Cable network reliability has many important dimensions, but operators are all too familiar with the significant cost of maintenance and repair, and some with the advantages of Proactive Network Maintenance (PNM). But not everyone has taken full advantage of PNM. Let’s have a look at some of the reasons for that, and what CableLabs is doing to address those needs as part of its PNM project.
The Proactive Operations Problem
CableLabs has been informally assessing the reasons why more operators don’t take advantage of the proactive gift that DOCSIS® provides: the ability to use PNM data to find problems in the network before they become impactful and costly.
It takes a lot of work to implement solid PNM solutions that keep working. A key task in operations is to make decisions based on data. That takes expertise and time. Not every operator or vendor has an expert army in place to analyze all the available operations data to find proactive maintenance work worth doing. Machine learning is anticipated to help, but it will require a lot of work to apply those techniques successfully to an operations task like PNM, and even more to develop the needed controls. Likewise, not every operator or vendor has a statistical analysis or IT army in place to build enterprise tools to automate the process of turning data into action.
Some operators need to start small with testing PNM concepts to find a solution that fits their needs. That means many operators must experiment and learn first. But that requires basic, general tools in hand before experimentation can begin.
A ProOps Platform for Everyone
Figure 1. ProOps with its elements and workers in four layers, built on CCF, on top of the network.
CableLabs created a generalized process for translating data into operations actions and applied it to PNM. Then we built the Proactive Operations (ProOps) platform to enable this process, thus making it easy for everyone to try, develop, deploy and make full use of PNM.
ProOps translates network data into action through a framework that is not strictly enforced but is enabled and supported to better ensure effective proactive maintenance.
The steps we identify for turning network data into action are briefly as follows, moving up from the network, through data collection, and through the worker layers of ProOps in Figure 1.
- Extract Data from the Common Collection Framework (CCF)—ProOps uses CCF to extract the data it needs from the network, then applies basic analysis to translate the data into useful information.
- Analyze for Triggering—Next, the results are analyzed further to determine whether they are interesting or not; interesting results are “triggered” for deeper scrutiny. The data are looked at over time and across data sources to orient the information into context.
- Make It Actionable—Once we find the most interesting network elements to watch, we group network elements into network tasks and provide a measure of importance for the identified work.
Threshold Analysis—The best work opportunities get picked to become proactive work packages, which can be selected based on impact to customers, likelihood of becoming an emergency, and so on.
You ShOODA Get ProOps!
The steps we outline for turning network data into action—or in this case pro-action—align nicely with the well-known strategy of observe, orient, decide, act (OODA). This OODA loop, or OODA process, was created by U.S. Air Force Colonel John Boyd for combat operations. The operations of combating network failure aren’t much different! If you work as a cable operator, then you know.
ProOps is available upon request to any operator member or vendor of the CableLabs community. CableLabs supports users by helping them to deploy ProOps with an example application that shows how to configure it to a specific operator or use case, and we will help our members develop solutions in it, too. Just contact Jason Rupe to get your copy.
Our goal is to help operators provide highly reliable service, and efficient, effective operations is one proven way to do that. ProOps is the latest tool to combat network failures.
