Insurance and asset management services company Momentum has implemented technology from Sybase and IBM to create a data warehouse for its group benefits division, Momentum BenefitsAtWork.
The company is a provider of employee benefit products and services, including retirement, risk and investment benefits, and administrative and consulting services.
The new application - developed and rolled out over a period of three years by a team which included Sybase, Alacrity and BI in the Box - has integrated data across the business unit into a single repository that is used by 70 Momentum business intelligence consumers.
To provide a holistic view of the business, data is sourced from more than 18 systems, hosted on eight different technology platforms, into one data warehouse.
“When users saw the system for the first time, they were amazed at being able to view their world in one place and obtain all the information they wanted with one click and within seconds,” says Steve Farmer, BI Practice manager at Sybase Cape Town. “Prior to the rollout of the data warehouse, users would have had to go to several disparate systems, manually extract the data and then manually combine it to achieve the same view.”
Karen van der Walt, a BI architect with software development company Alacrity, says the project was initiated because the business unit needed to obtain accurate, relevant and timeous information. “Data was spread across the business, in various systems and databases,” she says. “Business-critical information was not being stored in a structured, accessible manner, and compilation of essential reports was tedious, time-consuming and error-prone due to manual processes.”
Van der Walt says the emphasis of the project from the outset was on the underlying data linkages instead of the generation of a multitude of seemingly isolated reports. There was a general understanding that unless the data from the many disparate systems were transformed into meaningful information, no reports would be reliable.
“Some of the primary users of the system are actuaries,” Van der Walt notes. “Their job is to interpret numerical data and use mathematical and statistical methods to assess risk in the business. Prior to implementation, they were spending their time collecting data and putting it together, leaving little time to analyse it. From that point of view, the data warehouse has transformed the business whereby it provides immediate answers to the 'who, what, where, why and how' of the business. Another major user is the finance area. With millions of financial transactions processed every month, it was difficult for these users to perform manual processing on the data and to search for possible duplicate transactions. Now, they too have this information available centrally.”
Solution structure
The solution comprises Sybase IQ 12.7, an analytics server designed for mission-critical business intelligence, analytics and data warehousing solutions; IBM InfoSphere DataStage 8.1, an ETL solution that supports the collection, integration and transformation of large volumes of data, and manages data arriving in real-time as well as data received on a periodic or scheduled basis; and PowerDesigner 15, a modelling and metadata management solution for data architecture.
Van der Walt says Momentum's project team chose to work with Sybase for a number of reasons. “The fact that Sybase IQ is a high-performance data warehouse engine designed for analytics was very important, due to the complex nature of the analytics that would need to be delivered. IBM DataStage was chosen for BI integration because of its industry-leading status in the field of integration. We also wanted to work with a single vendor that could provide all the critical BI components. Sybase's BI Practice is an IBM Business Partner so, in addition to procuring Sybase IQ and receiving training and consulting services on Sybase IQ from Sybase BI Practice, Momentum also received training and consulting skills on IBM DataStage from Sybase. Sybase also provided us with Business Objects, which was chosen as the report delivery mechanism. This meant that Momentum had a one-stop shop for all the critical components of the BI architecture stack.”
Momentum had a number of expectations for the project, all of which were met, according to Van der Walt. “These included a well defined capital expenditure outlay for the set-up of the BI infrastructure and architecture with no runaway project expenses, a cost-efficient, scalable BI environment that would allow for future expansion plans, and a personal relationship with the vendor as this enables instant access to support people.”
Phased implementation
The implementation of the BI environment and delivery of the initial functionality took place in phases. Phase one was the proof of concept (POC) which took place in February 2006 and consisted of loading a substantive set of data into Sybase IQ and producing a number of complex financial reports. The goal was to prove the ability of Sybase IQ to load large volumes of data in an acceptable time and to perform complex analytics. The POC was conducted by Sybase consultants, with business knowledge assistance from Momentum BI team members.
“The results of the POC showed that Sybase IQ was perfect for Momentum's requirements,” says Farmer. “The first project started in September 2006 and the goal was to deliver a fully functional set of financial reports. At the same time, Sybase provided user training for development and admin staff on IBM DataStage and Sybase IQ. We also provided a team of consultants to assist with the implementation of the architecture and deliver the initial BI functionality.”
After the successful financial reporting implementation, a series of projects followed. These were implemented by a team that included Alacrity (project management and business analysis), BI in the Box (full technical implementation and support), and Sybase. Together the team delivered, among others, financial reports, profit analysis capabilities, productivity statistics, and transformation statistics.
“Implementing Sybase technologies was instrumental in our success as it helped us to focus on the business requirements without it becoming all about the technology,” says Van der Walt. “So often in this industry people rationalise their decisions based on technology constraints. With Sybase IQ, there are none.”
She cites the following advantages:
* All data can be stored as a result of Sybase IQ's data compression capabilities and data load performance.
* Different database types are possible, including data stores, transformed data stored according to the business model, and data marts. Again, this is a result of the technology's compression and data load performance.
*.Momentum does not have to choose a particular design purely for performance. Momentum has both relational and dimensional tables depending on the solution requirement. This is possible as a result of Sybase IQ's column-based indexing that allows fast access to all stored data.
* With Sybase IQ converting keys to integers (even when stored as a character), we are able to link data using meaningful foreign keys, instead of being forced to use arbitrary integer keys without a degradation in performance. This is important as Momentum sources data from many standalone systems where the update frequency and mode differs vastly.
* Sybase IQ facilitates the effective utilisation of resources. Momentum does not need a full-time data modeller or administrator.
Van der Walt has some words of advice for those about to embark on similar projects: “Ensure that you have the correct resources and that the people you work with understand data warehousing. Patience is key as it takes time to get all of the source data. However, once this has been accomplished, business quickly sees the results. Set up a methodology and a charter of best practices for each technology. Finally, when you are designing a system, make sure you do it with your eye on the future.”
Karen van der Walt, 021 658 0212, karen.vanderwalt@alacrity.co.za
Clint Vosloo, BI in the Box, 082 678 6263, clint@biitb.com
Share