Michael Abebe


Headshot of Michael Abebe.

I now work on distributed databases at Salesforce.

I have recently (2022) defended my PhD from the School of Computer Science at the University of Waterloo. I was advised by Khuzaima Daudjee and was a member of the Data Systems Group.

Generally, my research interests lie in the intersection of distributed systems, databases, and machine learning. I am particularly interested in building adaptive and distributed data systems. My thesis was on the design and development of adaptive data storage and placement in distributed databases.

I received a Bachelor of Computer Science degree from Waterloo in 2016, with a focus on bioinformatics.


Projects

Physical Design within Distributed Data Systems: Modern data systems exist within a cluster of servers, with data distributed across the cluster. Therefore, how data is placed and accessed is critical to application performance. My research focuses on building adaptive data systems that make intelligent and dynamic physical design decisions improve performance. I have applied these techniques in distributed database systems (MorphoSys and DynaMast), HTAP databases (Proteus) and erasure coded storage systems (EC-Store).


Publications

DBLP | Google Scholar

2022

Adaptive Data Storage and Placement in Distributed Database Systems
Michael Abebe
PhD Thesis (Slides)

Tiresias: Enabling Predictive Autonomous Storage and Indexing
Michael Abebe, Horatiu Lazu, Khuzaima Daudjee
Proceedings of the Very Large Databases Endowment (PVLDB) (Slides) (Talk) (Poster) (Poster Talk)

Proteus: Autonomous Adaptive Storage for Mixed Workloads
Michael Abebe, Horatiu Lazu, Khuzaima Daudjee
ACM SIGMOD/PODS International Conference on Management of Data (SIGMOD) (Tech Report) (Slides) (Talk)

2020

MorphoSys: Automatic Physical Design Metamorphosis for Distributed Database Systems
Michael Abebe, Brad Glasbergen, Khuzaima Daudjee
Proceedings of the Very Large Databases Endowment (PVLDB) (Tech Report) (Slides) (Talk)

DynaMast: Adaptive Dynamic Mastering for Replicated Systems
Michael Abebe, Brad Glasbergen, Khuzaima Daudjee
IEEE International Conference on Data Engineering (ICDE) (Tech Report) (Slides) (Talk)

Sentinel: Universal Analysis and Insight for Data Systems
Brad Glasbergen, Michael Abebe, Khuzaima Daudjee, Amit Levi
Proceedings of the Very Large Databases Endowment (PVLDB) (Tech Report)

ChronoCache: Predictive and Adaptive Mid-Tier Query Result Caching
Brad Glasbergen, Kyle Langendoen, Michael Abebe, Khuzaima Daudjee
ACM SIGMOD/PODS International Conference on Management of Data (SIGMOD)

Demo: Sentinel: Understanding Data Systems
Brad Glasbergen, Michael Abebe, Khuzaima Daudjee, Daniel Vogel, Jian Zhao
ACM SIGMOD/PODS International Conference on Management of Data (SIGMOD) Demo Track (Demo video)
BEST DEMO WINNER!

2019

WatDFS: A Project for Understanding Distributed Systems in the Undergraduate Curriculum
Michael Abebe, Brad Glasbergen, Khuzaima Daudjee
ACM Technical Symposium on Computer Science Education (SIGCSE) (Slides)

2018

EC-Store: Bridging the Gap Between Storage and Latency in Distributed Erasure Coded Systems
Michael Abebe, Khuzaima Daudjee, Brad Glasbergen, Yuanfeng Tian
International Conference on Distributed Computing Systems (ICDCS) (Slides)

Tutorial: Adaptive Replication and Partitioning in Data Systems
Brad Glasbergen, Michael Abebe, Khuzaima Daudjee
ACM/IFIP International Middleware Conference (Middleware) (Slides)

Apollo: Learning Query Correlations for Predictive Caching in Geo-Distributed Systems
Brad Glasbergen, Michael Abebe, Khuzaima Daudjee, Scott Foggo and Anil Pacaci
Extending Database Technology (EDBT)

2017

Diversity-generating retroelements: natural variation, classification and evolution inferred from a large-scale genomic survey
Li Wu, Mari Gingery, Michael Abebe, Diego Arambula, Elizabeth Czornyj, Sumit Handa, Hamza Khan, Mingshun Liu, Mechthild Pohlschroder, Kharissa L. Shaw, Amy Du, Huatao Guo, Partho Ghosh, Jeff F. Miller and Steven Zimmerly
Nucleic Acids Research

2013

A pipeline of programs for collecting and analyzing group II intron retroelement sequences from GenBank
Michael Abebe, Manuel A Candales, Adrian Duong, Keyar S Hood, Tony Li, Ryan A E Neufeld, Abat Shakenov, Runda Sun, Li Wu, Ashley M Jarding, Cameron Semper and Steven Zimmerly
Mobile DNA


Experience

Salesforce

I am a Researcher and Senior Member of Technical Staff at Salesforce. I research, design and implement techniques to improve the performance, scalability and efficiency of Salesforce's distributed database.

Facebook

I worked on two different infrastructure teams during my internships (2015) at Facebook.

On the Data Infrastructure team I worked on the next generation data warehouse system. I helped improve operational workflows by building a custom log search tool for distributed debugging. I additionally worked on a shared memory file transfer between the client interface and the storage interface.

On the Service Infrastructure team I worked on service discovery and the internal ZooKeeper system. My work focused on moving service discovery from a daemon process to a service itself. In the process I modified the Thrift compiler to provide transport level request caching.

Palantir

As a Software Engineering Intern (2014) at Palantir I prototyped a web application that allowed users to collaborate during investigations within the Palantir web applications. The application allowed users to directly share the state of their investigations as a first class citizen. I also worked on the maps application within Palantir.


Service

Program Committee Member

VLDB (2022-2023)

External Reviewer

VLDB (2017-2022), SIGMOD (2018-2021), TKDE (2021), ICDE (2017-2020), Middleware (2017)


Awards

I was awarded the Cheriton Distinguished Dissertation Award for my thesis.

I was named a Distinguished Reviewer at VLDB 2023.

I was previously supported by a NSERC CGS-D scholarship and a Facebook Emerging Scholar award.


Teaching

I was a TA for CS 454 (Distributed Systems) in the winter of 2017, 2018, 2019, 2020 and the summer of 2018. In these offerings I gave guest lectures on erasure coded storage systems. I also helped develop a new assignment on distributed file systems, which is described in our WatDFS SIGCSE paper.


Contact

Email: mtabebe@uwaterloo.ca

GitHub: @mtabebe