My research interests focus on different areas of computer networking, specifically, routing scalability in the context of IP and BGP protocols, verification of forwarding correctness of a router and in a private network, the programmability of routers' control and data planes for designing more dynamic, application-driven networks, and Remote Direct Access Memory data transfer for minimizing the CPU overhead when processing network packets at end-hosts. Throughout the years of graduate studies at Clarkson University and Rochester Institute of Technology, I have co-authored multiple full conference papers at venues such as IEEE INFOCOM, ICC, and NCA, ACM ANCS, HPDC and CoNext, and poster papers at IEEE ICNP and ACM ANCS.

In my research, I study different approaches to mitigating the negative effect of the global routing table (or FIB, a Forwarding Information Base) growth, such as TCAM memory overflow and increased operational expenses for routers storing that table. Together with my first advisor, Dr Yaoqing Liu, we designed an algorithm called "FIB Aggregation with Quick Selections" (FAQS) with reduced BGP update processing time and churn in the data plane. The corresponding paper for FAQS has been published at IEEE NCA in 2018 [1] and was presented at the conference by myself. Besides, a poster paper for FAQS has been published at ACM ACNS in 2018 [2].

We have also studied the routing scalability problem from a different aspect, specifically, the high skewness between the number of popular and unpopular routes in the global routing table. We leveraged the novel concept of the programmable data planes to design a Programmable FIB Caching Architecture to address the main concerns regarding FIB caching, such as cache victim selections and cache-miss lookup latencies. The corresponding paper has been published at ACM Symposium on Architectures for Networking and Communications Systems (ANCS) conference in 2018 [3] and was presented by my co-author, Dr. Yaoqing Liu. Finally, I proposed a new algorithm that combines FIB aggregation and caching in a single architecture for higher cache-hit ratios (up to 99.9%).

Modifying a FIB by aggregating it or distributing it into separate memory units (cache and slow memory) should not change the forwarding behavior of a router. To quickly and efficiently verify the correctness of FIB aggregation and caching, I designed VeriTable algorithm that can simultaneously verify the equivalence of multiple routing tables. We extended VeriTable to perform network diagnosis, e.g., finding loops or leaks in a private network and verifying the incremental BGP update handling at a router. The corresponding paper was published at IEEE INFOCOM in 2018 [4] and was presented by myself. Additionally, the extended version of this paper was published in the "Computer Networks" journal. Finally, a US patent was issued for "VeriTable".

I have been working on the problem of the gap between the application layer attacks and Network Intrusion Prevention Systems. We presented a Software-Defined Network (SDN) based architecture that would enable cooperative intrusion prevention through an SDN controller. This work was part of a joint research project with the Air Force Research Lab, Rome (NY). The corresponding paper with my co-authorship has been published at IEEE International Conference on Communications in 2018 [5] and was presented by my co-author, Dr Yaoqing Liu. To offload the SDN controller from managing the blacklist at the data plane, we leveraged the programmable data plane concept and designed a new protocol for in-network layer-7 attack mitigation, where the end-hosts under attack send an alert message into the network for edge switches to block the attack. We designed LAMP architecture, specifically, Prompt Layer 7 Attack Mitigation with Programmable Data Planes. The corresponding short conference paper was published at IEEE Network Computing and Applications (NCA) 2019 [6] and presented by myself. Also, a related poster paper was published at ACM ANCS in 2019. Another research direction is leveraging deep packet inspection in a programmable data plane to perform packet filtering. A demo paper related to this research has been published at a prestigious IEEE ICNP conference [12].

We study different applications of Software-Defined Networking and the Programmable Data Plane for in-network engineering of traffic forwarding. More specifically, we designed GRASP, A green energy-aware SDN platform, for smart data center scheduling based on green energy supplies at the end hosts. The corresponding paper for that work was published at the IEEE INFOCOM CNERT workshop [7] in 2018 and presented by myself, while the poster was presented at the GENI Engineering Conference in Miami (2017). In addition to that, I am currently working on the methods of in-network end-host-driven load balancing via a programmable data plane without an SDN controller involvement. Several papers related to this research direction have been published at flagship conferences [9][10]. One of these papers was presented by myself at the 2023 IEEE International Conference on High Performance Switching and Routing in Albuquerque, NM.

During my studies at the Rochester Institute of Technology, I was working on solving the challenges for Remote Direct Memory Access (RDMA) in a virtual containerized environment. This project has been funded by Cisco Systems and led us to design a novel architecture for integrating the container orchestrator Kubernetes and RDMA. The outcome of this work has been accepted as a poster paper at the IEEE International Conference on Network Protocols in 2019 [8] and was presented by myself. A full conference paper has been published at the CANOPIE-HPC conference [11].

[1] Y. Liu and G. Grigoryan, "Toward Incremental FIB Aggregation with Quick Selections (FAQS)", IEEE NCA, 2018.

[2] G. Grigoryan and Y. Liu, "Toward Incremental FIB Aggregation with Quick Selections (FAQS)", 14th ACM/IEEE ANCS poster paper, 2018.

[3] G. Grigoryan and Y. Liu, "PFCA: A Programmable FIB Caching Architecture", 14th ACM/IEEE ANCS, 2018.

[4] G. Grigoryan, Y. Liu, M. Leczinsky, and J. Li, "VeriTable: Fast Equivalence Verification of Multiple Large Forwarding Tables", IEEE INFOCOM 2018.

[5] G. Grigoryan, Y. Liu, L. Njilla, C. Kamhoua and K. Kwiat , "Enabling Cooperative IoT Security via Software Defined Networks (SDN)", IEEE ICC 2018.

[6] G. Grigoryan and Y. Liu, "LAMP: Prompt Layer 7 Attack Mitigation With Programmable Data Planes", IEEE NCA, 2018.

[7] G. Grigoryan, K. Bahmani, G. Schermerhorn, Y. Liu, "GRASP: a GReen energy Aware SDN Platform", IEEE INFOCOM, 2018.

[8] C. Link, J. Saran, G. Grigoryan, M. Kwon, and M. M. Rafique, "Container Orchestration by Kubernetes for RDMA Networking", IEEE ICNP poster paper, 2019.

[9] G. Grigoryan, Y. Liu and M. Kwon, "iLoad: In-network load balancing with programmable data plane", ACM CoNEXT Students Workshop, 2019.

[10] G. Grigoryan and M. Kwon, "Towards Greener Data Centers via Programmable Data Plane", IEEE HPSR, 2023.

[11] G. Grigoryan, M. Kwon and M. M. Rafique, "Extending the Control Plane of Container Orchestrators for I/O Virtualization", IEEE Canopie-HPC, 2021.

[12] S. Gupta, D. Gosain, G. Grigoryan, M. Kwon and H. B. Acharya, "Demo: Simple Deep Packet Inspection with P4", IEEE ICNP 2021.