Another way of tackling integration
The purpose of repurposing
Of course, there are a variety of different types of integration and there are a range of different things that you can do with data integration. I am not here to suggest that there is another way of tackling data integration in general. However, there are specific aspects of data integration (and in this context I am talking specifically about the data movement aspects of integration) where there may be an alternative that is worth considering.
In particular, think about a call centre application or similar CRM (customer relationship management) requirements, or even analogous supply chain functions. In most circumstances a customer calls up (or you call him or her) and what you need to know is how much they have spent over a period of time and what products they have bought. Based on that information you want to try to up-sell or cross-sell this customer.
So far, so easy. There are traditionally two ways to do this, especially bearing in mind that you may have multiple applications within which this customer data is stored. The first is that you put all the information into a data warehouse (or some other suitable storage) and extract the relevant information as and when needed. Alternatively, you can use federated queries that access the live database systems and bring all the relevant data together in real-time. Typically, preference is for the former because of the performance implications of using a federated approach. However, there is a third alternative.
The alternative is based on the fact that all the information that you need in the call centre application is contained within the customer's statements. If you could read the customer's statements from the archive in which they have been stored, and retrieve the information you need and combine it in appropriate ways, then you would fulfil the needs of the call centre.
Having described this potential approach, which is technically known as repurposing, it will come as no surprise that it is actually available from a number of vendors, though Xenos is perhaps the most well-known, and it certainly offers the best performance.
Basically, it can read any print format from either a document or report and convert it into more or less whatever you want, including XML, or you can present the data within a portal (either Xenos' own or a third party's). Of course, the software has the ability to access multiple content repositories and collate information across these. Moreover, it can index content such as statements to enable rapid retrieval.
There are two further points that we need to make. The first is that repurposing is not going to be suitable for all sorts, even of call centre, applications. For example, suppose you want to include such things as customer lifetime value within your up-sell or cross-sell calculations: you could not do this simply by repurposing statement data.
On the other hand, you can do quite complex calculations, such as "which customers have spent over $1,000 a month for the last three months on their credit card and have bought from PCWorld?".
In other words, if the data can be derived from the statement (or other printed form) then you can do it using repurposing. So, if you wanted to include customer lifetime value data you would need to create a report that (potentially) printed this out and then the repurposing software could use that information: except that that report might need to derive from the data warehouse, thereby defeating the object.
The final point is that this technology is great for delivering self-service capability. This is not an area where you would expect to see data integration vendors playing (perhaps they should?) and so Xenos and its competitors have a more or less open goal here, as the alternative is conventional development. Repurposing is hardly a technology on every IT manager's lips: maybe it should be.
Copyright © 2006, IT-Analysis.com
Sponsored: Hyper-scale data management