From POLICLOUD to DATACLOUD: Cloud, Data Science, Big Data

The DataCloud Interdepartmental Laboratory Project focuses on data analysis and machine learning techniques to find, propose, and apply new, precise solutions to complex problems in various scientific and application fields.

At the Politecnico di Milano, a methodological, technological, and application research institution, a wide variety of skills exist side-by-side that can contribute to data science applications in engineering and architecture, and to the development of algorithms, organisation of data, and technologies for science data itself. Coordination initiatives are in place among various Departments and cooperation with other academic, industrial, and medical/healthcare institutions.

It is from the multi-disciplinary work, generated by the combined expert skills, that the DataCloud Lab produces important benefits for the community of Polimi and outside researchers. These benefits include increased visibility, collaboration projects with other institutions, acknowledgement of scientific contributions, and publications.

DataCloud’s goal is to set up a shared, high-performance calculation infrastructure, open to all applications. DataCloud meets the need to improve and extend approaches and technologies, to support data analysis. The Lab provides opportunities for exploring many application fields that draw a benefit from using data science.

The goals identified when structuring the proposal included modelling and optimising the environment set up, processing urban transformation scenarios, and managing risks, estimating needs and services, and analysing medical and biological data. DataCloud is open to partnerships over different multidisciplinary research questions, and with outside Bodies.

The Laboratory operates in several locations:

• DEIB Server Halls (floor -1 and floor 1), Department of Electronics, Information and Bioengineering, Building 20, Via Ponzio 34/5, Milan
• ASICT Z3 server hall, Bovisa

• The equipment that we envisage acquiring is a Dell PowerEdge R940xa Small Server Rack. This is a high-density server, with 28 processing units, 2 GP-GPU, 1.5 TB of memory and 8 TB of disk space. This equipment’s high density means that it occupies limited space, making installation possible in the PoliCloud machine rooms which are now assigned to DataCloud. The new server integrates with and completes the Policloud Laboratory equipment and the calculation resources made available over time by the University’s DEIB Department.
• DataCloud will use and extend resources acquired over time and dedicated to the Clothing/PoliCloud Interdepartmental Lab. Equipment acquired using the financing for the Policloud Interdepartmental Laboratory in 2014 was later extended with new machines purchased in 2019. It was reorganised into a new structure which provides stability, versatility, security, and increased management simplicity.
• The machines used are four (of ten) Dell R630, four Dell R640, four T640 and two (of six) SuperMicros, dedicated to data storage.
• The pre-existing equipment includes IBM Cluster; 5G testbed (under construction as part of the MSCA Spotlight project, and partnership with Vodafone in the 5G trials).

The DataCloud Lab’s aim is to aggregate technology and algorithms expressed by DMAT and a part of DEIB, with those expressed by the DABC, DASTU, and DEIB departments.
The high-level goal is to build up effective methods and techniques for managing and manipulating data and evaluating approaches for solving scientific and practical problems, with the utmost focus on ethical and privacy aspects.
The equipment financed directly by DataCloud will make PoliCloud functioning more technologically powerful and reliable. Integrated use with the data analysis cluster resources offered by IBM and the 5G testbed is possible during the development phase (i.e. the MSCA SpotLight project).
DataCloud will rely on the following expertise offered by the DEIB and DMAT departments:

• Storage and Computation capacity planning
• Event processing
• Software Engineering
• Machine Learning
• Data Mining
• Privacy, Safety, Security
• Ethics
• Networking
• Statistics and mathematics

The Lab offers computational and data storage solutions, as well as algorithms for processing and analysing large amounts of data. Part of the work relates to theoretical and methodological development of advanced computing solutions. This will be associated with a data storage activity service. This expertise will be made available and exploited in many application and scientific contexts, such as:

• Patient risk stratification in cardiovascular, respiratory, and nervous systems diseases. Tailored therapy. Remote monitoring in fragile patients (DEIB - Bioengineering area)
• Genomic computing, Personalised medicine (DEIB - Informatics area and Bioengineering area)
• Analysis of hydrogeological risks, mobility, accessibility analysis and mapping, urban analytics (DASTU)
• Smart Cities, Buildings and infrastructures (DABC)

DataCloud Lab is open to any other application issue identified by Departments that wish to associate with the proposal later.

The access methods and costs of the related services are being finalised and are based on the PoliCloud experience.
DATACLOUD Lab users are structured personnel, research fellows, PhD students, and thesis writers, who intend to use the infrastructure. No costs for using the resources are planned during the construction phase. Once the initial phase has ended, costs will be determined for the various services, which will be used by outside users, as part of project partnerships.

Management of the DataCloud Lab is done collectively, by the Management Committee which is made up as follows:

• Elisabetta Di Nitto (DEIB Department)
• Maria Gabriella Signorini (DEIB Department)
• Ilenia Epifani (DMAT Department)
• Paola Pucci (DASTU Department)
• Massimo Tadi (DABC Department)

Based on the involved departments’ instructions, the Management Committee appoints an Operating Committee.