DBOS: A Database-oriented Operating System
Computer Society and GBC/ACM
Speaker: Michael Stonebraker
Please register in advance for this seminar even if you plan to attend in person at
https://acm-org.zoom.us/webinar/register/1717017075074/WN_RKdRzZZdRPuT0LqrY14JLw
After registering, you will receive a confirmation email containing information about joining the webinar.
Indicate on the registration form if you plan to attend in person. This will help us determine whether the room is close to reaching capacity.
We may make some auxiliary material such as slides and access to the recording available after the seminar to people who have registered.
For the last three years, a team of us at MIT, Stanford, CMU, Google and VMware have built a new operating system stack, based on a high-performance distributed DBMS. In other words, all OS state (files, messages, scheduling information, etc.) is stored in the DBMS and all OS services are written in SQL plus stored procedures. At the present time, we have a secured venture capital funding and are about to release a commercial open source version. In this talk, I report on aspects of our system, including:
Performance: DBOS is competitive with the state of the art concerning file system performance, message performance and scheduling performance. All of these are implemented in SQL.
Provenance: Because all OS state is in DBMS tables, DBOS change capture moves all state to a warehouse DBMS (currently Vertica or Redshift) with a runtime overhead of about 5%. In this case, security queries (e.g. looking for outliers) supporting the “right to be forgotten” in GDPR-style systems and other provenance operations can be coded in SQL. Experiments with a current security product show that DBOS can both capture provenance data and query it with higher performance.
Serverless environment: We have written a Java serverless environment on top of DBOS. It is an order of magnitude faster than current systems (AWS Lambda, Open Whisk) because it co-locates computation and data whenever possible. Also, provenance facilitates a novel time-travel debugger.
Early enterprise usage: I will report on early DBOS usage in enterprise environments at three large enterprises.
Commercialization changes to DBOS: These include moving to TypeScript and open source DBMSs.
Bio: Dr. Stonebraker has been a pioneer of database research and technology for more than forty years. He was the main architect of the INGRES relational DBMS, and the object-relational DBMS, POSTGRES. These prototypes were developed at the University of California at Berkeley where Stonebraker was a Professor of Computer Science for twenty-five years. More recently at M.I.T. he was the co-architect of the C-Store column- oriented DBMS, the H-Store transaction processing engine, the Data Tamer data integration system, the SciDB array processing engine, the Kyrix visualization system and the operating system DBOS. He is the founder of ten venture-capital backed startups which have commercialized his prototypes.
Professor Stonebraker is the author of scores of research papers on database technology, operating systems and the architecture of system software services. He was awarded the ACM System Software Award in 1992, for his work on INGRES. Additionally, he was awarded the first annual Innovation award by the ACM SIGMOD special interest group in 1994 and was elected to the National Academy of Engineering in 1997. He was awarded the IEEE John Von Neumann award in 2005, and the ACM Turing Award in 2014. Presently he is an Adjunct Professor of Computer Science at M.I.T., where he is working on a variety of future-generation data-oriented projects.
This joint meeting of the Boston Chapter of the IEEE Computer Society and GBC/ACM will be hybrid (in person and online), part of getting back to normal after the COVID-19 lockdown.
Up-to-date information about this and other talks is available online at https://ewh.ieee.org/r1/boston/computer/. You can sign up to receive updated status information about this talk and informational emails about future talks at https://mailman.mit.edu/mailman/listinfo/ieee-cs, our self-administered mailing list.