STREAMSTORE's targeted scientific breakthrough is to deliver a highly elastic (rapidly scalable) system, integrating
and coordinating stateful stream processing and key-value storage. This vision requires extending state-of-the-art software techniques
(such as incremental elasticity, measurement-based performance prediction, and lightweight
checkpointing) in the context of distributed stream processing and key-value storage, and aligning them with novel hardware capabilities of emerging micro-server clusters.
An overall view of STREAMSTORE is shown below.
STREAMSTORE advocates better integrating and coordinating scalable data stores and stream-processing systems.
The following poster presentation and extended abstract appearing in IEEE Cluster 2020 provides initial experimental evidence in this direction:
STREAMSTORE builds upon previous research work in a number of technical areas:
Rapid, smooth elasticity and monitoring in multi-layer systems
Incremental elasticity for NoSQL data stores, Antonis Papaioannou and Kostas Magoutis. In 36th IEEE International Symposium on Reliable Distributed Systems (SRDS 2017), Hong Kong, China, September 26-29, 2017. [pdf]
Towards cross-layer monitoring of multi-cloud service-based applications, Chrysostomos Zeginis, Kyriakos Kritikos, Panagiotis Garefalakis, Konstantina Konsolaki, Kostas Magoutis, and Dimitris Plexousakis. In Proceedings of European Conference on Cloud and Service Computing (ESOCC'13), Malaga, Spain, September 11-13, 2013. [pdf]
Predicting performance under specific workload specifications
A measurement-based approach to performance prediction in NoSQL systems, Flora Karniavoura and Kostas Magoutis. In 25th IEEE International Symposium on the Modeling, Analysis, and Simulation of Computer and Telecommunication Systems (MASCOTS 2017), Banff, Canada, September 20-22, 2017. [pdf]
Managing service performance in the Cassandra distributed storage system, Maria Chalkiadaki and Kostas Magoutis. In Proceedings of 5th IEEE International Conference on Cloud Computing Technology and Science (CloudCom 2013), Bristol, UK, December 2-5, 2013. [pdf]
Rapid recovery of streaming operator state
On the impact of log compaction on incrementally checkpointing stateful stream-procesing operators, Aris Chronarakis, Antonis Papaioannou and Kostas Magoutis. In Proceedings of 1st Workshop on Distributed and Reliable Storage Systems (DRSS'19); held in conjunction with SRDS 2019, Lyon, France, October 1-4, 2019.
CEC: Continuous eventual checkpointing for data stream processing operators, Zoe Sebepou and Kostas Magoutis. In Proceedings of 41st IEEE/IFIP International Conference on Dependable Systems and Networks (DSN'11), Hong Kong, China, June 27-30, 2011. [pdf]