Skip to content

EURA NOVA Publications in 2011

As we usually do at the end of the year,  we will try to summarize the activities led at EURA NOVA R&D during this year. In this post I come back on our scientific publications we had this year. Those publications cover distributed data storage, model management & governance in finance and, finally, elastic architectures for cloud infrastructures. Let me briefly introduce you those papers.

Elasticity of distributed storage

T. Dory, B. Mejías, P. Van Roy, N-L Tran. Measuring Elasticity for Cloud Databases,  Cloud Computing 2011 (Second International Conference on Cloud Computing, GRIDs, and Virtualization), Rome, Italy, Sep. 2011.

This paper is the outcome of a Master thesis work we proposed and managed in collaboration with the Université Catholique de Louvain (UCL). We proposed to study and to quantify the notion of elasticity of distributed storages used in cloud. Indeed, the ability to provision dynamically new storage nodes is not enough to qualify a storage as elastic. The add/removal of nodes involves a reaction of the storage software that can significantly impact the performance. The student, T. Dory, proposed a model and a set of metrics/analytics that define this notion. In this paper, he presents his approach and apply his methodology on HBase, MongoDB and Cassandra.

Model management & Governance

S. Skhiri, M. Delbaere, Y. Bontemps, G. de Hemptinne, N-L Tran, Governance issues on heavy models in an industrial context. Advances in Conceptual Modeling. Recent Developments and New Directions ER 2011, Brussels, Belgium, Nov. 2011.

This paper summarizes our contribution in a project of model management we led at SWIFT. SWIFT is responsible for transporting messages between financial institutions, worldwide. They are also responsible for defining the content of those messages through an ISO specification. One of the most important challenge in standard message management is “how can we manage efficiently this base standard for providing efficient standard integration to customers (Banks, and financial institutions)”.  In this paper we describe the solution we put in place for handling the six major governance issues in model management:

  1. Complete life cycle management of messages and market practices
  2. Versioning
  3. Link between the semantic layer (Taxonomy) and message layer
  4. Knowledge and formalization of the market practices
  5. Content management
  6. Auditing tools

 

Elastic architecture

N-L. Tran, S. Skhiri, E. Zimány. EQS: An Elastic and Scalable Message Queue for the Cloud, 3rd International IEEE conference on Cloud computing technology and science (IEEE CloudCom 2011), Athens, Greece Nov. 2011.

With the emergence of cloud computing, on-demand resources usage is made possible. This allows applications to elastically scale out according to the load. One design pattern that suits this paradigm is the event-driven architecture (EDA) in which messages are sent asynchronously between distributed application instances using message queues. However, existing message queues are only able to scale for a certain number of clients and are not able to scale out elastically.

We designed a novel approach for implementing highly scalable and de-centralized message queues. This paper described the architecture, the scaling algorithm and the tests we performed on Amazon EC2.


 

Releated Posts

Calibrate to Interpret

Trustworthy machine learning is driving a large number of the ML community works in order to improve ML acceptance and adoption. In this paper, we show a first link between uncertainty and explainability, by studying the relation between calibration and interpretation.
Read More

Mass Estimation of Planck Galaxy Clusters using Deep Learning

Galaxy cluster masses can be inferred indirectly using measurements from X-ray band, Sunyaev-Zeldovich (SZ) effect signal or optical observations. Unfortunately, all of them are affected by some bias. Alternatively, we provide an independent estimation of the cluster masses from the Planck PSZ2 catalogue of galaxy clusters using a machine-learning method.
Read More