Predicting future IT outages using AI

FixStream, an artificial intelligence company for IT, has introduced an advanced version of its product, an AIOps platform to predict business application issues across an enterprise’s entire hybrid IT stack. With new Machine Learning (ML) algorithms, advanced multi-layer correlation from business transactions to application services and infrastructure, FixStream can rapidly identify issues that can significantly impact business.

With the rapid pace of digital transformation and the need for on-demand access 24/7, companies rely on their applications to do business more than ever before. When an application goes down even for just a few minutes, IT can take days or weeks to resolve issues, costing millions of dollars in revenue. It is very challenging for IT teams to pinpoint and resolve these issues due to the increasing growth of real-time business transactions, and the complex nature of dynamic hybrid and software-defined technologies.

"We are modernizing IT, giving it a seat at the executive table,” said Sameer Padhye, Founder and CEO of FixStream. “FixStream’s multi-layer correlation, visualization and machine learning capabilities across business KPI’s, applications and infrastructure is a game-changer, reducing the complexity of how enterprises identify issues across their IT infrastructure and saving millions in revenue by predicting outages in the future.”

FixStream applies a machine learning algorithm to contextually correlated data to automatically discover patterns, so IT operations can predict and visualize future outages. As a result of deep learning insights, IT teams can significantly reduce troubleshooting efforts from weeks, months, or days down to minutes.

New FixStream ML capabilities include:

• Dynamic Thresholding and Multivariate Anomaly Detection: IT staff can now identify a sequence or group of anomaly events for transactions, applications and infrastructure entities that may impact a business application (i.e.-unplanned events such as DDOS attack or better plan for Black Friday)
  
  
• Sequential Pattern Analysis for Incident Prediction: IT staff can now identify and visualize in seconds the sequence of correlated events that have impacted a specific application or business process and identify patterns with probabilities to predict future incidents (i.e.-  predict eCommerce transactions such as ordering new service or paying bills will stop working after one hour due to a detected alert to prevent the incident from happening).
  
  
• Disk/Network Bandwidth Predictive Analytics: IT teams can now analyze historical trends around the utilization of an infrastructure resource and predict when the infrastructure entity will run out of capacity. Now IT operations can proactively plan to add more capacity before an event negatively impacts the business.