I am a postdoctoral researcher at the Max Planck Institute for Software Systems (MPI-SWS), Germany. I work with Jonathan Mace in the Cloud Software Systems Group. I recently submitted my Ph.D. thesis titled -- Towards “Ultra-Reliable” CPS: Reliability Analysis of Distributed Real-Time Systems -- under the supervision of Björn B. Brandenburg, head of the Real-Time Systems Group at MPI-SWS.
Prior to joining MPI-SWS, I was a software development engineer in the Cloud Networking Group at Citrix R&D, India. I finished my undergraduate studies in computer science from Birla Institute of Technology and Science (BITS Pilani), India.
Our paper on Serving DNNs like Clockwork: Performance Predictability from the Bottom Up will appear at OSDI 2020!
Our idea paper New Wine in an Old Bottle: N-Version Programming for Machine Learning Components will appear at WoSoCER 2020, co-located with ISSRE 2020!
I will be joining Universitiy of British Columbia's CS department as a Research Associate starting January 2021!
I am broadly interested in real-time systems, distributed systems, fault tolerance, reliability analysis, and scheduling problems.
At present, I am working on building efficient systems for machine learning in the cloud.
As part of my Ph.D. thesis, I worked on quantitative reliability analysis of distributed real-time systems, such as networked control systems, in the presence of environmentally-induced transient faults. I have had the pleasure of working closely with Rupak Majumdar and Mitra Nasri on this project.
During the initial years of my Ph.D., I worked with my colleague Felipe Cerqueira on formalizing the processor affinity concept from the perspective of hard real-time operating systems, and on providing schedulability guarantees for the same.
I have also worked briefly on scheduling problems in the cloud domain. During my internship at Microsoft Resarch, I worked on the problem of distributed auto-scaling for Microsoft Azure's inference serving platform in collaboration with Kathryn S. McKinley, Sameh Elnikety, and Yuxiong He. More recently, I worked on the problem of virtual machine scheduling for datacenter applications with my colleague Manohar Vanga.
Until last year, I used to organize the Real-Time Systems Reading Group at MPI-SWS. You may find the reading list here.
Arpan Gujarati, Reza Karimi, Safya Alzayat, Wei Hao, Antoine Kaufmann, Ymir Vigfusson, and Jonathan Mace, "Serving DNNs like Clockwork: Performance Predictability from the Bottom Up", Proceedings of the 14th USENIX Symposium on Operating Systems Design and Implementation (OSDI 2020), to appear, November 2020. PDF | Artifact
Arpan Gujarati, Sergey Bozhko, and Björn B. Brandenburg, “Real-Time Replica Consistency over Ethernet with Reliability Bounds”, Proceedings of the 26th IEEE Real-Time and Embedded Technology and Applications Symposium (RTAS 2020), April 2020. Distinguished Paper Award. PDF | Extended Version | Slides
Arpan Gujarati, Mitra Nasri, Rupak Majumdar, and Björn B. Brandenburg, “From Iteration to System Failure: Characterizing the FITness of Periodic Weakly-Hard Systems”, Proceedings of the 31st Euromicro Conference on Real-Time Systems (ECRTS 2019), July 2019. PDF | Extended Version | Slides
Arpan Gujarati, Mitra Nasri, and Björn B. Brandenburg, “Quantifying the Resiliency of Fail-Operational Real-Time Networked Control Systems”, Proceedings of the 30th Euromicro Conference on Real-Time Systems (ECRTS 2018), July 2018. Best Presentation Award. PDF | Extended Version | Slides | Poster
Manohar Vanga, Arpan Gujarati, and Björn B. Brandenburg, “Tableau: A High-Throughput and Predictable VM Scheduler for High-Density Workloads”, Proceedings of the 13th European Conference on Computer Systems (EuroSys 2018), April 2018. PDF
Arpan Gujarati, Sameh Elnikety, Yuxiong He, Kathryn S. McKinley, and Björn B. Brandenburg, “Swayam: Distributed Autoscaling to Meet SLAs of Machine Learning Inference Services with Resource Efficiency”, Proceedings of the 18th International Middleware Conference (Middleware 2017), December 2017. Best Student Paper Award. Paper | Appendix | Slides
Arpan Gujarati, and Björn B. Brandenburg, “When is CAN the Weakest Link? A Bound on Failures-In-Time in CAN-Based Real-Time Systems”, Proceedings of the 36th IEEE Real-Time Systems Symposium (RTSS 2015), pp. 249–260, December 2015. PDF | Slides | Poster
Felipe Cerqueira, Arpan Gujarati, and Björn B. Brandenburg, “Linux's Processor Affinity API, Refined: Shifting Real-Time Tasks towards Higher Schedulability”, Proceedings of the 35th IEEE Real-Time Systems Symposium (RTSS 2014), pp. 249-259, December 2014. PDF | Slides
Arpan Gujarati, Felipe Cerqueira, and Björn B. Brandenburg, “Schedulability Analysis of the Linux Push and Pull Scheduler with Arbitrary Processor Affinities”, Proceedings of the 25th Euromicro Conference on Real-Time Systems (ECRTS 2013), pp. 69-79, July 2013. Outstanding Paper Award. PDF (revision 1) | Slides | Poster
Erratum: the paper has been revised to account for inaccuracies in Lemmas 2 and 4 in Section IV of the original ECRTS 2013 paper. A detailed erratum is provided in Appendix A of the revised paper.
Arpan Gujarati, Felipe Cerqueira, Björn B. Brandenburg, and Geoffrey Nelissen, “Correspondence article: a correction of the reduction-based schedulability analysis for APA scheduling”, Real-Time Systems, August 2018. PDF | Springer Link
Arpan Gujarati, Felipe Cerqueira, and Björn B. Brandenburg, “Multiprocessor Real-Time Scheduling with Arbitrary Processor Affinities: From Practice to Theory”, Real-Time Systems, Volume 51, Issue 4, pp. 440–483. Springer Verlag, 2015. PDF | Springer Link
Arpan Gujarati, Sathish Gopalakrishnan, and Karthik Pattabiraman, “New Wine in an Old Bottle: N-Version Programming for Machine Learning Components”, Proceedings of the 10th IEEE International Workshop on Software Certification (WoSoCer 2020), October 2020. PDF | Slides
Arpan Gujarati, Malte Appel, and Björn B. Brandenburg, “Work-in-Progress: Achal: Building Highly Reliable Networked Control Systems”, Proceedings of the 15th ACM International Conference on Embedded Software (EMSOFT 2019), October 2019. PDF | Slides | Poster
Mitra Nasri, Arpan Gujarati, and Björn B. Brandenburg, “Using Schedule-Abstraction Graphs for the Analysis of CAN Message Response Times”, Proceedings of the 3rd Workshop on Security and Dependability of Critical Embedded Real-Time Systems (CERTS 2018), June 2018. PDF | Slides
Arpan Gujarati, Mitra Nasri, and Björn B. Brandenburg, “Lower-Bounding the MTTF for Systems with (m, k) Constraints and IID Iteration Failure Probabilities”, Proceedings of the 2nd Workshop on Security and Dependability of Critical Embedded Real-Time Systems (CERTS 2017), December 2017. PDF | Technical Report | Slides
Malte Appel, Arpan Gujarati, and Björn B. Brandenburg, “A Byzantine Fault-Tolerant Key-Value Store for Safety-Critical Distributed Real-Time Systems”, Proceedings of the 2nd Workshop on Security and Dependability of Critical Embedded Real-Time Systems (CERTS 2017), December 2017. PDF | Slides
Program Committee Member: RTEST WiP 2018, RTAS BP Track 2019 and 2020, ECRTS AE 2019, Middleware DW 2020
Journal Reviewer: TECS (2019), TDSC (2019)
External Reviewer for Conferences: ECRTS (2013-2015, 2019), EuroSys (2013, 2016, 2019), EMSOFT (2020), Middleware (2018), RTAS (2013, 2014, 2016), RTNS (2014-2016), RTSS (2013, 2016, 2018, 2020), Systor (2015, 2016)
Teaching Assistant: Data Structures and Algorithms (BITS Pilani, 2010), Foundations of Cyber-Physical Systems (MPI-SWS & TU-Kaiserslautern, 2014), Distributed Systems (MPI-SWS & Saarland University, 2016), Operating Systems (MPI-SWS & Saarland University, 2017)
arpanbg [AT] mpi-sws [DOT] org
MPI-SWS
Paul-Ehrlich-Strasse
Building G 26
D-67663 Kaiserslautern
Germany
+49 681 9303-8410