Home AI News Enhance Network Observability with Selector AI

Introduction to Enhanced Observability into Your Network

In today's discussion, we will explore how you can use selected AI to achieve enhanced observability into your network. As part of the solutions team at Selector, I have firsthand experience with the challenges faced by network operators and the solutions that can address them. In this article, I will introduce you to Selector's product and its customer deployments. Before diving into the specifics, let's highlight some key challenges that network operators face and the need for enhanced observability.

Key Challenges in Network Operations

One of the primary challenges faced by network operators is the need to analyze data from multiple sources to understand network outages or issues. Metrics, configuration events, alerts, and logs are valuable sources of contextual information, but they are often siloed across different monitoring tools. Network operators find themselves browsing through a multitude of dashboards, performing manual analysis to identify the root cause of the problem. This process is time-consuming and hinders efficient communication and collaboration among teams. Additionally, once the issue is identified, the time spent on analyzing data often far exceeds the time taken to implement the fix.

The Solution: Selected AI for Enhanced Observability

To address these challenges, Selector offers a comprehensive solution that enables network operators to have enhanced observability into their networks. The product focuses on three core pillars: data collection, correlation, and insights sharing. By collecting heterogeneous data from various sources such as metrics, logs, events, alerts, and routing information, Selector's platform enables automatic correlation of this data to provide actionable insights.

The Three Pillars of the Product

Selector's product is built on three main pillars: data collection, correlation, and insights sharing. By seamlessly collecting data from diverse sources and correlating it in real-time, network operators gain valuable insights into the health and performance of their networks. The platform integrates with collaborative tools like Slack and Microsoft Teams, facilitating efficient communication among teams and enabling collaborative debugging.

Use Cases and Deployments

Selector's product has been successfully deployed across various industries, including service providers, retail, streaming TV, and finance. In a large retailer, for example, the product provides insights into the health of individual stores, helping identify issues in the backbone network or specific VRFs. For streaming TV providers, Selector's platform offers performance monitoring and helps identify Patterns of issues at the subscriber or geographical level. In the finance sector, the product helps analyze transaction failures and provides insights into application or underlying network issues.

Deployment Models: Public Cloud, On-Premises, and Hybrid

Selector provides flexible deployment options to cater to the diverse needs of its customers. The product's containerized architecture allows for easy deployment in public clouds, making it available as a SaaS offering. Additionally, Selector can be deployed on-premises or within a customer's VPC if they prefer to keep their data locally. This flexibility ensures that customers have control over their data and can comply with data sovereignty requirements.

Pricing Model Based on Devices and Entities

Selector's pricing model is based on the number of monitored devices and entities, providing predictability and transparency for customers. Unlike other tools that charge based on data ingestion, Selector's pricing focuses on the tangible metrics of the network, ensuring that customers pay for the value they receive rather than the volume of data being analyzed.

Data Collection and Correlation

Selector's platform is capable of collecting data from a wide range of sources, including metrics, logs, events, alerts, and routing information. The platform supports multiple protocols, enabling seamless integration with different networking devices and systems. Whether it's streaming telemetry, SNMP, CLI scraping, or REST API connectors, Selector can adapt to the data collection requirements of its customers. The platform's ability to selectively collect and correlate Relevant data ensures efficient use of resources and minimizes noise.

Storing Insights for Analysis

Selector's platform focuses on storing and providing insights rather than raw logs. This approach allows for faster processing, as insights are compressed and provide the right signals for analysis. By storing insights for a reasonable period of time, such as six months, network operators can dive into past incidents, correlate data, and gain valuable historical context. The platform acts as a "DVR" for network operations, allowing users to review and understand previous events.

Integration with Telemetry and Monitoring Tools

Selector seamlessly integrates with existing telemetry and monitoring tools, leveraging the data already collected by these tools. Whether it's using gNMI or gRPC protocols or connecting to a Kafka message bus, Selector can adapt to the customer's existing infrastructure. The platform complements and enhances existing monitoring systems, providing a unified view of network health and performance.

ML and Training for Baseline Determination

Selector leverages machine learning (ML) algorithms to determine normal baselines and detect anomalies in network data. By training its models on historical network data, Selector can identify patterns and behaviors indicative of network issues. This capability enables proactive identification of potential problems, reducing mean time to innocence and facilitating faster incident resolution.

Selective Data Collection and Kafka Integration

Selector offers selective data collection, allowing network operators to choose the specific data they want to collect from devices. This approach eliminates the need to send all data from devices, reducing bandwidth and storage requirements. In some cases, customers already have a Kafka message bus deployed, making it the preferred distribution mechanism. Selector can seamlessly integrate with Kafka, extracting the necessary data and providing valuable insights for analysis.

Conclusion

Enhanced observability into your network is essential for efficient network operations and incident management. Selector's AI-powered platform enables network operators to Collect, correlate, and analyze data from various sources, providing actionable insights and facilitating collaboration among teams. With flexible deployment options, transparent pricing, and a focus on storing insights, Selector offers a comprehensive solution for enhanced network observability. Whether it's optimizing performance, identifying anomalies, or improving incident response, Selector empowers network operators to proactively manage and enhance their networks.

Highlights

Selector's AI-powered platform enables enhanced observability into networks.
The platform collects and correlates data from multiple sources, providing actionable insights.
Deployment options include public cloud, on-premises, and hybrid models.
Pricing is based on the number of monitored devices and entities.
Selector integrates with existing telemetry and monitoring tools.
ML algorithms enable baseline determination and anomaly detection.
Selective data collection and Kafka integration optimize resource usage.

FAQ

Q: Can Selector integrate with existing monitoring tools? A: Yes, Selector seamlessly integrates with existing telemetry and monitoring tools, leveraging the data collected by these tools.

Q: How does Selector ensure data privacy and compliance? A: Selector provides deployment options that allow customers to keep their data locally, ensuring compliance with data sovereignty requirements.

Q: Can Selector handle data from different networking vendors? A: Yes, Selector is vendor-agnostic and can collect and analyze data from a wide range of networking vendors, including Cisco, Juniper, and Arista.

Q: How does Selector determine normal baselines for anomaly detection? A: Selector applies machine learning algorithms trained on historical network data to identify normal baselines and detect anomalies.

Q: Does Selector support deployment in Europe due to GDPR concerns? A: While Selector is not currently deployed in Europe, the company has plans to expand its deployment capabilities to comply with GDPR requirements in the future.