While companies are struggling with the implementation of Big Data Architecture, all software vendors are bombarding the CIO with Cloud, Cloud and Cloud. CIO still haven’t found enough Business Cases and Business Values in implementing Predictive Models and they are facing a new question on the Cloud.

SDG Augmented Business Architecture [ABA] has been designed to highlight the benefits obtained by implementing the most innovative digital platform and should enable our customer to gain the expected benefit.

The Traditional Bi Architecture is not able to satisfy the following requirements;

 

Real-time Analytics

Real Time Analytics does not have the same meaning of Real Time Integration, quite often companies are asking why they should have a Real Time Analytics Front End if they have already adopted some Change Data Capture (CDC) Components. The role CDC plays is to integrate data when they are changed, rather than an ETL Logic, but traditionally the CDC technology does not have rich transformation libraries, so without a Real Time Analytics Front End, Infrastructures have to ingest data and then analyse them, resulting in the storage of a big amount of data without any business use.

Real-time Analytics are solutions that would allow companies to Analyse Data before storing them, and would allow to store only useful data.

 

Real Time Ingestion

Traditional Data Base DWH is not able to ingest high volumes of data at a high frequency in an agile manner. Massive Parallel Processing (MPP) infrastructures are designed to be an excellent High Performance Database for fast and performing queries, however they risk to be completely stuck with generated SQL for frequent transactions.

Having a file-based PSA would allow companies to ingest real time data in an Analytic Environment.

 

External or Unstructured Data

MPP and Database DWH are designed to manipulate highly structured data coming from the various Source Systems. However, companies are starting to have the need to integrate internal data with documents coming from external sources that not necessarily come from a database. Having a file-based PSA would allow companies to integrate and Analyse any typology of Data.

 

Scalability

Traditional DWH does scale vertically and not horizontally, meaning that is not easy to add computing or storage power to an existing Analytical Framework. Companies are requesting to increase Frequency of Data Loading, to increase the level of detail of the Analysis, to keep a longer history of data, to integrate new companies within the DWH, and Traditional DWH Architecture would require a relevant infrastructure intervention before enabling users to meet the mentioned expectations.

 

Predictive

Traditional BI Framework, even if Software Vendors declare to have specific tools or declare R integration or In-Data Base Analytics, they do not enable companies to deploy a flexible and powerful Predictive Modelling Solutions. These Models require computing power, long history of data, detailed data that would only be available if the companies do enrich their architecture with a file-based PSA.

 

Self Service Analytics

Big data Framework does need to provide Analytics to a high level of detail of information, without restricting users to adopt Ad-Hoc Analytics Software with prebuilt reports, but still needs to provide users only with the relevant data to the their processes and business, not with “every column of any table of any source system”. In addition, If you think of doing Self Service BI when you provide a Bi tool dataset on top of an MPP you will see that the Architecture will fail and collapse and Users will start saying that the data do not match.

Users of an SDG Scalable Architecture, in addition to the PSA and HPD, would guarantee excellent performance to Self Service analysis.