Orange integrates Augtera Network AI platform to its NOC tools to leverage AI/ML in daily network operations. This will reduce Network Operation Centers alarms by 70% and prevent failures.

FAQ

The FAQ ( frequently asked questions ), provides answers to commonly asked questions about the company and platform.

Q. What does Augtera Networks do?

A. Augtera Networks designs, develops and sells a software platform that detects network incidents that have already occurred or may occur in the future.

Responses to incidents can be automated, including, but not limited to, monitoring console alert, trouble ticket creation, automation system trigger, and the platform’s own automation playbooks.

Q. What stage of development is Augtera Networks at?

A. Augtera Network AI went into production in 2020 and has been large scale production deployments for over two years. Product deployments include large scale eCommerce Enterprises, Financials, Retailers, Telecom SPs, Cable Operators, and MSPs.

Q. Where is Augtera Networks located?

A. Augtera Networks is headquartered is in Palo Alto, California where it draws core engineering and networking talent. Augtera Networks also has locations in New York, France, Turkey, and Columbia.

Q. How is the software delivered?

A. Augtera Network AI is best consumed by most network operations teams as SaaS. Hybrid cloud and on-prem options are also available.

Q. How is Augtera Networks different than other network management offerings?

A. Augtera Networks is disrupting network management by not just receiving and visualizing alerts, but by eliminating the noise associated with alerts so that network operations teams can focus on what is most important to them, act proactively to address issues before they are noticed, and in some cases even prevent them from occurring at all. Multi-layer, topology-aware network model and autocorrelation groups related events and anomalies together, as well as identifying the root of an incident.

FAQ - Impact of Augtera Network AI. Before and After.

Q. Does Augtera have an AIOps offering?

A. Yes, Augtera Network AI is a Network AIOps platform. It provides monitoring, observability, and automated ingestion, detection, noise elimination, incident root identification, and incident response (automation, notification, ticketing…). Central to the offering is 9+ purpose-built anomaly detection algorithms for Networking use cases. The platform also implements multi-layer, topology-based, autocorrelation, natural language processing, and other AI/ML technologies.

Q. Does Augtera provide monitoring?

A. Yes, Augtera must ingest data and monitor as a foundation for more advanced capabilities. However, Augtera does much more than monitoring. Customers do not need to sit in front of monitoring terminals hoping to pick out critical issues in a sea of alerts. Instead, they can experience Network AI as automated trouble tickets or automation responses.

Q. Does Augtera provide observability?

A. Yes, Augtera has sophisticated deep dive and query capabilities for triage and to identify anomaly signatures that can be reduced to classifiers for future high-fidelity alerts. However, Augtera Network AI does more than observability. Customers do not need to sit in front of observability terminals hoping to pick out critical issues in a sea of alerts or perform manual correlation. Instead, they can experience Network AI as automated trouble tickets or automation responses.

Q. What is the difference between a threshold and an algorithm?

A. A threshold has a binary relationship with data, either current levels are below or above the threshold. This leads to false positives and/or false negatives in addition to significant ongoing tuning and maintenance. Some network properties, for example latency, are almost impossible to manage with thresholds because appropriate ranges vary from link to link and over time. Purpose-built algorithms understand patterns in networks and only create alarms when there is a significant departure from those patterns. In addition, algorithms can detect gray failures and trigger action before a complete failure.

Q, How is noise reduced?

A. Purpose-built algorithms for networking are inherently less noisy than thresholds. Correlation groups related events and anomalies together so when a switch fails, for example, instead of getting 30-100 separate alarms, a single incident is created with child records for related data. Augtera Network AI also allows customers to define what is operationally relevant for them: types of errors, parts of the network, etc.

Q. What outcomes are customers realizing?

A. Number of incidents / tickets are being reduced by 10 times, time between incidents is increasing by 4 times, detections times are reducing by 90%, mitigation by 50%+, and remediation by 40%+

Q. Can data be ingested from devices and logs?

A. Augtera Network AI can ingest data directly from devices, messages buses such as Kafka, APIs, and other collection solutions like Splunk. Currently supported data types include SNMP, sFlow, IPFiX, VPC Flow Logs, Syslog, and agents measuring latency and packet loss. For a full list of supported data types, see data sheet.

Q. Are agents required?

A. No agents / probes, hardware or software, are required for Network AI. Significant results can be provided with Syslog alone. SNMP can be used for topology auto-discovery, or topology can be provided through an API. SNMP and gRPC/gNMI each provide many metrics. Most existing telemetry data does not provide end to end latency and packet loss, so agents are required for those metrics.

Q. Does Augtera have a Data Center solution?

A. Yes, Network AI provides an extensive multi-layer, multi-vendor, topology-aware network model for Data Center operations including common constructs such as ToR, Spine, POD and more. For more information on the Data Center solution, see here.

Q. Does Augtera have a Syslog solution?

A. Network AI is a high-performance, high-efficiency real-time syslog analysis solution, capable of processing more than a billion messages per day. Classifiers identify known anomaly signatures, algorithms identify unusual bursts, and natural language processing identifies rare/unknown messages that sometimes precede equipment / network failures. For more information on the syslog solution, see here.

Q. Does Augtera have a WAN solution?

A. Yes, we are currently engaged with multiple Service Providers. Please contact Augtera Networks Sales for more information.

Q. Does Augtera have a SD-WAN solution?

A. Yes, we are currently engaged with multiple Service Providers and Managed Service Providers. Please contact Augtera Networks Sales for more information.