Migrating legacy source code is a time-consuming and complicated business, as discussed in my previous article, "Saving Money with Legacy Source Code." The same is often true for the migration of legacy data, but as for the source code case, there are some useful techniques that can reduce the cost. The use of more modern tools (in this case, SAX and XSLT) can greatly assist in this process.
Legacy Data in General
Demand for new features, network devices, and technologies combine to drive the need for ongoing changes in network management technology. This demand translates into a steady stream of vendor updates to the many commercial network management system (NMS) applications. As code is changed, the legacy data must generally be either discarded (a bad practice) or migrated/upgraded (good practice).
Other industries (such as financial services) are beginning to make far more use of legacy customer data than before. One practice, known as customer lifetime management (CLM), is becoming the norm: Customers are grouped by value and then targeted for specific products and services. CLM aims to extract the maximum value from customers over their lifespan with the service provider.
Examples of customers with different CLM values include the following:
- A final year college student with substantial debts
- A retiree with a paid-up mortgage and a fixed future income
The student has no assets, but may have a higher lifetime value to the service provider than is the case for the retiree. CLM can be used to assist the financial service provider to decide which customers are more likely to consume additional products and services. Legacy data is the cornerstone of any such effort.
Getting CLM right requires a huge amount of data mining and analysis. It marks an interesting trend: Financial service providers work with their existing customers rather than exclusively trying to win new customers. Traditionally, the networking industry (including telecom) has not placed a great deal of financial value on legacy data. This may be about to change as we see the following important trends:
- Wafer-thin margins on network hardware and software
- Commoditization of network hardware
- Major hardware vendors building consultancy service divisions
Just as PCs have become commodities, the same is now happening with network hardware. Even the announcement of a new multimillion dollar router in 2004 did little to increase Cisco's share price. It seems as if the next wave in networking will be software—and service-related—as evidenced by the growing interest in vendors creating divisions devoted to consultancy and service.
As you'll see, legacy network management data is an important element in network service provisioning, so its preservation and migration is very important. Let's take a look at this now.