Professor Saurabh Bagchi Ph.D. | |
---|---|
Add a Photo | |
Born | Calcutta, India |
Nationality | Indian, American |
Citizenship | United States |
Education |
|
Alma mater |
|
Occupation | Computer science researcher |
Scientific career | |
Fields |
|
Institutions | Purdue University |
Notable students | Gunjan Khanna, Issa Khalil, Nipoon Malhotra |
Website | engineering |
Saurabh Bagchi is an Indian American computer science researcher, currently a Professor at Purdue University in Indiana, USA. There, he leads a university-wide center on resilience called CRISP.[1]. Bagchi is widely cited in the field of dependable computing[2], particularly in the areas of IoT reliability and security, high performance computing reliability, and distributed system reliability. He received his undergraduate degree at the Indian Institute of Technology, Kharagpur and his M.S. and Ph.D. degrees from the University of Illinois at Urbana-Champaign. He also serves as the inaugural International Visiting Faculty at the Indian Institute of Technology, Kharagpur.
Born in Calcutta, India (now Kolkata), Bagchi is an Indian American computer science researcher specializing in the areas of reliability and security of distributed systems. He received his B. Tech. degree in Computer Science and Engineering from the Indian Institute of Technology Kharagpur in 1997[3], MS and PhD degrees in Computer Science from the University of Illinois at Urbana-Champaign in 1998 and 2001, respectively[4]. He spent two years at IBM T. J. Watson Research Center in New York working on the distributed messaging system called Gryphon. From 2002, he has been a faculty member of Electrical and Computer Engineering and Computer Science at Purdue University, becoming tenured in 2008 and Full Professor in 2013.
From 2017, he has served as the Founding Director of the university-wide center called CRISP (Center for Resilient Infrastructures, Systems, and Processes)[1]. CRISP involves 30 faculty members across the Colleges of Engineering, Science, and Agriculture. From 2018, he co-leads the Wabash Heartland Innovation Network (WHIN)[5] project, funded by a $20M grant from the Lilly Endowment. WHIN is an alliance of 10 counties in north-central Indiana headquartered in Tippecanoe County and its mission is to develop this region as a very large scale, living laboratory for education and scientific research related to Internet of Things (IoT) technologies.
Bagchi’s research group called the Dependable Computing Systems Lab[6] has made contributions to dependability of distributed systems by introducing data analytics-based algorithms, moving away from rule-based solutions. To commercialize the technology, he along with two students and a faculty colleague started a company called SensorHound Inc.[7] in 2013.
Bagchi serves on the Editorial Board of IEEE Transactions on Dependable and Secure Computing (TDSC) and regularly serves on the Program Committees of premier Computer Science conferences (DSN, Usenix ATC, ICDCS, Middleware, etc.).
Bagchi’s work on design of practical dependable distributed systems has found expression in various domains: embedded “bare metal” systems, networks of interdependent assets from multiple ownership domains, cellular and wireless systems, enterprise networks, and supercomputing clusters.
Bagchi writes a blog called “Distant Whispers”[8] on matters that relate to technology and society. These posts also appear in the Communications of the ACM Blog[9].
Bagchi's work on a security problem in wireless networks, specifically targeted to multi-hop wireless networks, solved a long-standing problem[10]. The open nature of the wireless communication, the fast deployment practices, and the fact that these networks are deployed in areas that are not physically secured, make them vulnerable to a wide range of security attacks against both control and data traffic. Bagchi’s group was the first to leverage the fact that most wireless embedded networks use omni-directional antennas and therefore neighbors can overhear communication in and out of nodes[11]. They used this to develop a powerful primitive called “local monitoring” which they used for detection of sophisticated attacks, like blackhole or wormhole, which degrade the throughput of large networks to close-to-zero. This primitive has since been used by many researchers[12][13] and commercial wireless packet sniffers[14][15].
Progressing logically to the comparatively resource-rich devices, they unveiled the common failure modes for the OSes on mobile phones (Android)[16][17] and on wearable devices (Wear OS)[18][19]. The first work led to fundamental changes by Google to the exception handling architecture of Android (in Android 4.0, Ice Cream Sandwich, 2011). More recently, the work has led to understanding of vulnerability in wearable OSes (Wear OS) and the vendor patching the vulnerability[20].
Distributed systems and applications are pervasive in today’s world providing the core infrastructures for the largest commercial and scientific applications. With their increasing complexity and scale, it becomes challenging to have efficient error detection without significantly slowing down the main applications and that scale to the size of the largest systems, namely, the country’s largest supercomputers at the Department of Energy (DOE) labs. Debugging production systems also becomes increasingly challenging as the number of concurrent tasks increases overwhelming human cognitive abilities.
The joint work between Lawrence Livermore National Lab (LLNL) and Purdue University developed a targeted approach to allow large-scale runtime systems to isolate regions where faults occur and replicate only those parts of the system, rather than the standard practice of replicating entire computations. In 2010 and 2011, they developed an error-detection and problem-localization technique, called AutomaDeD, that helps application developers find the period of time, task, and code region where a fault is first manifested in an application[21][22].
This technique was used at LLNL to set a new supercomputing record in fluid dynamics running on its Sequoia supercomputer. The team went on to win the 2013 Gorden Bell Prize for outstanding achievement in HPC. The software artifact was released as an open source project jointly by LLNL and Purdue University[23] and has become part of the standard software stack that runs on DOE supercomputers.
A principle from this work has also been shown by us to be effective in recovering from failures in storage systems in data centers[24] and this work has been patented and commercialized by AT&T[25] through incorporating the technique in the ceph file system that is part of its commercial offerings in software-defined storage for the cloud.
{{cite web}}
: CS1 maint: url-status (link)
{{cite web}}
: CS1 maint: url-status (link)
{{cite web}}
: CS1 maint: url-status (link)
{{cite web}}
: CS1 maint: url-status (link)
{{cite web}}
: CS1 maint: url-status (link)
{{cite web}}
: CS1 maint: url-status (link)
{{cite news}}
: CS1 maint: url-status (link)
This article "Saurabh Bagchi" is from Wikipedia. The list of its authors can be seen in its historical. Articles taken from Draft Namespace on Wikipedia could be accessed on Wikipedia's Draft Namespace.