Life inside the network operating center (NOC) at a mid-sized communications service provider (CSP) is a pressure-packed environment, as engineers and operators are bombarded with upwards of 30,000 alarms per day, 24 hours a day, 7 days a week.

Each of these alarms — triggered by devices across the network — represents the possibility of significant network outages, which can impact service level agreements, customer satisfaction and the provider’s profitability and overall success.

Operators and engineers in the NOC must quickly triage the constant inflow of alarms, identifying the greatest threats to uptime and deploying the right resources to neutralize or mitigate the problem.

This is all easier said than done, as network administrators work to conduct root cause analysis in real time, identify a path to resolution, and solve the issue as quickly as possible — then repeat the process.

This assess-diagnose-resolve loop has proven itself to be an ideal use case for the infusion of AI across a variety of network applications, which is why forward-thinking CSPs are deploying IBM Cloud Pak for Watson AIOps across their networks to:

  • Increase uptime
  • Reduce mean time to response
  • Improve their customer experiences
  • Improve the productivity and effectiveness of operators and engineers
  • Reduce the level of expertise required to perform operations tasks effectively
  • Reduce overall costs
Transforming your operations with IBM Cloud Pak for Watson AIOps

Automating your operations with Watson AIOps transforms and streamlines the way your NOC engineers and operators handle inbound alerts by automatically filtering and correlating events by:

  • Timing, including repetition or seasonality by day, hour, or month.
  • Location, including the ability to aggregate multiple alerts triggered by the same device.
  • Relationship, identifying root cause based on the impact of one device on peer or neighbor devices.

Watson AIOps can also learn from these patterns over time, alerting network operators about a potential device failure before it happens. This insight allows the provider to conduct preventative maintenance and address the situation before the device fails, improving uptime and reducing expedited support and parts expenses.

Of course, transforming your network operations with Watson AIOps isn’t as easy as plugging a new rack of servers into your data center or activating a SaaS application, but that barrier to entry only increases the durable competitive advantage that committed CSPs will receive for their investment.

Accelerating your AIOps time to value and ROI

Persistent Systems has successfully integrated IBM Cloud Pak for Watson AIOps and its predecessors for decades across a variety of networks, industries, organizations, and regions.

Here are a few of the best practices we’ve identified to help accelerate your time to value:  

  • IBM Cloud Pak for Watson AIOps isn’t a solo effort: AIOps consists of four or five disparate products that address different functions of the network, including Fault Management, Configuration Management, Automated Event Enrichment and predictive Insight. While these are all part of a single suite, they don’t work magically together upon installation. Each of these must be integrated fully and tailored to the provider’s needs before Watson AI can be infused throughout.
  • Focus on documenting your alert correlation policies: The ability for Watson AIOps to correlate alerts and events to reduce complexity, improve decision making accelerate root cause analysis must be customized for each CSP’s needs. So focus attention on determining how you want to have your alerts correlated, which ones you want to ignore, and how you want them presented to your network operators.
  • Develop the full topology: If Watson AIOps doesn’t know a device exists or how it relates to other devices around it, there’s no way to monitor and analyze it. Watson AIOps needs a complete picture of your network topology, including interrelations between devices, peer-to-peer connections and the like. This can take some doing, but it’s a critical step to accelerating the root cause analysis capabilities and sorting which alerts are the root cause and which are just symptoms.
  • Don’t forget about post-installation support: Watson AIOps isn’t a “set-it-and-forget-it” solution; it requires ongoing management and support to learn and adapt to your dynamic network environment. Be sure to find a systems integration partner that can provide timely first-, second- and even third-level technical support for all of your users on a 24/7 basis. Because your network never sleeps, and neither should your network automation support.

Persistent Systems can help transform your telecommunications network operations with IBM CloudPak for Watson AIOps and improve your uptime and service levels while decreasing your costs, visit our Telecom & Media page to find out how.