[Case Study] Automating bandwidth optimization and SRTE policy management

When a prominent networking innovator needed to turn theoretical research into a rugged, real-world application, they faced tight timelines and architecture hurdles. 

PANTHEON.tech stepped in to engineer a durable, portable application capable of processing Cisco XR router telemetry data in real time. In just one quarter, we transformed raw data pipelines into an automated real-time bandwidth optimization solution ready for customer deployment.

Project Snapshot

[The Challenge] Turning research code into production-grade telemetry networks

Telecommunications networks generate massive volumes of telemetry data. Extracting actionable insights in real time remains a significant bottleneck. 

Our client had advanced research and algorithms, designed to optimize network bandwidth. But this theoretical framework lacked the engineering durability required for live deployment. 

Manually handling data population into a timeseries DB, mining that data for network topology mapping, and provisioning segment routing traffic engineering policies created severe delays. Without an integrated application, the business suffered from slow development cycles and an inability to dynamically mitigate network congestion - risking SLA breaches.

Our Approach: Engineering a portable, automated bandwidth optimization app

PANTHEON.tech applied its expertise in cloud-native networking and Agile development to architect a robust, production-ready solution from the ground up. 

We focused on durability, portability, and rapid iteration. Our team systematically tackled the data pipeline to ensure a quick implementation.

1. Building a portable app foundation

Our first priority was ensuring that the application architecture aligned with enterprise-grade standards. We designed the core software framework to be fully compatible and easily portable to the timeseries database. This involved containerizing services and implementing design principles to guarantee high availability and durability.

2. Streamlining Telemetry Ingestion and Database Mining

Data without structure is a liability. We engineered robust data pipelines to populate real-time telemetry into a high-performance timeseries DB. This included:

• Automating the extraction and ingestion of streaming telemetry metrics

• Developing specialized mining algorithms to extract relationship data from the timeseries DB to populate the topology DB

• Creating a unified data layer that allows for rapid querying and real-time analysis of changing network states

3. Implementing the RTBO Algorithm and SRTE Management

With the data foundation secured, we shifted focus to the execution of the client’s core business logic during successive sprints. Over the three-month timeline, we progressively rolled out advanced automated traffic engineering capabilities.

• May (Data Collection): Integrated real-time congestion detection mechanisms alongside additional data gathering optimized for Real Time Bandwidth Optimization (RTBO).

• June (Policy Management): Programmed systems for the precise determination and placement of RTBO SRTE policies, ensuring automated maintenance and automated cleanup of stale policies.

• July (Algorithm Implementation): Fully coded the core RTBO algorithm into the application, transitioning it from a research design into an actionable system ready for real-world lab assessments.

4. Developing API Access and a Functional GUI

To make the system accessible to both external services and network operators, we designed a robust interface layer. We created clean, high-throughput API access models to expose both the timeseries data and topology data. Additionally, our team built a basic GUI (Graphical User Interface) to provide administrators with immediate visibility into current optimization metrics and network health.

[The Results] Production-Ready Automation Delivered in One Quarter

By combining our deep networking domain knowledge with Agile execution, PANTHEON.tech successfully turned an unoptimized research project into a fully functional network optimization app in just three months.

[Business] Accelerated Time-to-Market: Successfully transitioned a complex research project into an executable application within a single quarter, saving months of internal development overhead.

[Business] Lab Deployment Readiness: Delivered a robust software package prepared for immediate deployment in an end-customer lab environment for real-world validation.

[Technology] Automated Congestion Mitigation: Enabled real-time congestion detection and automated SRTE policy management (placement, maintenance, and cleanup), replacing manual intervention.

 [Technology] Portability & API Access: Achieved seamless portability to existing product platforms using timeseries database while delivering clean API endpoints for timeseries and topology database mining.

Ready to automate your network telemetry and traffic engineering pipelines?

Contact our Agile engineering experts today to discover how we can transform your complex research concepts into durable, production-ready enterprise solutions.