Moving to a Network Datawarehouse

By | April 15, 2011

Communications Service Providers(CSPs), have lots of OSS/BSS applications that are running to support their business and operations. Those applications do manipulate some kind of data based on the domains they address. Obviously, applications touching the same domains cause data duplication. They may also maintaining the same data with separate representations.When one domain’s data is updated by it’s manager application, the other applications’ data that touches the same domain should also be updated. Weakly designed business processes cause inconsistencies between data sources. Synchronization mechanisms that are built to solve these problems bring additional integration tax and maintenance resources.

The solution is to build a centralized place to maintain the domain data and share this data among the OSS/BSS applications. We call this “Master Data Management” or MDM and it deserves a separate topic to be explained in detail. In today’s topic however, I will discuss the benefits of such kind of a centralized scenario and possible practical implementations of it.

Most of the CSPs, already invested on datawarehouses in order to run some business analytics functions that will help them to offer campaigns that will increase their sales. Mainly stemmed from CRM perspective, these datawarehouses are maintained by IT departments. That’s why, they are also called IT Datawarehouse systems. IT Datawarehouses maintain transactional data (new product purchase, up-sell, cross-sell, ask for information, churn) and the customer data that include the demographics information. This datawarehouse shares it’s data with all the BSS systems (Billing, CRM, Campaign Management, Order Management). Unfortunately not all BSS applications are capable to use an external data model and sometimes the central data needs to be duplicated to the applications database via integration enabler like ESBs. Another approach is using this datawarehouse just for reporting purposes and not as an MDM.

Regardless of it’s usage type (as an MDM, and/or reporting solution), the missing part of IT datawarehouse is the lack of network related information. IT datawarehouses are fed by the CDR information to understand the usage patterns but this data gives information about the successful usage information and mainly useful for basic churn analysis and CRM campaigns.

However, apart from the CDRs (or xDRs) there is a huge data that is waiting to get analyzed. The network data include performance management KPIs, Faults, Signalling information, network facing trouble tickets, service activation and resource provisioning results, test results, SQM performance KQIs etc. This data is maintained by several OSS systems and these systems deliver advanced reporting and analytics capabilities in the domains that they address.

Is it possible to merge these data sources? There are multiple quick wins. First quick win would be a common reporting platform that allows the CSP to remove the reporting functionality from the application level. This will allow the CSP to get rid of the licensing of these reporting systems and their databases. Second, the maintenance cost would dramatically reduced. 3rd and the most important reason would be the cross-domain or cross-technology analytics that would enable the CSP to understand the usage patterns and combine these with the analytics data that reside on the IT Datawarehouse side. This will also support our vision to reach to 360 degrees customer view , where we understand the customer not only the services we are running.

A question comes up in mind again : Is it possible to utilize the current IT Datawarehouse system in place? The answer is: It depends. Network Datawarehouse systems utilize the Massive Parallel Computing (MPP) technology to cope with the huge amount of network data that is flowing into the system. Traditional datawarehouses cannot handle this load. So, it would be wise to invest on a separate platform without changing (and impacting the performance of) the IT Datawarehouse on hand.

A successful CEM strategy will need to include NW datawarehouse concept. Removing the boundaries between IT and the Network means Network Aware Business. Up to now, business business were asking network related data from the network/IT departments and the those departments were going to the related OSS Applications to grab some reports. After that, it was again technical team’s responsibility to translate these reports to the business language. This process is a waste of time for both business and the technical teams. However, with the help of a Network Datawarehouse platform, business people will be able to reach the data via advanced reporting in their “language”.

Network Aware Business will be in the center of the  next generation CSP architecture, which is  designing and delivering business processes that targets the positive customer experience at all times.

Category: CEM