Zsolt István

Research Profile

The fundamental question I want to answer in my research is how to use specialization to speed up database operations and data processing in the datacenter. See presentations about my work on this Youtube playlist.

My dissertation explores how we can reduce data movement bottlenecks in large distributed systems by pushing computation closer to storage. The outcome is Caribou: a distributed key-value store that runs entirely on FPGAs. It provides replication for fault tolerance and near-data filtering while meeting the network's line-rate requirements. Caribou is open source and can be used as a starting point for exploring near-data processing for emerging workloads.

Previously I have been working on new ways of implementing high-throughput hash tables in FPGAs. I contributed to our IBEX project (database storage engine running on an FPGA) a hash table which performs fast Group By aggregation, and I worked on an FPGA-based memcached pipeline while on an internship at Xilinx Labs (see my Master thesis).

Prior to that I worked on annotation-based parallelization of objects in a distributed computing framework on top of Java and on distributed random number generation using the BlobSeer storage engine.

Team at IMDEA Software

Previosly working with us:
  • Matei Istoan -- Visiting Post-Doc Researcher until April 2019 (from Imperial College London)
  • Eva Garcia -- Intern until February 2019 (from Bachelor's at UAM)
I am looking for talented people, especially for PhD positions, to grow the team. See the recruitment website for more detail.



Specialize in Moderation -- Building Application-aware Storage Services using FPGAs in the Datacenter.   L. Kuhring, E. Garcia, Zs. István. 11th USENIX Workshop on Hot Topics in Storage and File Systems (HotStorage'19), Renton, WA, USA, July 2019. (to appear)

Something New Under The Sun: Thoughts on Optimizing the Performance of Blockchains. (Position paper).   Zs. István. 9th Workshop on Systems for Multi-core and Heterogeneous Architectures co-located with EuroSys'19, Dresden, DE, 2019. [pdf]

The Glass Half Full: Using Programmable Hardware Accelerators in Analytics.   Zs. István. IEEE Data Engineering Bulletin, March 2019. [pdf] [slides: IMDEA seminar]

Design Patterns for Code Reuse in HLS Packet Processing Pipelines.   H. Eran, L. Zeno, Zs. István and M. Silberstein. 27th IEEE International Symposium on Field-Programmable Custom Computing Machines (FCCM'19), San Diego, USA, 2019. [pdf]


StreamChain: Do Blockchains Need Blocks? (Workshop).   Zs. István, A. Sorniotti, M. Vukolić. 2nd Workshop on Scalable and Resilient Infrastructures for Distributed Ledgers (SERIAL 2018) [pdf] [slides]

Providing Multi-tenant Services with FPGAs: Case Study on a Key-Value Store.  Zs. István, G. Alonso. A. Singla. 28th International Conference on Field Programmable Logic and Applications (FPL'18), Dublin, Ireland, August 2018. [pdf] [slides]
Code on Github: [code]

A Flexible K-Means Operator for Hybrid Databases.  Z. He, D. Sidler, Zs. István, G. Alonso. 28th International Conference on Field Programmable Logic and Applications (FPL'18), Dublin, Ireland, August 2018[pdf]

Active Pages 20 Years Later: Active Storage for the Cloud.  Zs. István, D. Sidler, G. Alonso. In IEEE Internet Computing July/Aug 2018[pdf]


Caribou: Intelligent Distributed Storage. Zs. Istvan, D. Sidler, G. Alonso. To appear in VLDB 2017, Munich, Germany. [pdf] [slides]
Resources for the larger project: [Code on Github] [Short Video]

Accelerating Pattern Matching Queries in Hybrid CPU-FPGA Architectures. D. Sidler, Zs. Istvan, M. Ewaida, G. Alonso. 2017 ACM SIGMOD/PODS Conference (SIGMOD'17), Chicago, US. [pdf]


Low-Latency TCP/IP Stack for Data Center Applications. D. Sidler, Zs. Istvan, G. Alonso. 26th International Conference on Field Programmable Logic and Applications (FPL'16), Lausanne, Switzerland, September 2016. [pdf] 

Runtime Parameterizable Regular Expression Operators for Databases. Zs. Istvan*, D. Sidler*, G. Alonso. (*=equal contribution).  The 24th IEEE International Symposium on Field-Programmable Custom Computing Machines (FCCM'16), May 2016. [pdf] 

Consensus in a Box: Inexpensive Coordination in Hardware. Zs. Istvan, D. Sidler, G. Alonso, M. Vukolic. 13th USENIX Symposium on Networked Systems Design and Implementation (NSDI '16), March 2016. [pdf] [slides+audio] [slides]


A Hash Table for Line Rate Data Processing. Zs. Istvan, G. Alonso, M. Blott, K. Vissers. ACM Transactions on Reconfigurable Technology and Systems (TRETS) - Special FPL'13 Issue, March 2015. [pdf]


Ibex -- An Intelligent Storage Engine with Support for Advanced SQL Off-loading. L. Woods, Zs. Istvan, G. Alonso. VLDB 2014, Hangzhou, China, September 2014. [pdf]

Histograms as a Side Effect of Data Movement for Big Data. Zs. Istvan, L. Woods, G. Alonso. 2014 ACM SIGMOD/PODS Conference (SIGMOD'14), Snowbird, Utah, US. [pdf]


A Flexible Hash Table Design For 10Gbps Key-value Stores on FPGAs. Zs. Istvan, G. Alonso, M. Blott, K. Vissers. 23rd International Conference on Field Programmable Logic and Applications (FPL'13), Porto, Portugal, 2-4 September 2013. [pdf]

Achieving 10Gbps Line-rate Key-value Stores with FPGAs. M. Blott, K. Karras, L. Liu, K. Vissers, Zs. Istvan, J. Bar. 5th USENIX Workshop on Hot Topics in Cloud Computing (HotCloud'13), San Jose, CA, 25-26 June 2013. [pdf] [slides]

Multi-threaded Active Objects. L. Henrio, F. Huet, Zs. Istvan. 15th International Conference on Coordination models and Languages (COORDINATION 2013), Firenze, Italy, 3-5 June 2013. [pdf]


Adapting Active Objects to Multicore Architectures. L. Henrio, F. Huet, Zs. Istvan, G. Sebestyen. International Symposium on Parallel and Distributed Computing (ISPDC 2011). [pdf]


Systems and Methods for Providing Distributed Tree Traversal Using Hardware-Based Processing (US 20160147779 A1). Kenneth H. Eguro, Zsolt Istvan, Arvind Arasu, Ravishankar Ramamurthy, Kaushik Shriraghav. Patent application filed 11/26/14.

Demos and Posters

Enzian: a Research Computer for Datacenter and Rackscale Computing.
Poster for European Conference on Computer Systems (EuroSys'18), Porto, Portugal.

Caribou: A Platform for Building Smart Storage
[Poster] for European Conference on Computer Systems (EuroSys'17), Belgrade, Serbia, 24-26 April 2017.

doppioDB: A Hardware Accelerated Database
[Demo][Poster] for SIGMOD 2017, Chicago IL, 2017

Specialized Microservers for the Data Center
[Poster] for European Conference on Computer Systems (EuroSys'15), Bordeaux, France, 21-24 March 2015.
[Demo] for 25th International Conference on Field Programmable Logic and Applications (FPL'15), London, UK, September 2015.

Hybrid FPGA-accelerated SQL Query Processing
[Demo] for 23rd International Conference on Field Programmable Logic and Applications (FPL'13), Porto, Portugal, 2-4 September 2013.

Service and Events

2019: Co-chairing SFMA Workshop @Eurosys'19, PC member for Doctoral Workshop @Eurosys'19.

2018: Co-organized birds of a feather session @Middleware'18.


At UPM, Master Universitario en Software y Sistemas:

Building Data Processing Systems with FPGAs, Spring 2019 [website] [seminars@UPM]

Performance Analysis and Modeling of Software Systems, Fall 2018 [website] [seminars@UPM]

In the past, at ETH Zurich:

Advanced Systems Lab, Fall 2017 [website]

Data Modelling and Databases, Spring 2017 [website]

Advanced Systems Lab, Fall 2016 (Head Teaching Assistant) [website]

Data Modelling and Databases, Spring 2016 [website]

Advanced Systems Lab, Fall 2015 [website]

Programmieren und Problemlösen, Spring 2015 [website]

Advanced Systems Lab, Fall 2014 [website]

Programmieren und Problemlösen, Spring 2014 [website]

Advanced Systems Lab, Fall 2013 [website]

Data Modeling and Databases, Spring 2012 [website]