Publication

Publications

The concepts of our novel and innovative BaaS platform are build upon over 15 scientific publications of the research areas Cloud Computing, Databases and Performance Engineering by Jörg Domaschka (external link: > Scholar -Jörg Domaschka)  and Daniel Seybold (external link: > Scholar – Daniel Seybold)

The following publications present the key scientifc concepts of our BaaS platform.

2019: ACM/SPEC International Conference on Performance Engneering (ICPE)

Mowgli: Finding Your Way in the DBMS Jungle

Authors: D. Seybold, M. Keppler, D. Gründler and J. Domaschka

Our „Mowgli“ publication highlights the challenges in performance and scalability benchmarking of cloud-hosted databases with respect to the cloud and database domain. For the cloud domain, challenges such as heterogenous resource types, cloud resource performance and resource inferences are discussed. For the database domain, heterogonous database types and database-specific runtime parameters are discussed.

As a result, „Mowgli“ presents a novel evaluation framework for cloud-hosted databases that fully automates the benchmarking process for any cloud provider and database type.

External link: 

2019: IEEE Iternational Conference onCloud Computing Technology and SCIENCE (CloudCom)

Kaa: Evaluating Elasticity of Cloud-hosted DBMS

Authors: D. Seybold, S. Volpert, S. Wesner, A. Bauer, N. Herbst and J. Domaschka

Our „Kaa“ publication highlights the need for evaluating the elasticity of distributed databases to handle fluctuating workloads.

Therefore, „Kaa“ defines a comprehensive evaluation framework that is capable of evaluating the elasticity of any distributed databases by adapting the cluster topology by considering multiple adaptaitons triggers such as time- or metric-based events.

External link: 

2020: ACM/SIGAPP Symposium on Applied Computing (SAC)

King Louie: Reproducible Availability Benchmarking of Cloud-hosted DBMS

Authors: D. Seybold, S. Wesner and J. Domaschka

High-availability is a crucial requirement of any database but cloud resources might fail over time. Therefore, our  „King Louie“ publications establishes a novel concept to evaluate the high-availability mechanisms of distributed databases in the case of cloud resource failures. The resulting evaluation framework enables the injection of cloud resource failures into database clusters and their availability evaluation based on a novel set of availability metrics.

These concepts are integrated into the BaaS platform and extended with respect to supported failure types and recovery actions.

External links 

2020: MASCOTS: International Symposium on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems

Baloo: Measuring and Modeling the Performance Configurations of Distributed DBMS

Authors: J. Grohmann, D. Seybold, S. Eismann, M. Leznik, S. Kounev and J. Domaschka

Our „Baloo“ publication presents a first approach towards establishing holistic performance models of cloud-hosted databases. These performance models consider not only the raw performance metrics but also database and cloud resource characteristics. Building upon such novel performance models enables more significant benchmarks results and the opportunity to predict the performance of additional database configurations.

These concepts are integrated into the BaaS platform to ensure you reliable benchmark results and an increasing space of evaluation configurations..

External links: 

2020: ACM/IEEE INTERNATIONAL CONFERENCE ON UTILITY AND CLOUD COMPUTING (UCC)

Hathi: An MCDM-based Approach to Capacity Planning for Cloud-hosted DBMS

Authors: J. Domaschka, S. Volpert and D. Seybold

Selecting the optimal database and cloud resource is a complex challenge, due to heterogenous performance, scalability, elasticity and availability metrics as well as operational cloud costs. Therefore, our „Hathi“ publication applies established Multi-Criteria-Decision-Making (MCDM) concepts to unify multiple evaluation metrics and cloud costs into a novel and unified score based on weighted preferences.

These concepts are part of our BaaS platform and ease your decision process by allowing you to specify and weight relevant benchmark metrics and our BaaS platform will calculate the unified score.

External links: 

> available 11. December 2020

Ready to use the optimal Cloud resources?