The 2023 Internet2 Technology Exchange was the event where global leaders convened to ensure the R&E community moved forward, together.

TechEX23 boasted an exceptional lineup of expert-led sessions, tutorials, working meetings, and co-located events, all designed to tackle the most pressing challenges and discover solutions in the realm of research & education.

The event featured five tracks for knowledge exchange & growth: Advanced networking, Cloud technology, Federated trust, identity, and access management, Information security, and Research engagement and innovation.

The Power of Community-Driven Content: The topics featured in the sessions were thoughtfully submitted and presented by the community, for the community. This ensured that the program was tailored to address the real-world issues faced by professionals, offering insights and strategies that made a genuine impact.

Jeronimo Bezerra and Renata Frez presented several updates on AmLight infrastructure.

Title: A Scalable Solution to Detect Microbursts
Short abstract: Detecting microbursts presents a persistent challenge for research and education networks (REN) and commercial internet service providers due to their sporadic and short-lived nature, often slipping past conventional network monitoring tools. These sudden bursts of traffic, lasting mere milliseconds, significantly affect data transfers and can cause costly performance issues in both long-haul and regional networks. The rise of programmable network devices with Intel Tofino ASIC and P4 has sparked an increasing interest in addressing this issue. While many solutions leverage programmable devices, implementing additional stages to the forwarding pipeline may not always be feasible, particularly for RENs supporting a range of protocols and functions. Consequently, out-of-band solutions employing technologies like In-band Network Telemetry (INT) and traffic mirroring have gained traction, enabling microburst detection through bandwidth utilization measurements at sub-second intervals. To enhance detection, an adaptive and efficient solution has been developed, aiming to evaluate and store granular counters selectively based on operator-defined metrics, effectively identifying microbursts as short as 20 milliseconds while managing resource usage. Read the full abstract here.

Title: DDoS Detection/Mitigation @ AmLight
Short abstract: Recent years have seen a rise in cyber attacks targeting Government and Research & Education facilities, as highlighted by Cloudflare. As an International Research and Education Network, AmLight provides vital academic and commodity connectivity across the USA, Latin America, and Africa. The network supports diverse science drivers with varying levels of infrastructure robustness, ranging from multiple 100 Gbps connections to 10 Gbps or less. However, even a 1Gbps attack can disrupt operations for some connectors within AmLight. To counter these threats effectively, a multi-tier DDoS Detection/Mitigation Model has been implemented. The model includes three levels: Tier 1 focuses on large-scale attack mitigation, leveraging a scrubbing service provided by AmLight’s primary upstream provider. Tier 2, on the other hand, detects and mitigates small-scale attacks through a cloud-based DDoS detection solution, allowing customization and federated management by individual connectors for enhanced resilience. Read the full abstract here.

More details about the event can be found here: