Oracle Analytics Cloud (OAC): Comprehensive Analytics platform on cloud with more control

As we are eagerly waiting for this year's Oracle Open World for new announcement in Analytics space, it is worth to recap what has happened in the last year.  As per its pledge during last open world Oracle released its new generation comprehensive PaaS version (Platform as a Service) of Analytics with options to configure and manage the instance by customer. I got a chance to build a POC for one of our customers using Oracle Analytics cloud and would like to share my first-hand experience through this blog.

Product is available in below two versions with mentioned components in it.

1)      Standard Edition :  Data Visualization + Essbase (Standard Edition) + Data Visualization desktop

2)      Enterprise Edition : Data Visualization + Modelling & Enterprise Reporting  (BICS)+ Essbase (Enterprise Edition) + Data Visualization desktop + Day by Day mobile App

       Fig1: Schematic diagram of Oracle Analytics Cloud (OAC) 

As we can see from the product offering options, OAC is like an OBIEE12c  version available on cloud with Data Visualization + Essbase with enhanced  features.Standard version is more suitable for Departmental Data Analysis /visualization needs whereas Enterprise version is a complete package to meet Organization wide Analytics needs. I am sharing the key differentiating features here, which I feel more appealing for the existing OBIEE customers and for the prospects.

Control over configuration and maintenance

OAC provides complete control over the sizing of their cloud environment and as well as the maintenance part of it.  OAC licensing is unbundled compared to "Business Intelligence Cloud Service" or "Data Visualization Cloud Service", customer need to buy a separate IaaS (Infrastructure as a Service) and Oracle Database cloud service. It means, customers can choose the number of OCPUs-RAM combination depending on their requirement.

We can now connect to server through SSH and perform required configuration file changes as similar to on-premises environment.  We can also decide when to apply patches and carry out maintenance activity.

Complete Application Life cycle management activities such as starting/stopping, provisioning of services, backup can be managed either through front end or via REST API based scripting. We can control the OCPUs allocation to OBIEE, DVCS and Essbase.

Customers can bring their own on premises DB license and host data warehouse on cloud using Oracle Database cloud or a buy a new license. Most fascinating feature of OAC is that it can connect to on-premises database through VPN connection.So one can get the benefits of cloud Analytics without any need for changing their ETL/ Data warehouse architecture. 

Flexible Pricing

As stated earlier, to run OAC, customers need to buy additional IaaS and database license as per their convenience. Licensing of OAC platform itself is based on number of OCPU-RAM combination instead of user based licensing as in the case of SaaS (Software as a Service) cloud products. Customers can thus provide access to unlimited number of users similar to on-premises environment with appropriate IaaS and Database sizing (for DVD there is a cap of 50 users/OCPU).

OAC can be subscribed either through metered or as non-metered OCPU license. Metered license itself has got Monthly/Annual hour rates or Pay as You Go option. Customers can now run their environment when it is actually needed and turn it off when it is not. Non metered license can be bought on monthly subscription rates, and it is more suited for global organizations working across multiple time zones. OAC thus provides flexibility to scale up/descale the capacity as per customer's need with no lead-time.  For more details, please refer Oracle web site. 

Essbase on cloud

OAC does come with Essbase cloud component which is mostly tailored for what-if or MOLAP analysis kind of needs rather than like a standalone, full-fledged on premises Essbase Application. It has a cube designer Excel add-in along with pre-defined Excel workbook templates enabling single step cube creation and easy data loading. One will still miss the control and management option available through Essbase Admin Console (EAS).  We need to define data load rules in the templates and there is no other way to change them. Just to assure Smartview Excel add-in is still available for reporting on cubes.  I strongly feel Oracle will enhance this service in upcoming OAC releases.

Easy Data Modelling and Deployment

We can either use BICS data modeler for simple data modelling needs or we can use OBIEE12c Admin Tool to build and manage more complex RPD models. Access to RPD through Admin Tool provides full-fledged design capabilities similar to that of on premises environment.

Customers can easily archive catalog, Security Model and RPD in the BAR (BI Application Archive) file and migrate to another environment in a single click. Similarly the RPD can be deployed from one environment to another. OAC thus provides seamless migration capability between cloud instances and also from on-premises to cloud or vice versa.

Ability to prepare data by end users easily

OAC offers enhanced out of the box connectors to 25+ on premises and cloud databases /SaaS applications in DVCS. It is more helpful in case of SaaS applications as it provides direct access to business entities of those applications (for e.g. Incidents, Organization in RightNow/Service Cloud).We can expect more new additions to the current list in upcoming releases.

Fig2: OOB Data souces and Data flow in Oracle Analytics Cloud (OAC)

User can also create light ETL operation such as joining data sets, concatenating columns, aggregations etc. on the DVCS. It will greatly help end users in their data preparation needs reducing dependency on IT.

Extensive Visualization and Advanced Analytics Capability

OAC comes with added list of pallets and charts such as Sankey, Chord Diagram and Network etc. in the DVCS. Users also get single click access to add advanced analytics features such as Clusters, Outliers, Reference, Trend Lines and Forecasts directly on the graph/charts.

Fig3: Visualizations available on OAC

Most promising feature of OAC is availability of Advanced Analytics functions such as Regression, Cluster, Outlier and Trend Line as out of the box for creating calculated fields. We can now embed custom R Script or Python script leveraging Evaluate_Script function, which opens gateway to machine learning, advanced Statistical modelling, text analysis etc.

Missing dots

I observed few important gaps in current version of OAC which temporarily stops it to meet overall BI and Analytics needs of customers, though Oracle has marked some of them on their roadmap to be available by next year or so.

• Business Intelligence Publisher (BIP): As we know BIP is an integral part of Oracle BI offerings and it's an industry leading product for pixel perfect document reporting and its currently missing in OAC. For my current POC exercise, I had to install it on Java Cloud Service by following a complex installation process.  
• Delivers: OAC does not offer full delivery feature such as attachments in different formats, it does offer the report/analysis content to be deliver in the email body itself.
• Security: OAC does not provide options in terms of configuring security such as enterprise LADP integration or configuring a custom credential provider.
• Action Framework: OAC only provides within catalog and URL based navigation actions. Customer do expect JavaScript Action, EJB Action, Web service based Actions etc.
• Customization on VM: Currently there is no mechanism to retain the custom configuration changes on the VM, which gets overwritten during patching /upgrade.
• Custom Map Layer in Answer: Users cannot add custom map layers in the Answers though same can be added in DVCS. 

I have not covered one of the very exciting features of OAC, Oracle Day by Day mobile App. I will cover the same in my next blog as it requires more details.