Anomaly Detection

IBM Predictive Insights leverages IBM’s Watson machine learning algorithms to analyze millions of metrics across thousands of resources to identify anomalous behavior in real time. We typically see anomalous behavior begin on service supporting resources before service degradation is realized. The solution can help you avoid your next outage.

  • NOC operators will be alerted of anomalous behavior and notify a Teir 2 SME
  • Tier 2 SME can drill down into the anomalous resources to troubleshoot the cause of the problem
  • Monitoring SME can leverage Predictive Insights baseline capabilities to discontinue manual threshold setting, thus saving manual work efforts.
Contact Us
  • James Anderson III
    "Since deploying the Assure1 solution, we have remediated the dependency and cost of operating with legacy tools and plan to reduce OPEX in the months and years to come. In addition to the direct savings, Assure1’s unified platform for fault, performance and topology is easy to use and provides our engineering team with the speed and agility to deliver a high-degree of service quality to our NOC and end-customers."
    — James Anderson III, HTC

Your early warning system for IT Operations

Today’s IT and Network Operations teams routinely find themselves in an unenviable position. Despite growing complexity caused by “software-defined-everything” and exponential growth in operational data to sift through, end users are demanding zero downtime and faster SLAs. Whether you’re an energy utility or a Tier 2 service provider, service impacting events are no longer tolerated. Customers have a choice and regulators are poised now more than ever to complicate your business.

IT and network performance management tools have been around for a long time. Besides the complexity of administering them, operations teams often don’t have the bodies or intimate knowledge of these point tools to watch dozens of KPI dashboards much less take action.

Most performance management tools have the ability to send northbound SNMP traps when KPI thresholds are met. However, as any operations person would tell you, setting and managing thresholds is an interesting balancing act that often results in important alarms missed or causing so much noise that that alarms ignored.

Suffice it to say, the value promised by most performance management tools – giving you proactive response to problems rather than reactive – often goes undelivered.

Over the past few years, leading big data and analytics vendors have made tremendous strides in delivering commercially available solutions for real-time structured data analytics. Perhaps one of the biggest leaders in this space is IBM and they have not overlooked IT and network operations.

With IBM Operations Analytics – Predictive Insights (PI), IBM’s structured data analytics delivers on three key areas:

  1. Once implemented correctly, it automatically generates and maintains its own thresholds based on self-learning so you can forget threshold management.
  2. Often, KPIs begin acting anomalously long before an impact to end user experience. Kind of like the “Butterfly Effect” in the Chaos Theory. Small variations often lead to big effects down the road. PI looks for this anomalous behavior and alerts operations teams using northbound SNMP traps. PI gives you the “early warning” you’ve always been looking for.
  3. Unique to PI, PI can derive mathematical correlations between KPIs and resources regardless of known topological relationships. This allows SMEs the ability to gain valuable insight into resources that impact each other without the use of topology discovery tools and/or a CMDB.

In summary, PI is helping organizations shift their response left of end-user-impact.

Besides being one of the only professional services organizations in the world who has successfully designed and implemented PI for customers, AccuOSS has also developed a methodology that allows clients to drive more actionable and consumable insights from the solution.

So if you’re interested in IBM Operations Analytics – Predictive Insights, please feel free to contact us by clicking the “contact us” button below.