Professors Rejaie and Durairajan Receive $1M+ from NSF to Push the Frontiers of Runtime Programmable Dataplanes

Reza Rejaie and Ram Durairajan composite headshot

Professors Reza Rejaie and Ram Durairajan, co-directors of Oregon Networking Research Group (ONRG) and faculty members in the CS department, recently received a competitive $1M+ grant from the NSF Computer and Network Systems program in support of their research project titled: “Distributed Runtime Dataplane Telemetry as an Adaptive Query Scheduling Problem: Algorithms and Applications.” This research will have a transformative impact on improving the performance and security of tomorrow’s campus and enterprise networks.

Modern network management and security solutions increasingly rely on programmable dataplane systems. These systems offer the ability to launch telemetry queries, gain visibility into the state of the networks, and support a wide variety of applications (e.g. supporting the stringent requirements for streaming applications, detecting the onset of cyber attacks, etc.). However, such systems are designed for static queries and traffic workloads, do not scale with the number of queries or traffic rate, and assume that tasks are simple (e.g., stateless, single switch).

To tackle these challenges, this project envisions the next generation of network telemetry systems to accommodate traffic and query dynamics and adaptive telemetry applications over multiple switches across a network. A key enabler of the project is the emergent runtime programmable switches that allow reprogramming dataplanes “on the fly.” The novel scientific contributions of this project include (i) scheduling mechanisms to manage resources on switches while controlling accuracy-latency-reporting load; (ii) methods to support queries that require traffic visibility across multiple switches; and (iii) a novel interface for facilitating stateful and iterative execution of adaptive telemetry applications, designing a test suite to enable purposeful evaluations of telemetry systems, and developing post-mortem analysis to properly diagnose and tune these applications in a specific operational setting. The project builds on top of an earlier effort from ONRG.

Other participants in this four-year project are, Chris Misa, their senior Ph.D. student, their collaborator Walter Willinger from NIKSUN Inc., and Shahram Davari from Broadcom Inc.