As the world's largest pure online bank, Egg plc provides banking, insurance, investments and mortgages to more than 3 million customers through its Internet site and other distribution channels. Launched in 1998, Egg pioneered online banking in the UK and also helps its customers understand and manage their money more effectively.

Sun Microsystems has played a large role in Egg's success, beginning with Egg's online infrastructure and most recently with Egg's Customer Data Warehouse.

The Foundation

In 1998, Egg selected Sun technologies as the foundation of a Web infrastructure to enhance security and improve performance so that they could provide outstanding service to their customers. Egg required an infrastructure that was extremely reliable, scalable, secure and able to support the company's more than 2.5 million transactions per day. Sun ably met those demands. The workhorses for this traffic-heavy Web site are high-end Sun Fire V880 servers running the Solaris Operating System. Sun Java System Application Server software runs the application layer of Egg's Web site.

With approximately 85 percent of Egg's customer interaction conducted via the Internet, Egg could take no chances with product selection. To continue to provide services to millions of users over the Internet, Egg required an environment that would provide a secure and compelling online user experience. The hugely secure Sun environment is tremendously efficient and barely breaks sweat, even with more than 250,000 visitors a day.

The combination of Sun hardware and software has provided Egg with enhanced security, tremendous efficiency, scalability and confidence that Egg's customers will have favorable experience when visiting www.egg.com. But, that's only the beginning of the Sun story at Egg.

Building on the Success of Egg's Sun Web Infrastructure

Given the great success achieved by Egg with their Sun Web infrastructure, it was natural for Sun to be top of mind again in 2001 when Egg was looking to solve a data latency problem.

Because the interaction between Egg and its customers is almost exclusively an online experience, the offering of products and services to customers must hit the mark the first time. Egg must offer the correct mix of products and services to the appropriate customers at the appropriate times. Originally, Egg was able to obtain the customer information required to do this by outsourcing the company's sales and marketing data warehouse to Experian. However, the latency between the outsourcing and the provision of data eventually became an issue, and Egg was forced to take action. Jay Parmar, Head of Data, Customer Technology Team at Egg, elaborates, "In 2001, as Egg expanded and our requirements for marketing also expanded, we made the decision to build our own data warehouse using Sun, Oracle and SAS." The goal for the data warehouse was to provide Egg with a single source of truth regarding their customers. Egg continued its relationship with Experian while the new data warehouse was being designed, built and tested.

The Sun Environment

Because Sun was Egg's preferred choice for the heavy, industrial-strength data services and industrial-strength dot-com services that Egg provides, Sun was also the obvious choice for Egg's Customer Data Warehouse (CDW). "In terms of data and specifically our Customer Data Warehouse, the strategy that I own and drive is also centered around Solaris and Sun," says Parmar.

Parmar's data warehouse team deployed the first iteration of the Customer Data Warehouse on a Sun Fire 6800 Server. "We specifically chose the Sun Fire 6800 Server because we wanted the ability to have an ETL [extract, transform and load] box, an end-user layer box and a test environment all contained within one environment - just partitioned into domains. However, as the service provided by our CDW expanded and the use of our system grew beyond our expectations, the Sun Fire 6800 was fully expanded, and we found that the hardware infrastructure was not big enough to cover all the services that we wanted to provide, so we purchased a Sun Fire 15K," notes Parmar.

CDW Architecture

Egg's Customer Data Warehouse now resides on a Sun Fire 15K with 16 CPUs and approximately 10 terabytes of storage. "That is essentially our core system," says Parmar, "and all of the data feeds into that system. Our storage application is an EMC SAN," he adds.

The system runs on Solaris 9 operating system, and the database is currently Oracle9i. Parmar notes, however, that Egg is considering version upgrades. "We're looking very closely at Solaris 10 and Oracle 10g. We want to take advantage of Solaris 10 features such as virtual domaining, and we are currently prototyping our data warehouse to use virtual domaining. We also want to explore some of the new features of Oracle 10g such as grid computing," he adds.

Egg internal customers use a variety of tools. "Once we've presented the data to our customers within Egg, they can use basically whatever technologies they choose for their particular needs. SAS is our major end-user tool. After presenting the data in Oracle, SAS is used to extract, join and mine the data for either credit decisioning or campaign management. SAS BASE, SAS Connect, SAS Share, SAS SPDS and SAS Stat are used, along with a tool written in SAS called Comms Builder. We use Oracle Discoverer as well," comments Parmar.

Expanded Usage

During the course of going live, Egg found other people within the business wanted to use the data warehouse service as well. Therefore, while they call their creation a CDW, it is really more of a service that serves data to customers within Egg. "Thus," notes Parmar, "we provide transactional services where necessary, and we provide data warehouse and data mining services where necessary."

CDW Functionality

Egg takes data for running the bank from a variety of sources - internal and external - to populate the Customer Data Warehouse. Thus, if a person has a credit card, a loan and an insurance service with Egg, that person is seen as one entity having three Egg products. "The major benefit Egg has received from the Customer Data Warehouse," notes Parmar, "is knowing every way that each customer is involved with Egg. This enables us to do things such as insight analysis to determine the propensity of a customer to buy additional products so we can offer that customer the right products at the right time."

According to Parmar, the Customer Data Warehouse provides great benefit to the company. "Within sales and marketing, we generate nearly 120 campaign selections per month using our CDW. It also provides a daily credit decisioning capability to manage risk. For instance, if someone has an Egg credit card, we are actually able to make a decision using the data in the data warehouse on a transactional basis to determine whether that person's account requires collection, is just slightly overdue or, conversely, that the person is a good credit risk and should be offered a pre-approved loan. Our CDW is also used to input some of the data into the risk modules for our Basel II reporting."

Egg's Customer Data Warehouse is currently just under two terabytes and growing by approximately 10 gigabytes each month. The campaign managers, creative teams and phone sales teams use the data within the CDW and targeting system to generate incremental sales. All campaigns are run on Sun Solaris systems, specifically on a domain on the Sun Fire 15K. "If we didn't have this CDW," notes Parmar, "we couldn't do the 120 campaigns per month across six channels that we currently conduct."

"In terms of our customers in credit," Parmar adds, "all credit decisioning is run on Sun Solaris systems, with a number of Sun Fire V440 servers. In terms of our customers in finance," says Parmar, "the CDW is used as a key system for customer numbers reporting. Without the CDW, we would have a lot more work to do to report to the city and the regulator. The CDW also supports the generation of the Basel II models."

Internally and Externally Secure

Egg has implemented a number of security controls to ensure that data warehouse access conforms to internal policy as well as external regulatory requirements. The Egg data warehouse team is responsible for protecting the data in the data warehouse. The team consists of a data warehouse manager and ten other members - specifically Oracle developers, Oracle DBAs and meta data analysts.

Only the data warehouse team can make changes to the code of the data warehouse, emphasizes Parmar. "Additionally, DUG - the Data Users' Group - is the only group within Egg that can determine what changes are made to systems and what accesses are given. Our data warehouse manager makes sure that Egg is in compliance with the various regulations set forth by the Data Protection Act of 1998, the Financial Services Authority within the UK and Section 39 of the Banking Act in the UK. Our extended team keeps the lights on to Oracle, Solaris and the UNIX systems," he adds.

In Real Time

Egg's Customer Data Warehouse is currently refreshed in near real time. "If someone visits Egg.com, applies for credit with us, is accepted for credit and issued a card, by the time the card is issued, we would see that person in our data warehouse," Parmar explains.

Currently, data is cleansed, matched and used to populate the core data warehouse. Then, data marts are published from the CDW for Egg's internal departments as required. The BED (back-end decisioning credit data mart) is refreshed every day, as is the Basel II data mart. The financing materialized views are refreshed every day for financial reporting and counting of customer numbers, and the marketing data mart is refreshed three times every week.

While Sun infrastructure enables a real-time environment, there are, of course, some exceptions to having all of the data in real time. "One of our biggest issues," explains Parmar, "is orchestrating all of the third-party arrivals of data. Third parties send the data at their own pace, and their processes are such that you have a service-level agreement that the data will arrive at a certain time. If, for whatever reason, we don't get that data, then obviously we have some latency within the system. Back in the days of Experian, for instance, we were happy with our six-week latency on data, and we were doing marketing campaigns using data that was six weeks behind. We're now doing marketing campaigns using data that is one week behind, but we have the capability to conduct daily campaigning and that data is only 24 hours behind. But, even that is sometimes not good enough. Egg's internal customers are asking for the data in 12 hours, 6 hours or even 4 hours. What I've done is drawn a line in the sand and said that we need to invest in our applications architecture to actually provide data in real time because from the moment that we go into real time, we get into the realms of actually being a corporate data warehouse. Potentially, we're a corporate operational data store, so we're having those conversations now," says Parmar.

Egg's Customers Benefit Too

The CDW is now an integral and valuable asset for Egg, yet Parmar points out that Egg customers also receive value from the Customer Data Warehouse. "If we didn't have a data warehouse, we would be severely restricted in the service we provide to customers," notes Parmar. "At Egg, we take the approach of caring about our customers and helping them avoid financial difficulties. The data warehouse plays quite a strong part in providing the information and data necessary to help our customers make an informed choice on their finances. Outwardly, the customers don't know that we use our CDW to help them, but there's no question of its positive impact on our external customers."

System Reliability

"We have an infrastructure and environment where we can cover any failures with hot pluggable CPUs and multiple failsafes on the SAN. We are actually working to also provide a disaster recovery system. We are not going to throw away the Sun Fire 6800 server that originally housed our data warehouse. Rather, we are going to architect it to handle our disaster recovery via our sister site in Dudley," Parmar says.

"Basically we are looking at 90% uptime for our system due to scheduled downtimes. In all the time I've been working with the Sun hardware in the last three years, I don't believe that the system has ever gone down. We have never had a failure on the Sun hardware that warranted anything to be escalated above and beyond the engineers," he adds.

Sun Satisfaction

Egg's Sun-powered Web infrastructure is more than capable of steadfastly handling the heavy traffic at Egg, and visitors report their experiences with Egg to be extremely favorable. As noted in Egg's 2004 Annual Report, even though the company is only seven years old, nearly three-quarters of UK consumers now report that they would consider buying a financial product from Egg.

"In terms of enabling our data warehousing services, what we've found is that our choice of Sun as the provider of the service has been very good," states Parmar. "We were able to take the success of our existing Sun-centric Web infrastructure and parlay it into our invaluable Customer Data Warehouse. We already had a Sun engineer on site, 24x7 support contracts in place, and the experience gained through our large bandwidth usage and industrial-strength applications powered by Sun. We hopped on to the back of that, and we've not regretted it at all."

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access