Skylight Performance Analytics

What is Skylight Performance Analytics?

A unique view of service performance - in physical, virtual, and cloud infrastructures - keeps network and IT operations teams synchronized and proactive, providing an exceptional customer experience.

More than 30% of client churn is due to network issues that have not been identified or could not be resolved. And it's 50 times more expensive for service providers to win over a new customer than to keep a customer happy.

With complex networks, it's getting harder to get insights quickly to solve problems. How about stopping application and network issues in your tracks before your customers notice?

Skylight generates billions of unique performance measurements per day, providing excellent visibility into the performance of your network, services, and applications.

Skylight analytics, developed with big data and machine learning, filters these billions of data points into valuable information where you can take action.

How it works

Skylight analytics is a cloud-native platform designed from the outset to analyze large volumes of time series data in real time, providing levels of flexibility and speed that are not possible with legacy performance management tools.

• Skylight marks the stored data with contextual metadata through the user interface (UI) or RESTful APIs. For example, performance metrics can be enriched with inventory, location, or customer data to accelerate root cause detection and analysis
• Leverages machine learning (ML) and artificial intelligence (AI) algorithms to handle billions of data points and baseline performance data dynamically, detect anomalies, detect patterns, correlate performance data and metadata, and provide insight into service performance.
• Uses enhanced data in queries, dashboards, and reports in Skylight to provide additional network and service visualizations, and leverage correlation and prediction algorithms.
• Analyzes in detail user application issues in network infrastructure with root cause analysis, configurable alerts, and machine learning troubleshooting tools.

Capacity and data flow

Uses the power of Apache Druid and Apache Spark to manipulate petabytes of data.

Any type of data entry

Skylight performance management KPIs, SNMP and CSV/XML/JSON files, streaming telemetry data using gRPC and other similar data formats and protocols.

Correlation and contextual metadata

Performance data can be correlated with data from other systems, for example, inventory, topology, vendor, etc.

Machine learning

Baseline performance data dynamically detects anomalies, detects patterns, correlates performance data and metadata, and provides an insight into service performance.

Reporting and Visualization

Drag-and-drop report dashboards, custom formats, and ad-hoc data exploration and analysis.

Multivendor & Open

RESTful and JSON open and documented to and from third-party systems, e.g. OSS, orchestrators, etc.

Cloud-native microservices architecture

Runs on Linux / x86 / Docker

Multi-catalog platform

Logical instances in the installation of shared physical software

Flexible deployment

Public or private cloud SaaS

Resources

Deep, layer-to-end visibility into end-to-end service performance

Skylight discovers the underlying problem to see the "invisible" cause of service performance issues using advanced machine learning and correlating performance management KPIs with other metadata sources (for example, network topology).

Real-time contextual view to enable closed loop automation

Skylight automatically cleans and removes "suspicious data" to ensure highly accurate reporting and valuable information that enables network management automation and service assurance.

Consolidate multiple reporting tools into a single unified analytics platform

Skylight facilitates digestion of third-party data sources and acts as a complete feature for network and service performance intelligence.

Use cases

Troubleshooting performance issues

Skylight analytics processes millions of data points per hour and details it in actionable insights into not only where, but also how to solve performance issues, in three simple steps:

• Create dynamic dashboards that give you a unique view of network performance, using the power of machine learning to alert you to active and potential network issues.
• Drill down to determine the impact of an event alert (such as packet loss and latency) on your network and customer experience, and quickly identify the root cause.
• Harness the power of performance data, using machine learning, to determine what other events are impacting the network and where the root cause of the problem is located. Easily visualize and correlate metadata to understand the root cause of the problem and how to resolve the problem.

Closed-loop automation

With the shift to software-defined networks and increasingly native cloud network functions, there is a pressing need for new analytics-based monitoring software that can provide a multi-layered view of service availability and performance – in real time and in context. Skylight analytics transmits real-time data to monitor service performance and whether it's time to make changes or adjust policies.

• Easily integrates with any software-defined network (SDN) or orchestration workflow that can react in real time to events as they occur on the network.
• Enables closed loop automation capabilities with granular networking and service visibility. Events can be processed in real time to increase granularity and frequency of KPIs collection, or start collecting other KPIs to improve the ability to correlate and perform root cause analyses.
• Uses machine learning to identify patterns and anomalies and recommend bug fixes that, with human training, can become automated fixes and how to resolve the problem.

System Requirements

Runs on Linux / x86 / Docker