Email: ghobadi at csail dot mit dot edu
Office: 32 Vassar Street, Cambridge, G940
Research Interests
Efficient Systems for Machine Learning
Large-Scale Reconfigurable Networks
High-Performance Cloud Infrastructure
Hardware-Software Co-design
Data Center Networks
Network Optimization
Optical Networks
Teaching
Fall 2024, 2023, 2022, 2021, 2020, 2019: 6.582 (6.829) Computer Networks
Spring 2024, 2023, 2020, 2019: 6.033 System Engineering
Spring 2021: 6.009 Fundamentals of Programming
Accessibility
|
Associate Professor
Computer Science & Artificial Intelligence Laboratory
Department of Electrical Engineering and Computer Science
Massachusetts Institute of Technology
CV
Bio
Manya Ghobadi is an associate professor in the EECS department at MIT. Her research spans different areas in computer networks, focusing on building efficient systems for machine learning and high-performance cloud infrastructure.
Her work has been recognized by the Sloan Fellowship in Computer Science, ACM SIGCOMM Rising Star award, ACM-W Rising Star Award, NSF CAREER award, Optica Simmons Memorial Speakership award, best paper award at the Machine Learning Systems (MLSys) conference, as well as the best dataset, and best paper awards at the ACM Internet Measurement Conference (IMC). Manya received her Ph.D. from the University of Toronto and spent a few years at Microsoft Research and Google prior to joining MIT.
Publications
|
Rail-only: A Low-Cost High-Performance Network for Training LLMs with Trillion Parameters
W. Wang, M. Ghobadi, K. Shakeri, Y. Zhang, N. Hasani
IEEE Hot Interconnects 2024
[paper]
|
|
NetBlocks: Staging Layouts for High-Performance Custom Host Network Stacks
A. Brahmakshatriya, C. Rinard, M. Ghobadi, S. Amarasinghe
ACM PLDI 2024
[paper]
|
|
Cassini: Network-Aware Job Scheduling in Machine Learning Clusters
S. Rajasekaran, M. Ghobadi, A. Akella
USENIX NSDI 2024
[paper]
|
|
On-Fiber Photonic Computing
M. Yang, Z. Zhong, M. Ghobadi
ACM HotNets 2023
[paper]
|
|
Lightning: A Reconfigurable Photonic-Electronic SmartNIC for Fast and Energy-Efficient Inference
Z. Zhong, M. Yang, J. Lang, C. Williams, L. Kronman, A. Sludds, H. Esfahanizadeh, D. Englund, M. Ghobadi
ACM SIGCOMM 2023
[paper]
|
|
TopoOpt: Optimizing the Network Topology for Distributed DNN Training
W. Wang, M. Khazraee, Z. Zhong, M. Ghobadi, Z. Jia, D. Mudigere, Y. Zhang, A. Kewitsch
USENIX NSDI 2023
[paper][slides][demo][code][video]
|
|
InfoShape: Task-Based Neural Data Shaping Via Mutual Information
H. Esfahanizadeh, W. Wu, M. Ghobadi, R. Barzilay, M. Medard
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2023
[paper][slides]
|
|
Emerging Optical Interconnects for AI Systems (Invited) (Optica Jane Simmons Speakership Award)
M. Ghobadi
OFC 2022
[paper]
|
|
Delocalized Photonic Deep Learning on the Internet's Edge
A. Sludds, S. Bandyopadhyay, Z. Chen, Z. Zhong, J. Cochrane, L. Bernstein, D. Bunandar, P. Ben Dixon, S. Hamilton, M. Streshinsky, A. Novack, T. Baehr-Jones, M. Hochberg, M. Ghobadi, R. Hamerly, D. Englund
Science 2022, Vol. 378 (6617)
[paper]
|
|
Congestion Control in Machine Learning Clusters
S. Rajasekaran, M. Ghobadi, G. Kumar, A. Akella
ACM HotNets 2022
[paper]
|
|
Using Trio -- Juniper Networks' Programmable Chipset -- for Emerging In-Network Applications
M. Yang, A. Baban, V. Kugel, J. Libby, S. Mackie, S. S. R. Kananda, C. Wu, M. Ghobadi
ACM SIGCOMM 2022
[paper][slides]
News article: Building a Programmable Platform for Tomorrow's Applications
|
|
ABM: Active Buffer Management in Datacenters
V. Addanki, M. Apostolaki, M. Ghobadi, S. Schmid, L. Vanbever
ACM SIGCOMM 2022
[paper][code]
|
|
Demonstration of WDM-Enabled Ultralow-Energy Photonic Edge Computing
A. Sludds, R. Hamerly, S. Bandyopadhyay, Z. Zhong, Z. Chen, L. Bernstein, M. Ghobadi, D. Englund
OFC 2022
[paper]
|
|
WDM-Enabled Photonic Edge Computing (Best Paper Award)
A. Sludds, R. Hamerly, S. Bandyopadhyay, Z. Chen, Z. Zhong, L. Bernstein, M. Ghobadi, D. Englund
OptoElectronics and Communications Conference (OECC) and International Conference on Photonics in Switching and Computing (PSC) 2022
[paper]
|
|
Cerberus: The Power of Choices in Datacenter Topology Design (A Throughput Perspective)
C. Griner, J. Zerwas, A. Blenk, S. Schmid, M. Ghobadi, and C. Avin
ACM SIGMETRICS 2022
[paper]
|
|
Performance Tradeoffs in Reconfigurable Networks for HPC
M. Y. Teh, Z. Wu, M. Glick, S. Rumley, M. Ghobadi, K. Bergman
Journal of Optical Communications and Networking, 2022
[paper]
|
|
BOW: First Real-World Demonstration of a Bayesian Optimization System for Wavelength Reconfiguration
(Postdeadline Paper)
Z. Zhong, M. Ghobadi, M. Balandat, S. Katti, A. Kazerouni, J. Leach, M. McKillop, Y. Zhang
OFC 2021
[paper] [slides] [code]
|
|
ARROW: Restoration-Aware Traffic Engineering
Z. Zhong, M. Ghobadi, A. Khaddaj, J. Leach, Y. Xia, Y. Zhang
ACM SIGCOMM 2021
[paper][slides][code]
News article: Who Can Bend Light for Cheaper Internet?
|
|
SiP-ML: High-Bandwidth Optical Network Interconnects for Machine Learning Training
M. Khani, M. Ghobadi, M. Alizadeh, Z. Zhu, M. Glick, K. Bergman, A. Vahdat, B. Klenk, E. Ebrahimi
ACM SIGCOMM 2021
[paper][slides][code]
|
|
IOI: In-Network Optical Inference
Z. Zhong, W. Wang, M. Ghobadi, A. Sludds, R. Hamerly, L. Bernstein, D. Englund
ACM SIGCOMM 2021 Workshop on Optical Systems (OptSys 2021)
[paper][slides]
|
|
In-network Aggregation for Shared Machine Learning Clusters (Best Paper Award)
N. Gebara, P. Costa, M. Ghobadi
MLSYS 2021
[paper] [slides]
|
|
A Social Network Under Social Distancing: Risk-Driven Backbone Management During COVID-19 and Beyond
Y. Xia, Y. Zhang, Z. Zhong, G. Yan, C. L. Lim, S. S. Ahuja, S. Bali, A. Nikolaidis, K. Ghobadi, M. Ghobadi
USENIX NSDI 2021
[paper]
|
|
NeuraCrypt: Hiding Private Health Data via Random Neural Networks for Public Training
A. Yala, H. Esfahanizadeh, R. G. L. D'Oliveira, K. R. Duffy, M. Ghobadi, T. S. Jaakkola, V. Vaikuntanathan, R. Barzilay, M. Medard
[paper][code]
|
|
Challenging the Stateless Quo of Programmable Switches
N. Gebara, A. Lerner, M. Yang, M. Yu, P. Costa, M. Ghobadi
ACM HotNets 2020
[paper][slides]
|
|
PINE: Photonic Integrated Networked Energy efficient datacenters (ENLITENED Program) [Invited]
M. Glick, N. C. Abrams, Q. Cheng, M. Teh, Y. Hung, O. Jimenez, S. Liu, Y. Okawachi, X. Meng, L. Johansson, M. Ghobadi, L. Dennison, G. Michelogiannakis, J. Shalf, A. Liu, J. Bowers, A. Gaeta, M. Lipson, and K. Bergman
Journal of Optical Communications and Networking, 2020
[paper]
|
|
Photonic Memory Disaggregation in Datacenters
J. Shalf, G. Michelogiannakis, B. Austin, T. Groves, M. Ghobadi, L. Dennison, T. Gray, M. Glick, K. Bergman
OSA Advanced Photonics Congress 2020
[paper]
|
|
On the Complexity of Traffic Traces and Implications
C. Avin, M. Ghobadi, K. Griner, S. Schmid
ACM SIGMETRICS 2020
[paper]
|
|
Enabling Programmable Transport Protocols in High-Speed NICs
M. T. Arashloo, A. Lavrov, M. Ghobadi, J. Rexford, D. Walker, D. Wentzlaff
USENIX NSDI 2020
[paper]
|
|
OptFlow: A Flow-based Abstraction for Programmable Topologies
K. Foerster, L. Luo, M. Ghobadi
ACM SOSR 2020
[paper]
|
|
FAB: Toward Flow-aware Buffer Sharing on Programmable Switches
M. Apostolaki, L. Vanbever, M. Ghobadi
Buffer Sizing Workshop 2019
[paper] [slides]
|
|
FlexEnt: Entropy Coding to Curb Stragglers in Large-Scale Distributed Machine Learning
J. Salamy, A. Sharma, M. Ghobadi, M. Medard
SOSP's AI Systems Workshop 2019
[paper] [poster]
|
|
TeaVaR: Striking the Right Utilization-Availability Balance in WAN Traffic Engineering
J. Bogle, N. Bhatia, M. Ghobadi, I. Menache, N. Bjorner, A. Valadarsky, M. Schapira
ACM SIGCOMM 2019
[paper] [slides]
[code] [live demo]
[poster]
News article: Wall Street Secrets Could Cut Cloud Costs
|
|
Bandwidth Steering for HPC using Silicon Nanophotonics
G. Michelogiannakis, Y. Shen, M. Y. Teh, X. Meng, B. Aivazi, T. Groves, J. Shalf, M. Glick, M. Ghobadi, L. Dennison, K. Bergman
ACM/IEEE Supercomputing Conference 2019
[paper]
|
|
RADWAN: Rate Adaptive Wide Area Network
R. Singh, M. Ghobadi, K. Foerster, M. Filer, P. Gill
ACM SIGCOMM 2018
[paper] [slides]
|
|
Characterizing the Algorithmic Complexity of Reconfigurable Data Center Architectures
K. Foerster, M. Ghobadi, S. Schmid
ACM/IEEE ANCS 2018
[paper] [slides]
|
|
Beyond SmartNICs: Towards a Fully Programmable Cloud
A. Caulfield, P. Costa, M. Ghobadi
IEEE High Performance Switching and Routing 2018
[paper]
|
|
PINE: An Energy Efficient Flexibly Interconnected Photonic Data Center Architecture for Extreme Scalability
K. Bergman, J. Shalf, G. Michelogiannakis, S. Rumley, L. Dennison, M. Ghobadi
IEEE Optical Interconnects 2018
[paper]
|
|
Understanding and Mitigating Packet Corruption in Data Center Networks
D. Zhuo, M. Ghobadi, R. Mahajan, K. Foerster, A. Krishnamurthy, T. Anderson
ACM SIGCOMM 2017
[paper][slides]
|
|
RAIL: A Case for Redundant Arrays of Inexpensive Links in Data Center Networks
D. Zhuo, M. Ghobadi, R. Mahajan, A. Phanishayee, X. K. Zou, H. Guan, A. Krishnamurthy, T. Anderson
USENIX NSDI 2017
[paper][slides][code]
|
|
Run, Walk, Crawl: Towards Dynamic Link Capacities
R. Singh, M. Ghobadi, K. Foerster, M. Filer, P. Gill
ACM HotNets 2017
[paper][HotNets dialogue] [Press]
|
|
HotCocoa: Hardware Congestion Control Abstractions
M. T. Arashloo, M. Ghobadi, J. Rexford, D. Walker
ACM HotNets 2017
[paper][code]
|
|
Optical Layer Failures in a Large Backbone (Best dataset award)
M. Ghobadi, R. Mahajan
ACM IMC 2016
[paper][dataset]
|
|
ProjecToR: Agile Reconfigurable Data Center Interconnect
M. Ghobadi, R. Mahajan, A. Phanishayee, N. Devanur, J. Kulkarni, G. Ranade, P. Blanche, H. Rastegarfar, M. Glick, D. Kilper
ACM SIGCOMM 2016
[paper]
[slides.pdf][slides.pptx][Disco Ball tech. report]
|
|
ECN or Delay: Lessons Learnt from Analysis of DCQCN and TIMELY (Best Paper Award Finalist)
Y. Zhu, M. Ghobadi, V. Misra, J. Padhye
ACM CoNEXT 2016
[paper][slides]
|
|
Evaluation of Elastic Modulation Gains in Microsoft's Optical Backbone in North America
M. Ghobadi, J. Gaudette, R. Mahajan, A. Phanishayee, B. Klinkers, D. Kilper
Optical Fiber Communication Conference (OFC)'16
[paper][slides]
|
|
Elastic Optical Networking in the Microsoft Cloud
M. Filer, J. Gaudette, M. Ghobadi, R. Mahajan, T. Issenhuth, B. Klinkers, J. Cox
Journal of Optical Communications and Networking (JOCN)'16
[paper][images/]
|
|
TIMELY: RTT-based Congestion Control for the Datacenter
R. Mittal, V. Lam, N. Dukkipati, E. Blem, H. Wassel, M. Ghobadi, A. Vahdat, Y. Wang, D. Wetherall, D.
Zats
ACM SIGCOMM 2015
[paper][public review][code]
|
|
Efficient Traffic Splitting on Commodity Switches
N. Kang, M. Ghobadi, J. Reumann, A. Shraer, J. Rexford
ACM CoNEXT 2015
[paper][slides]
|
|
TCP Pacing in Data Center Networks
M. Ghobadi, Y. Ganjali
IEEE Hot Interconnects 2013
[paper][slides]
|
|
Trickle: Rate Limiting YouTube Video Streaming (Google Research Excellent Paper Award)
M. Ghobadi, Y. Cheng, A. Jain, M. Mathis
USENIX ATC 2012
[paper][slides]
[demo]
|
|
Rethinking End-to-End Congestion Control in Software-Defined Networks
M. Ghobadi, S. Yeganeh, Y. Ganjali
ACM HotNets 2012
[paper][slides]
|
|
Caliper: Precise and Responsive Traffic Generator
M. Ghobadi, G. Salmon, Y. Ganjali, M. Labrecque, G. Steffan
IEEE Hot Interconnects 2012
[paper][slides][demo]
|
|
Proportional Rate Reduction for TCP (Google Research Excellent Paper Award)
N. Dukkipati, M. Mathis, Y. Cheng, M. Ghobadi
ACM IMC 2011
[paper][RFC]
|
|
OpenTM: Traffic Matrix Estimator for OpenFlow Networks
A. Tootoonchian, M. Ghobadi, Y. Ganjali
Passive and Active Measurement conference 2010
[paper]
|
|
Emulation of Optical PIFO Buffers
H. Rastegarfar, M. Ghobadi, Y. Ganjali
IEEE GLOBECOM 2009
[paper]
|
|
NetThreads: Programming NetFPGA with Threaded Software (2nd Place for Best Project Award)
M. Labrecque, J. Steffan, G. Salmon, M. Ghobadi, Y. Ganjali
NetFPGA Developers Workshop 2009
[paper][code]
|
|
NetFPGA-based Precise Traffic Generation
G. Salmon, M. Ghobadi, Y. Ganjali, M. Labrecque, J. Steffan
NetFPGA Developers Workshop 2009
[paper][tech. report][code][documentation]
|
|
Experimental study of router buffer sizing (Best Paper Award)
N. Beheshti, Y. Ganjali, M. Ghobadi, N. McKeown, and G. Salmon
ACM IMC 2008
[paper]
|
|
Resource Optimization Algorithms for Virtual Private Networks Using the Hose Model
M. Ghobadi, S. Ganti, G. C. Shoja
Computer Networks Journal 2008
[paper]
|
|
Performing Time-sensitive Network Experiments
N. Beheshti, Y. Ganjali, M. Ghobadi, J. Naous, N. McKeown, and G. Salmon
ANCS 2008
[paper]
|
|
Hierarchical Provisioning Algorithm for Virtual Private Networks Using the Hose Model
M. Ghobadi, S. Ganti, G. C. Shoja
IEEE GLOBECOM 2007
[paper]
|
|
Resource Optimization to Provision a Virtual Private Network Using the Hose Model
M. Ghobadi, S. Ganti, G. C. Shoja
IEEE ICC 2007
[paper]
|
Postdocs
Dr. Lingling Fan (Ph.D. Stanford University, co-advised with Prof. Dirk Englund)
Dr. Zhizhen Zhong (Ph.D. Tsinghua University)
Dr. Homa Esfahanizadeh (Ph.D. UCLA, co-advised with Prof. Muriel Medard)
Current Students
Mingran Yang (Ph.D. candidate)
Sudarsanan Rajasekaran (Ph.D. candidate)
Weiyang (Frank) Wang (Ph.D. candidate)
Benoit Pit--Claudel (Ph.D. candidate, co-advised with Prof. Muriel Medard)
Charlie Liu (Ph.D. candidate, co-advised with Prof. Hari Balakrishnan)
Anton Zabreyko (Ph.D. candidate)
Sanjoli Narang (Ph.D. candidate)
James Lynch (Ph.D. candidate, co-advised with Prof. Hari Balakrishnan)
Andew Li (Undergraduate student)
Alumni
Jeremy Bogle (M.Eng. 2019, now at Apple, Thesis: A Fundamental Approach for Providing Service-Level Guarantees for Wide-Area Networks)
Ayush Sharma (M.Eng. 2020, Thesis: Aikido: Toward Straggler Mitigation for Distributed Machine Learning Training in Cloud Data Centers)
Amir Farhat (SuperUROP, Fall 2019 -- Spring 2020, Project: Communication Patterns in Distributed Deep Learning)
Tenzin Ukyab (UROP, Fall 2019 -- Spring 2020, now PhD student at UC Berkeley)
James Salamy (M.Sc. 2021, Thesis: Network Requirements for Distributed Machine Learning Training in the Cloud)
Dr. Moein Khazraee (Postdoc, Summer 2020 -- Summer 2022, now at NVIDIA)
Alex Mallery (Ph.D. student, Fall 2021 -- Summer 2022, now at Skydio)
Christian T Williams (M.Eng. 2023, now at Jane Street, Thesis: LightSpeed: A Framework to Profile and Evaluate
Inference Accelerators at Scale)
Yichen Gao (Undergraduate student)
Arthur Migdal (Undergraduate student)
Liam Kronman (Undergraduate student)
Professional Activities
Member of the Steering Committee, USENIX NSDI, 2024
Program Co-Chair, USENIX NSDI, 2023
General Co-Chair, ACM HotNets, 2023
Workshop Co-Chair, Optica OFC workshop on Machine Learning and Optics, 2023
Program Committee, ACM SIGCOMM, 2022
Program Committee, USENIX NSDI, 2022
Program Chair, ACM SIGCOMM Workshop on Optical Systems (OptSys), 2021
Program Committee, ACM SOSP, 2021
Program Committee, USENIX OSDI, 2021
Co-chair, Optics and Machine Learning Symposium, OFC, 2021
Program Committee, USENIX NSDI, 2021
Program Committee, OFC, 2021
Workshops Chair, ACM SIGCOMM, 2020
Program Committee, ACM SIGCOMM, 2020
Program Committee, USENIX NSDI, 2020
Program Committee, USENIX ATC, 2020
Program Committee, ACM HotNets, 2019
Program Committee, ACM SIGCOMM, 2019
Program Chair, New England Network and Systems (NENS), 2019
Program Chair, ACM SIGCOMM Workshop on Optical Systems (OptSys), 2019
Program Committee, USENIX NSDI, 2019
Program Committee, OSA Photonic Networks and Devices, 2018
Program Committee, ACM CoNEXT, 2018
Program Committee, ACM IMC, 2018
Program Committee, ACM HotNets, 2018
Program Committee, ACM SOSR, 2018
General Chair, ACM SOSR, 2018
Program Committee, APNet, 2018
Program Committee, IEEE HPSR, 2018
Program Committee, ACM HotMobile, 2018
Mentoring co-chair, N2Women workshop at ACM SIGCOMM, 2017
Program Committee, ACM SIGCOMM, 2017
Program Committee, ACM IMC, 2017
Program Committee, ACM SOSR, 2017
Program Committee, ACM CoNEXT, 2017
Program Committee, IEEE ICNP, 2017
Program Committee, IEEE/IFICO TMA, 2017
Program Committee, APNet, 2017
Program Committee, PAM, 2017
Program Committee, IEEE MASCOTS, 2017
Program Committee, USENIX NSDI, 2016
Program Committee, ACM CoNEXT, 2016
Program Committee, IEEE Hot Interconnects, 2016
Program Committee, ACM SOSR, 2015
Program Committee, ACM CoNEXT, 2015
Program Committee, IEEE Hot Interconnects, 2015
Program Committee, ACM CoNEXT student workshop, 2013
Copyright Notice
This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.
|