I think it’s fair to say service providers have been trying to reduce costs for years, but with the recession of 2009, providers were actively looking at every Opex dollar to understand how to shed more. This came in the form of staff reductions and re-negotiated agreements with suppliers to name a few. However, data hungry customers both on the fixed line and mobile side of the business challenged providers to keep costs in check. The introduction of technologies such as Ethernet and IP have given providers some cost relief to deliver services more cost effectively.
But, when we consider the projections on data volume growth, service providers will not be able to sit still. They will have to find new ways to save costs.
So how does a service provider become more cost efficient? If we were to take lessons from other industries we would see that “Automation” is key to efficiency.
Service providers will tell you they automate and so will the software vendors that provide the solutions. The reality is that the software will store, collect and report. Very little software will make a decision and act on it. Take several examples from the OSS perspective which I believe is perhaps the biggest problem area;
- Inventory Management – Managing network inventory is not new. But let’s face it, this is simply a tool that understands the relationships between how network devices relate to each other. Sure it will enforce rules on the use of these devices and in some cases may guide a user through a process such as a service design, but it is the end user who has to interact with it and make decisions. A very large number of inventory implementations have no way other than manual activity to ensure the data is accurate.
- Performance Management – Collecting real-time data from the network to understand data throughput is very important to a service provider. It helps them make decisions on what actions to take to balance capacity vs cost. But today, most performance solutions generate some form of report showing trends etc. It is up to the end user to figure what actions need to be taken. What is missing from performance management systems is a real understanding of network implementation. This makes it impossible to pinpoint the reason for a bottleneck.
- Alarm Management – Collecting alarms from the network helps a provider identify network troubles and depending on the level of customization that has been built, may trigger the generation of a Trouble Ticket. Tickets are then dispatched to users to resolve. Users are then left with the decision to prioritize which ticket to resolve first. If Alarm Management systems were built to focus on Customers and their Services rather than devices and trigger tickets based on number of customers impacted, or perhaps customers with stringent SLAs then troubles would be resolved in an other than best protects the companies revenue.
What has allowed other industries to manage costs is to have a clear concise understanding of inventory. This is true if you are an auto-maker, an online reseller, or a big distribution store. Most of the other supporting systems such as ERP are build around the inventory system. If an item has reached a low stock level, automation kicks in and re-orders more. When something is sold or used the inventory stock level is automatically updated.
Now compare that to a service provider. Most providers will have some form of inventory system which could have been in-house developed or purchased. These systems will have a level of data accuracy which is often poor to good. Providers struggle to integrate their systems and processes around the inventory system because of the data accuracy issue and the fact that most inventory system implementations were driven from a service fulfillment perspective vs considering what the operation requirements were. So performance and alarm data can not be integrated to the inventory data. Result, very little real automation.
Consider for a moment the amount of automation that service providers have put into place around voice services. Now imagine what the costs to both the service provider and the customer would be to handle a simple call if the automation didn’t exist. Now apply that to data, considering the growth demand we can quickly see that automation is a immediate requirement.
For more of my thoughts visit my Blog.