๐ฃ๐ฎ๐ฝ๐ฒ๐ฟ ๐๐ฐ๐ฐ๐ฒ๐ฝ๐๐ฎ๐ป๐ฐ๐ฒ ๐๐ป๐ป๐ผ๐๐ป๐ฐ๐ฒ๐บ๐ฒ๐ป๐ โYaksha-Prashna: Bridging the semantic Gap in eBPF Bytecode Analysis and Diagnostics
We are delighted to announce that the paper titled โ๐ฌ๐ฎ๐ธ๐๐ต๐ฎ-๐ฃ๐ฟ๐ฎ๐๐ต๐ป๐ฎ: ๐๐ฟ๐ถ๐ฑ๐ด๐ถ๐ป๐ด ๐๐ต๐ฒ ๐ฆ๐ฒ๐บ๐ฎ๐ป๐๐ถ๐ฐ ๐๐ฎ๐ฝ ๐ถ๐ป ๐ฒ๐๐ฃ๐ ๐๐๐๐ฒ๐ฐ๐ผ๐ฑ๐ฒ ๐๐ป๐ฎ๐น๐๐๐ถ๐ ๐ฎ๐ป๐ฑ ๐๐ถ๐ฎ๐ด๐ป๐ผ๐๐๐ถ๐ฐ๐โ has been accepted at the 22nd International Conference on emerging Networking EXperiments and Technologies (๐๐ผ๐ก๐๐ซ๐งโ๐ฎ๐ฒ).
๐ฅ ๐๐๐๐ต๐ผ๐ฟ๐: Animesh singh (IIT Hyderabad), K Shiv Kumar (IIT Hyderabad), VenkataKeerthy S (IIT Hyderabad), Pragna Mamidipaka (Carnegie Mellon University), R V B R N AASEESH (IIT Hyderabad), Sayandeep Sen (IBM Research India), Palanivel Kodeswaran (Walmart), Theophilus A. Benson (Carnegie Mellon University), Ramakrishna Upadrasta (IIT Hyderabad), and Praveen Tammana (IIT Hyderabad).
๐ ๐๐ผ๐ป๐ด๐ฟ๐ฎ๐๐๐น๐ฎ๐๐ถ๐ผ๐ป๐ ๐๐ผ ๐ฎ๐น๐น ๐๐ต๐ฒ ๐ฎ๐๐๐ต๐ผ๐ฟ๐!
๐ ๐๐ฒ๐ ๐๐ถ๐ด๐ต๐น๐ถ๐ด๐ต๐: The paper presents Yaksha-Prashna, a scalable system for validating eBPF network function bytecode and providing actionable diagnostic feedback. It bridges the semantic gap between low-level eBPF bytecode and high-level network behavior by using a novel control-flow-guided dataflow analysis framework to reconstruct packet-processing semantics directly from bytecode, representing them as a Network Context Model (NCM). The system also introduces a unified query interface that supports both validation and diagnostics, enabling users to understand not only whether a violation occurs, but also why it occurs. Overall, Yaksha-Prashna provides a practical and scalable approach to eBPF bytecode analysis and debugging, achieving 200โ1000ร faster validation than existing state-of-the-art works.