I'm often asked what is the best way that a communications company can manage their OSS metadata?
There seems to be an infinite number of approaches that can be taken each with their own merits and shortcomings. Based on my experience, the answer of course is not straightforward because every situation is unique and there really is no "one size fits all" approach. Every communications company has a unique set of circumstances that may make one approach more effective than the other. The appropriateness of an approach almost entirely depends on the context in which the OSS platform is used and delivered. There are however a common set of factors that seem to drive a company to a specific approach. In my experience, the major drivers tend to be:
Maturity of implementation
Often, a particular metadata management approach is driven by where you are in within the software development lifecycle. A company that is currently in the build phase of implementing a metadata driven OSS may find that managing metadata in very granular datasets that represent how the system is being delivered is most effective. In other words, it likely makes sense for a company in early stages of implementing an OSS to manage their datasets so that development is optimized. This might mean managing it in a way that permits things like parallel development and maximizes re-use. It might also be logical for an OSS delivery team to manage datasets according to the Work Breakdown Structure (WBS) of their project in order to simplify reporting and tracking.
Every company has its own preferred software delivery approach and the way datasets are managed may be impacted by this. For example, companies that follow agile or iterative development methodologies may find it cumbersome to manage their metadata datasets using a fine grained approach. For them, it might simplify the metadata management process to group datasets in a small number of large versions or drops that line-up with build and test iterations.
A company may decide to manage metadata according to the user communities that use the OSS application. Often, communications providers are organized in business units that align with, for example, the network assets that they are responsible for managing or the business processes that they support. Each of these business units often use the same OSS in a different way and that may drive the metadata management approach OSS in much the same way. For example, I've seen cases where the same network device type (metadata) may be modeled slightly differently across different business units for different requirements and therefore require multiple versions of the same dataset. Or perhaps one group requires a detailed logical inventory for port and capacity management, whereas another wants a high level topology solution that may have common components with the former group.
IT organizations within communications companies often make investments in software tools that can influence the way they decide to manage metadata. For example, there may be automated testing applications or configuration management tools used that require that metadata be organized in a certain way in order to make effective use of the tools.
The way an OSS is deployed can also drive the need for a specifically tailored metadata management approach. A communications provider that manages their network inventory in a federated manner or has multiple instances of an OSS across their enterprise will certainly require attention to detail during the deployment planning and execution. It may make sense, in this case that metadata is organized in datasets that align with the production instances of the OSS, or requires a common model for all instances for part of the solution, and then technology-specific models that differ per instance.
Composite Delivery Structure
At large communications companies it is not uncommon for multiple software vendors and professional services organizations to support a single OSS. They are often responsible for managing disparate pieces of the OSS or delivering specific pieces of functionality to the OSS. The way metadata datasets are managed can impact the success and effectiveness of such a composite delivery structure and it is crucial that a metadata management approach fit well into this delivery structure.
In a future article I'll look at some specific metadata management approaches I've seen at communications companies and some of the pros and cons of those approaches.