top of page

Abstract

Welcome to Connections 2026 and Introduction to IETF 126

Dhruv Dhody

Ever wondered how the Internet’s protocols and standards come to life? This talk dives into the IETF, its ethos, the process behind RFCs and Internet Drafts, and how the IETF community drives innovation in networking. We’ll also explore the role of organizations like IIESoc and INTC, highlighting their contributions and activities. Whether you’re new to the IETF world or just curious about how things work behind the scenes, this session will give you a solid introduction to the world of internet standards.

Post-Quantum Cryptography: Preparing for the Quantum Era

Nalini Elkins

Quantum technologies are rapidly transitioning from research laboratories into practical applications that will transform communications, computing, sensing, and cybersecurity. This introductory presentation provides a high-level overview of three key areas: quantum sensing, quantum computing, and post-quantum cryptography (PQC). It explains how quantum sensing may enable resilient navigation beyond GPS, how quantum computers can solve certain optimization problems more efficiently than classical computers, and why the eventual arrival of a Cryptographically Relevant Quantum Computer (CRQC) threatens today's public-key cryptography. The presentation concludes with an overview of the global effort to standardize and deploy post-quantum cryptography through organizations such as NIST and the IETF, highlighting why engineers, researchers, and technology leaders should begin preparing today for the quantum era.

IP-based Geolocation - The state of the craft, best practices, and the road ahead

Siddharth Mathur

Self-published IP-based geolocation feeds are gaining adoption from ISPs and cloud companies eager to wrestle back control over their IP address pools from "black box" geolocation data providers. What are RFC 8805 and 9632, their state of adoption, and the new JSON-based extensible format now under discussion. And how Indian network operators can publish their own.

The Evolution of BMP: Building the Next Generation of BGP Observability

Narasimha Prasad

As network scale and complexity grow, the need for real-time, vendor-neutral BGP control-plane observability has moved from a luxury to an operational necessity. The BGP Monitoring Protocol (BMP), originally designed to replace inefficient CLI scraping and proprietary telemetry formats, is currently undergoing a significant architectural evolution within the IETF

This presentation provides a deep dive into the foundational BMP RFCs and explores the cutting-edge drafts currently transforming the protocol into a comprehensive, high-scale observability framework.

Extensible Protocol Foundation
• BMPv4, Path Marking, Structured Timestamps

Event-Driven Observability & Synchronization
• Generic Event Notification (GEN), Route Event Logging, BMP Snapshots

Operational Scale
• New methods(3) to reduce CPU and message overhead on the router

Operational Manageability & Security
• Manageability
• Security & Emerging Frontiers – TCP-AO, RPKI, QUIC
• Operational Considerations

AI-Native Networking: Inference, Automation, and the Standards That Must Catch Up

Vinayak Hegde

The shift from training to inference and agentic AI is reshaping network requirements faster than standards bodies can respond. The talks covers emerging work across three layers: fabric standards for inference traffic, content-access protocols for training data , and agent communication protocols. We assess which problems these standards solve, which remain open - particularly agent identity and enforcement - and where IETF efforts lag behind faster-moving, Linux Foundation–governed alternatives.

AI Data Centers: Networking Problems and Solutions

Shraddha Hegde

AI data centers face unique networking challenges as model sizes and data volumes grow. This talk examines the key problems bandwidth bottlenecks, congestion, latency sensitivity, scalability limits,
and presents solutions being offered in this space for congestion mitigation including multi-planar architecture, deterministic forwarding, Multipath Traffic Engineering and Multipath Reliable Connection (MRC)and RNG.

​AI Inference Streaming: What's on the Wire

Santhosh Pallaghati

AI inference endpoints stream responses using a fragmented set of vendor-specific protocols and payload schemas — SSE with vendor-unique JSON (OpenAI, Anthropic, Google), Connect+proto with binary-encoded envelopes (Cursor, Windsurf), WebSocket with SignalR framing (Microsoft), and JSON-RPC over SSE (MCP). While the underlying transports (HTTP/2, WebSocket, SSE) are IETF/W3C standardized, none of the application-layer formats are. This presentation walks through production pcap analysis of these AI traffic patterns and quantifies the protocol divergence and metadata overhead observed across vendors.

AI in the IETF and IRTF: Current Work and What’s Next

Dhruv Dhody

This session will provide an overview of the AI-related work already underway in the IETF and IRTF, highlighting the existing working groups and research groups addressing AI-related topics. It will then examine the three AI-focused Birds of a Feather (BoF) sessions at IETF 126, outlining the problems they aim to solve, their scope, and how they fit into the broader standards landscape. The goal is to give participants a clear picture of the current state of AI standardization activities across the IETF and IRTF and the directions being explored for future work.

Internet Scale Notifications with HTTP Events Query

Rahul Gupta

Events Query is a minimal protocol built on top of HTTP that allows user agents to receive event notifications directly from any resource of interest using the QUERY method [RFC10008]. The Events Query Protocol (EQP) is predicated on the idea that the most intuitive source for event notifications is the resource itself.

High-Performance Networking with eBPF and XDP

Sushanth S. Rao

eBPF has emerged as one of the most transformative technologies in the Linux kernel, enabling safe, dynamic extensions to kernel behaviour without requiring kernel modifications.
In this talk, I will explore how eBPF and XDP can be applied to high-performance networking, with a focus on early packet processing, profound visibility, and rapid feature development. Demonstrate the practical introduction to eBPF fundamentals, including the execution model and safety guarantees. I will then walk through the Linux networking pipeline and explain the unique position of XDP in terms of performance and early packet interception. As part of the talk, I will demonstrate a firewall I built using eBPF, showcasing how packets can be parsed, filtered, and processed. Also, how eBPF makes it possible to quickly test new or changing protocols without having to recompile the kernel, which makes experimentation much more flexible. I will also touch upon traffic shaping using eBPF at the TC layer. You will gain a clear understanding of how eBPF and XDP can accelerate development and unlock new capabilities in the Linux networking stack.

 

Improvements to FQ-PIE

Vishal Kamath

FQ-PIE is an active Internet Draft in the tsvwg. Since the time it was first introduced in the working group, we have received a lot of feedback, suggestions, and improvements for the draft; having worked on FQ-PIE testing and analysis, there was now scope for improvements. This talk covers the continued work carried out for FQ-PIE, including support in a golang tc library, adding FQ-PIE support to a OpenWRT QoS tool, implementing class stats for FQ-PIE in the Linux Kernel, and most importantly - adding L4S support for FQ-PIE.

Scalable Simulation of QUIC Networks with picoquic and ns-3

Abhyuday K. Hegde

QUIC has emerged as a widely deployed transport protocol, yet support for simulating modern QUIC implementations remains limited. This talk presents a new ns-3 module that integrates the Picoquic QUIC implementation with ns-3, enabling simulation of QUIC within a network simulation environment. The talk will discuss the design and implementation of the module and demonstrate how it can be used to evaluate QUIC performance across a variety of network topologies.

Reducing Tail Latency by Optimising CoDel for High-BDP Networks

Mohit P. Tahiliani

Controlled Delay (CoDel) and Flow Queue CoDel (FQ-CoDel), as described in RFC 8289 and RFC 8290, are supported by several Linux distributions and enabled by default in many of them. The effectiveness of CoDel relies on its control law, which was designed under the assumption that the congestion control algorithms (CCAs) at the endpoints are based on Reno. However, with CUBIC being the most widely deployed CCA for over a decade, this assumption is no longer valid. In this work, we propose a new technique for CoDel to reduce the tail latency in high-BDP networks. Currently, we have implemented the proposed technique in both Linux and ns-3, with both implementations being available openly.

Coherent Pluggables: Transforming IP over DWDM Networks

Swamynathan Balasundaram

Network architectures using IP over DWDM has been deployed and operated for many decades now with independent technology evolutions. With the innovations in the coherent optical plugs the network operators now have a compelling opportunity to introduce cost and energy efficiencies in the network. This however brings in operational and cultural challenges. The presentation looks at the proposed challenged and options on how these can be overcome.

​VPN or Vpwn? How Afraid Should You be of VPN Traffic Identification?

Tanmay Rajore

Several governments are gradually choosing to monitor VPN traffic. In this paper, we explore how hard or easy it would be for large ISP-scale adversaries to identify and block VPN traffic. More specifically, we try to answer questions like whether ordinary netizens fear such decisions or whether it is not as trivial to identify and block all sorts of VPNs. A recent study found that blocking and identifying OpenVPN endpoints is feasible for small ISPs. We explored detecting Open- VPN and alternatives like TLS, SSH, IPSec/IKEv2, Wireguard, and proprietary VPNs. Analyzing seven popular commercial and open-source VPN services, we identified patterns for detection. While OpenVPN is easily spotted, many alternatives resist iden- tification, some using tactics like obscure TLS ClientHello SNI strings. We demonstrated evasion methods, including altering packet sizes, sending dummy traffic to confuse middleboxes, and obscuring plaintext strings. We also proposed a scalable mechanism for OpenVPN services to hide identifiable plaintext without affecting user or gateway scalability.

Building Internet Measurement and Engineering ecosystem from Classrooms to Hackathons

Dr. T. Venkatesh and 

Anand Raje

The evolution of Internet engineering requires the ability to observe how networks behave in practice, analyze operational evidence, and translate those insights into deployable solutions, operational improvements, and standards-related work. This session presents the Advanced Internet Operations Research in India (AIORI) initiative and its approach to building an ecosystem for Internet measurement, experimentation, education, and community participation. The session will also discuss integrating Internet measurement concepts and hands-on experimentation into an academic course at IIT Guwahati. The objective is to help students move beyond theoretical networking and develop practical capabilities in reproducible measurement, network troubleshooting, protocol analysis, and Internet operations research. A further focus will be the preparation for the 2026 AIORI Hackathons. These hackathons will bring together students, faculty members, researchers, and network operators to work on real Internet engineering challenges and develop practical solutions. The session will highlight a continuous model in which classroom learning, operational measurements, open testbeds, hackathon-based innovation, and engagement with the IETF reinforce one another. This model aims to develop a stronger and more practice-oriented Internet engineering community in India.

bottom of page