Infosys’ blog on industry solutions, trends, business process transformation and global implementation in Oracle.

« December 2011 | Main | February 2012 »

January 30, 2012

Oracle's Big Data Entry - Big Data Appliance

We recently heard and have been talking a lot about Exalytics "The speed of thought analytics" machine which has in-Memory features and is tuned for extreme performance. The question arises is this machine the mother of all solutions in analytics space, or there's more coming?

Firstly Exalytics is an analytical machine with in-Memory (TimeTen Database) capabilities, advanced visualizations, Columnar compression and hardware tuned for extreme performance. Depending on your appetite towards analytics, you can scale this machine to your needs and if still not satisfied Oracle gives you even better option to join it with your Exadata machine for even better peformance on the Oracle Database side of it with Exadata being tuned for RDBMS performance. Yes certainly it will be seperate license cost (separate SKU's) that you will have to shelve out.

Recently Oracle announced their entry into the Big Data World with Oracle Big Data Appliance which includes
1. An open source distribution of Apache Hadoop
2. Oracle NoSQL Database
3. Oracle Data Integrator Application Adapter for Hadoop
4. Oracle Loader for Hadoop
5. An open source distribution of R - for Big Data analysis

There's need for some bit of explaining about when and how those 3 machines will be used, as there overlaps in terms of capabilities like Cache management, in-memory database:
1. Oracle's design of those 3 appliances (4 actually if I include Exalogic too) is with idea of division of labour, distrubute analytical workload across appliance machines. The performance gains on analytics are possible only when caching is effective, & only if the data is in Exalytics DRAM the true performance gains can be realized
2. The hardware lock-in with Sun servers that Oracle's pushing for could be another decision factor for customers
3. The right use-cases of leveraging those 3 (or 4) boxes needs to be seen
4. One additional point to explore and ponder over would be how other vendor based tools would interface or fit with Exalytics - if that's a possibility

What is it competing against - SAP HANA (already in market for over a year: provides flexibility to choose their hardware), and IBM buying our Netezza in September 2010 building their integrated software-hardware appliances optimized for specific workloads (Smarter Systems initiative of IBM).

Interest space to watch how Oracle prices those appliances right, build right use-cases and agressive marketing strategy to penetrate with their in-memory TimeTen database. Keeping all doubts aside, end consumers are in for treat and sharpen their analytical capabilities with features which were never thought of possible before.

Will post more as we see the growth and progress of Exalytics and Oracle's Big Data Machine. Do share your experiences, challenges that you see in Big Data space.

January 24, 2012

Thinking beyond Hyperion Essbase for planning (Part 1)

Guest post by
Hari Ram, Senior Systems Engineer, Infosys


Oracle Hyperion Essbase provides the best of multidimensional database capabilities in terms of performance and features. The Essbase lovers and existing users may believe that Essbase is sufficient enough for their planning, budgeting and forecasting requirements. Hence most planning applications still exist on Hyperion Essbase instead of Hyperion Planning - the centralized MS Excel and Web-based planning, budgeting, and forecasting solution from Oracle's Hyperion Suite of Products.

Hyperion Planning & Essbase

Hyperion Planning uses Essbase as the database engine to store the numeric data and dimensions, while it needs a relational database for other repository needs. Hyperion Planning leverages Multidimensional functionalities and capabilities of Essbase to the fullest to provide varied planning options along with inherent Data Forms, MS Excel, Web based grid interface for the data entry and reporting needs. It also has Task lists to track various activities and their status during the Financial Planning process. The Process management feature provides an approval process independent of the Organizational hierarchy.

Why Hyperion Planning is not just {Essbase + Excel interface, Task, Approval.... features}?

Handling various types of planning

Typical Financial Planning systems would handle Expense, Revenue, Capex, workforce, Profit & Loss, Balance Sheet, Cash flow...etc. In planning systems built on Essbase, an application would have one database to handle each types of planning. Data, dimension member sharing between two databases would need additional development/maintenance effort.

While working with similar multi-planning type requirement with Hyperion Planning, I found that the idea of plan type reduces the complexity of the solution, user-appraisal when compared to a Standalone Essbase application. Hyperion planning's Plan Types treated as individual databases in underlying Essbase database comes with inherent facility to share data and dimension members across plan types in Hyperion Planning or equivalent databases in underlying Essbase. This reduces the need for additional development/maintenance effort when compared to Planning systems based on Standalone Essbase. The Hyperion Planning user would seldom realize that the dimension members belong to multiple databases. While this might be an area of concern while using Standalone Essbase to build planning system.

  • Multi-Currency application
    As a developer, I find it quite taxing to work with multi-currency applications built on Standalone Essbase. This requires a Currency database, few defined dimensions and also mapping between members in the Currency database with members in the planning database.

Designing a multi-currency application using Hyperion Planning reduces the developer's effort and complexity of thought. The pre-built Currency dimension, Currency Conversion scripts, Exchange rate table that may be specific to a Scenario and Version (allowing one to have different exchange rate for the same year for different scenario). The User interface for Exchange Rate Table makes it so convenient for the Financial Planners & Analyst to work with the system.

  • Varying attributes, Text/date measures
    Something that excited me about the version 11 of Essbase is "Varying attributes". The 'Varying Attributes' feature in Essbase provides a solution to situations, where the relationship between two dimensions varies with an independent dimension. Example, member of Salesman dimension changes with Period, while tracking Sales. Hyperion Planning doesn't support Varying attributes, but such cases can be handled using a flag member.

Another addition in the newer versions of Essbase is Text and Date Measures, in order to track textual values or measure (say Customer Feedback: 'Good', 'Average' and 'Bad'). Initially it sounded like an imitation of Hyperion Planning's Smart View to handle textual data, but Smart List still remains the favorite. Practically, most of the times, I came across cases where we need to capture textual information for more than one dimension (Accounts, Entity...Products, etc..,). Smart List in Hyperion Planning, which can be used with multiple dimensions of any type, scores higher than the Text/Date Measure in Essbase, which can be used with only one dimension.

  • Data Forms
    Financial Analysts and Planners, I worked with, always like to have set of standard data entry/reporting spreadsheet that complements MS Excel (though it has 'rigidity' as a freebie). We spend lot of time fixing issues with the data validation at individual cells, incorporated either by formula or VBA. But the validation and guiding rules would change with the dynamics of the market. Lack of central maintenance of standard spreadsheet, the need to circulate the revised spreadsheet carrying the newer validation rules, throws up a challenge to the efficiency of Change Controller.

Data Forms in Hyperion planning provides a standard centralized grid-based interface for data entry and data-checks. Customizing or fixing a Data Form issues take lesser time and effort compared to the Excel based spreadsheets. It also provides option to have textual comments, annotations, a built in calculator (Supporting Detail), various data allocation and spreading options like Grid Spread, Adjust, Mass Allocate, etc..,) to handle the daily needs of a Financial Planner &Analyst.

In this blog, I tried to share my view on the basic differentiators between Essbase & Hyperion Planning based Financial Planning systems. In Part-2, I will be sharing my thoughts on Data and Process management aspects.

Subscribe to this blog's feed

Follow us on

Blogger Profiles

Infosys on Twitter