Sun hardware, whereas SAP partners with a number of players.
But let’s not forget that IBM will also compete with Exalytics. The company has its own in-memory capabilities courtesy of TM1 and has a similar—although not in-memory enabled—offering called Smart Analytics System, an appliance that packages an optimised version of its data warehouse and analytics software on its hardware platform. The company is also rounding out its data warehousing appliance capabilities based on its acquisition of Netezza to compete against Exadata.
With so much focus in the market on in-memory and appliances it is easy to get lost in the hype and technobabble around “speeds and feeds”. However, organisations looking towards Exalytics as a way to super charge their existing BI, analytics and financial planning applications should continue to press Oracle about the real world benefits of better query responsiveness, analytic speed and faster decision making. This, we believe, is best demonstrated through active customer references, something that has been missing from Oracle’s story so far.
We look forward to hearing more from the Exalytics customer base and following its progress with interest.
We automatically stop accepting comments 180 days after a post is published. If you would like to know more about this subject, please contact us and we'll try to help.
Published by: electronicdawn Ltd.