A Cloud-aware distributed parallel compiler
Project Status: Set-up
Start Date: November 2016
End Date: February 2019
Budget (total): 5054 K€
Effort: 76.3 PY
Name: Jose Javier Garcia Aranda
Company: Alcatel-Lucent España S.A.
E-mail: jose_javier.garcia_aranda(Replace this parenthesis with the @ sign)nokia.com
Alcatel-Lucent España SA, Spain
Telefónica Moviles, Spain
INNOVATI SERVICIOS TECNOLÓGICOS, S.L., Spain
Bor Software, Turkey
DIA A.S, Turkey
Luxembourg Institute of Science and Technology, Luxembourg
BEIA Consult International S.R.L, Romania
Universidad Politécnica de Madrid, Spain
The Project CloudBook proposes to develop an Integrated Development Environment Platform including a core Compiler aware of cloud resources. Any Internet Service, any application, will be dynamically split it into hundreds of small pieces that will be deployed into the available cloud connected hardware belonging to a particular “circle” of available cloud resources. The proposed solution uses a combination of social networks, IoT, cloud computing and grid computing technologies. This new approach in programming technology will develop the tools that take care of the dynamically splitting and the distribution of the software functionality in distributable designs and its deployment operations while programmers can focus on developing and testing the software functionality and its components.
The Project CloudBook will develop innovating ideas on Parallel Programming including the new labelling proposal that allows programmers to label functions at invocation. This splits the original code into small pieces of “Micro services” that can be deployed using different strategies. Labelling functions at invocation is the key for distributed computing. In addition CloudBook considers the entire cloud as a unified machine, represented as a graph of resources belonging to a “circle” (such as the social network concept of “circle”) .This is used to manage the resources. The deployment of micro-services must minimize certain function “cost of execution”, which takes into account CPU timings, traffic latencies, bandwidth consumptions, energy footprint and reliability.
The Project Partners an International consortium that includes Telecom, Research Institutes and SME’s from four different countries propose multiple validation scenarios from Cloud Gaming to Network Management and Process intensive Supercomputing that will allow a clear exploitation path for the results of the Project.