Network inventory systems sit at the heart of most company’s OSS strategy: without a clear view of your network, none of the other systems can ever be assured of working properly. Yet most organizations will tell you that they aren’t happy with their current solution and neither are their users. This is due to a set of interrelated factors: improper or inadequate models and standards, poor or nonexistent integrations, and lack of customization.
The first issue that causes unhappiness with a network inventory platform is when the representations and/or naming conventions aren’t accurate, or don’t work out in the field. Every database is to some extent an abstraction of the real world, and network inventory is no exception. But creating that abstraction properly requires in-depth knowledge of the inventory system, the network and how that network information is being used. In the field, the system is used to find and record information about the network at a level that’s close to the hardware: find an open port and determine which fiber has enough available bandwidth to support a new service. This requires that the models and naming conventions be accurate and directly relatable to the physical network.
At the management level, the network inventory system is used to create reports on the state of the network with an eye towards cost control and capacity management. At the executive level, the network inventory system provides information about overall network health and helps to determine the ROI for specific hardware vendors and bandwidth providers. For these purposes it’s not only important that models and names be accurate, but that they are also consistent. Otherwise, the reports used to make critical business decisions may be inaccurate because inconsistent naming caused the query to match too few (or too many) network elements.
So if you have good models and good naming conventions, how do things go wrong and frustrate the users? Well, there are two ways: there’s already bad data in the system, and more bad data keeps being added to the system. For the bad data already there, a cleanup effort using bulk data manipulation tools like Sincera’s SASI Data Manager can fix the problem, but unless you have methods to keep the bad data from continuing to be added, it’s a short-term solution.
Keeping the bad data out is what the other two issues are all about. Too often the answer is to provide a set of standards to the staff using the network inventory system and simply tell them to follow it. This is inadequate for a couple of reasons:
- The standards constantly evolve as new equipment is added to the network, and new services are developed in the company. It’s a moving target that most users simply don’t have the cycles to keep up on.
- People are fallible. No matter how diligently they try to ensure that their data is solid, mistakes will happen.
One way to minimize the possibility of erroneous data in your network inventory system is to take the user out of the loop wherever practical. If part of their role involves gathering information from another system and entering into the network inventory system, a properly designed and implemented integration between the two systems can save time, money, and effort. It is a similar case when another system requires network inventory data. For example, if you have an automated provisioning system, an integration to the network inventory platform can ensure that the proper steps are accomplished in provisioning a circuit without requiring an engineer to type data into an interface.
Where a full integration isn’t the answer, system customization comes into play. Oftentimes we see cases where the company’s staff has developed complex processes to ensure that they are conforming to the requirements of the network inventory system’s interface. Instead, it’s a better solution to create interfaces that work the way the company does and have the users work in that. A properly designed interface can:
- Enforce naming conventions and modeling standards
- Ensure that essential steps aren’t skipped or mishandled
- Cut down on training time, as the interface works for the user and not vice versa
- Interact with multiple systems behind the scenes, pulling data from other systems to be added to the network inventory as required
- Allow changes to standards to be implemented at the back end, without any change being apparent to the end users
- Save time! A well-designed interface can cut down the time it takes to accomplish tasks like a complicated circuit design from hours to minutes, and that generates enormous ROI for the company.
A network inventory should be complete, correct, coherent, and consistent. And it should achieve all four of these traits without extraordinary effort on the part of the company. In fact, it should be the natural and inevitable outcome of proper standards, integrations that work, and most of all well-designed and well-implemented network inventory customizations that seamlessly guide the users along the way while allowing them to concentrate on their jobs and not on trying to make their inventory system work.
If you’re looking for some assistance in getting out of the “audit, update, degrade” loop with your Granite Inventory or Ericsson Adaptive Inventory system, contact Sincera Consulting today!