Dimokritos Stamatakis

Dimokritos Stamatakis (Dimos)

PhD candidate, Brandeis University





Office: Volen 111

Email: [my-first-name] at [brandeis] dot [edu]

Linkedin profile

Personal Profile

I am currently working toward my Ph.D. with Prof. Liuba Shrira on Main-Memory Databases, aka Transactional Memory Systems. I received my BSc and MSc degrees in Computer Science from the University of Crete, Greece, while working as a research associate at the Computer Architecture and VLSI (CARV) Laboratory at ICS-FORTH. My research interests include Parallel and Distributed Storage Systems, Transactional Memory Systems and Multicore Systems.

Education

PhD candidate, Computer Science Department, Brandeis University

2013 - Present

M.A. in Computer Science, Computer Science Deprartment, Brandeis University

2016

M.Sc in Parallel and Distributed Systems, Computer Science Department, University of Crete, Greece

2012

B.Sc. in Computer Science, Computer Science Department, University of Crete, Greece

2011

Publications

  • Flexible navigation through a multi-dimensional parameter space using Berkeley DB snapshots, Dimokritos Stamatakis, Werner Benger, and Liuba Shrira, in Proceedings of 25th International Conference in Central Europe on Computer Graphics, Visualization and Computer Vision (WSCG ’17), May 2017 [link]. Selected to be published at the Journal of WSCG ’17 [pdf].
  • A General-Purpose Architecture for Replicated Metadata Services in Distributed File Systems, Dimokritos Stamatakis, Nikos Tsikoudis, Eirini Micheli, and Kostas Magoutis, in Journal IEEE Transactions on Parallel and Distributed Systems, 2017 [link].
  • SLA-driven Workload Management for Cloud Databases, Dimokritos Stamatakis, Olga Papaemmanouil, in Proceedings of 6th International Workshop on Cloud Data Management (CloudDB ’14), April 2014 [pdf].
  • Scalability of Replicated Metadata Services in Distributed File Systems, Dimokritos Stamatakis, Nikos Tsikoudis, Ourania Smyrnaki, and Kostas Magoutis, in Proceedings of 12th IFIP International Conference on Distributed Applications and Interoperable Systems (DAIS 2012), Stockholm, Sweden, June 14-15, 2012. [link]
  • Real-time Analysis of Localization Data Streams for Ambient Intelligence Environments, Dimokritos Stamatakis, Dimitris Grammenos, and Kostas Magoutis, in Proceedings of International Joint Conference on Ambient Intelligence (AMI-11), Amsterdam, The Netherlands, November 16-18, 2011. [link]

Paper Presentations

Paper presentation at the 25th International Conference in Central Europe on Computer Graphics, Visualization and Computer Vision (WSCG ’17)

May 29-June 2 2017, Plzen, Czech Republic

Invited to give a talk at Texas A&M University by Dr. Jian Tao. Talk title: Flexible navigation through a high-dimensional parameter space using Berkeley DB snapshots

May 15 2017, College Station, Texas, USA.

Paper presentation at the 6th International Workshop on Cloud Data Management (CloudDB ’14)

March 31-April 4 2014, Chicago, IL, USA

Distinctions and Honors

Paper selected to be published at the Journal of the 25th International Conference in Central Europe on Computer Graphics, Visualization and Computer Vision (WSCG ’17).

Chaired a session at the 25th International Conference in Central Europe on Computer Graphics, Visualization and Computer Vision (WSCG ’17).

Received AWS Research Grant (AWS Cloud Credits for Research) for the Elastic Compute Cloud and Amazon S3 to conduct research with the HDF5 Big Data Project (2015).

Received Gerondelis Foundation Scholarship (2015).

Received scholarship to attend the ACACES Summer School of the EU Network of Excellence on High Performance and Embedded Architecture and Compilation (HiPEAC), 10-16 July 2011, Fiuggi, Italy.

Brandeis University

Graduate Research Associate

1/2013 - Present

I have been working as a research associate in projects involving Transactional Memory Systems, Big Data File Formats (HDF5) and Cloud Computing. Below is an overview of these projects.

Creating Transactional data structures with STO

This work focuses on making any data structure to easily support Database transactions with all ACI properties and leverage the increasing number of processing cores. This is achieved with the use of a Software Transactional Memory System that supports custom size of shared objects between processing threads (Software Transactional Objects - STO), allowing the programmer to control the granularity of conflicts and thus tuning the system to achieve the best performance. Our current focus is to use Bloom Filters and create a hybrid data structure with a normal part that allows updates, and a part being read-only.

Past Projects:

HDF5 - Big Data File Format

This project is developing support for snapshots in HDF5, designing new data structures and algorithms to scale HDF5 data access on modern storage devices to Bigdata. The project is designing several new HDF5 drivers: mapping objects to a Linux file system; storing objects in a database; and accessing data objects on remote Web servers. These improvements are evaluated using large-scale visualization applications with Bigdata, stemming from real-world scientific computations. Detailed description can be found here.

Cloud Databases

Work on a proposal to provide performance extensibility at next generation Cloud Data Processing services. Performance extensibility refers to the ability to declaratively define application-specific performance models and constraints and leverage them to automatically customize the functionality of core application tasks towards meeting the application’s performance objectives. The challenges that this project aims to tackle are:

  • Design performance specification mechanisms that allow application developers to express custom criteria and constraints for the effectiveness and efficiency of their data processing applications and queries, respectively
  • Exploit the properties of these specification mechanisms to design generic resource, workload and SLA management services that can be automatically customized to serve the application-defined performance goals.

ICS-FORTH

Graduate Research Associate

2/2011 - 11/2012

Member of the Computer Architecture and VLSI (CARV) Laboratory of the Institute of Computer Science (ICS), Foundation for Research and Technology – Hellas (FORTH); Contributions include:

  • Design of a general framework to building scalable replicated metadata services in distributed file systems. The overall approach is based on the use of a general-purpose Paxos-compatible replication primitive (the Oracle Berkeley database) along with a methodology for making it interoperate with existing distributed file system metadata services (such as PVFS and HDFS). Personal contributions include the implementation of enhancements to the Berkeley replicated database and the HDFS recovery path, and extensive large-scale evaluation of our versions of PVFS and HDFS on Amazon EC2 deployments under variety of benchmarks focusing on performance and availability under different replication factors.
  • Design and implementation of a monitoring and control system for a Cloud-deployed distributed stream-processing engine. The system measures resource usage and automatically scales up/down the overall system. When system load approaches capacity a new VM is allocated to take over part of the work. When load is reduced, work is redistributed to free up and de-allocate unused VMs. This work leverages a novel fault-tolerance mechanism for stream processing called Continuous Eventual Checkpointing (CEC), to seamlessly provide elasticity with minimal availability (and no data) loss.

Undergraduate Research Associate

6/2010 - 1/2011

Member of the Computer Architecture and VLSI (CARV) Laboratory of the Institute of Computer Science (ICS), Foundation for Research and Technology – Hellas (FORTH); Designed and implemented a novel methodology for performing real-time analysis of localization data streams produced by sensors embedded in ambient intelligence (AmI) environments. The methodology handles different types of real-time events, detects interesting behavior in sequences of such events, and calculates statistical metrics using a scalable stream-processing engine (SPE) that executes continuous queries expressed in a stream-oriented query language. Key personal contributions are (a) the integration of the Borealis SPE into a large-scale interactive museum exhibit system that tracks visitor positions through a number of cameras; and (b) the extension and customization of Borealis to support the types of real-time analysis useful in the context of the museum exhibit as well as in other AmI applications. The results of this research work were presented at the Ambient Intelligence 2011 conference.

Fall 2019

Head Teaching Assistant, CS-131a Operating Systems, Brandeis University

Spring 2019

Teaching Assistant, CS-146a Principles of Computer System Design, Brandeis University

Fall 2018

Head Teaching Assistant, CS-131a Operating Systems, Brandeis University

Spring 2018

Teaching Assistant, CS-147a Distributed Systems, Brandeis University

Fall 2017

Teaching Assistant, CS-131a Operating Systems, Brandeis University

Spring 2017

Teaching Assistant, CS-146a Principles of Computer System Design, Brandeis University

Fall 2016

Teaching Assistant, CS-131a Operating Systems, Brandeis University

Spring 2015

Teaching Assistant and contribution in course design of CS-146a Principles of Computer System Design, Brandeis University

Spring 2014

Teaching Assistant, CS-147a Distributed Systems, Brandeis University

Fall 2013

Teaching Assistant, CS-130a Networked Information Systems, Brandeis University

Fall 2012

Teaching Assistant, CS-590.45 Modern Topics in Scalable Storage Systems, University of Crete

Spring 2012

Teaching Assistant, CS-559 Infrastructure Technologies for Large-Scale Service-Oriented Systems, University of Crete; also offered as English-language course part of the Erasmus Mundus International Master in Service Engineering (IMSE) curriculum

Spring 2011,2012

Teaching Assistant, CS-335 Computer Networks, University of Crete

Spring 2010

Teaching Assistant, MATH-106 Computer Programming, University of Crete

Side Projects

I like to work with embedded systems to support automations for several tasks. I mainly focus in making my aquariums maintenance easier and my projects aim to remotely control and monitor the water conditions of the aquariums. I have also developed a library to stream images from a surveillance camera via OpenCV, connected to the Odroid-XU4. I have developed a compatible iPhone app to connect and view the stream, achieving a high frame rate given the low performance of the embedded systems. I am also voluntarily consulting my friend with his startup company about Home Services.

Hobbies

In my free time I like to do CrossFit and I participated in the 2019 Crossfit Games Open Competition with the CrossFit ONE Nation in Waltham. During the windy days I am out for KiteSurfing at a nearby beach and during the non-windy days I go biking or take care of my aquariums. I have a wide collection of freshwater fish and shrimps, saltwater fish and corals.