Titulo Estágio
Auto-Adaptable Cloud Big Data Processing
Áreas de especialidade
Engenharia de Software
Local do Estágio
DEI/CISUC/SSE and CMU (USA)/Institute of Software Research
Enquadramento
Big data brings new opportunities for discovering new values, helps us to gain an in-depth understanding of the hidden values, and incurs new challenges, e.g., on how to effectively organize and manage such data.
Every big data system faces one fundamental challenge: scalability. The common assumption that even the most obscure data might contain some value, and the resulting obsession to store as much as possible, make predicting data volumes essentially impossible. The evolution of big data was driven by the rapid growth of application demands and cloud computing developed from virtualized technologies. Therefore, cloud computing not only provides computation and processing for big data, but also itself is a service mode. However, as application developers began to leverage the cloud, many felt that standard databases were too inflexible to adapt to the cloud-scale setting.
Objetivo
Cloud computing is widely recognized as the most promising computing paradigm of the last several years. Cloud providers such as Amazon, Google, and Microsoft have promoted the emergence of large-scale data centers with thousands of machines, offering economies-of-scale, since multiple applications and businesses share a common infrastructure. They offer the following cloud platforms: Amazon’s AWS, Google’s AppEngine, and Microsoft Azure. All these providers have added scalable data management products to their offerings. An important issue in cloud data management is the need to provide scalability to both transactional and analytical loads. Cloud computing lets system administrators adjust resources in an on-demand fashion. However, as application developers began to leverage the cloud, many felt that standard databases were too inflexible to adapt to the cloud-scale setting.
The goal of this work is to propose and test solutions to adapt the system to guarantee scalability for Big Data in the Cloud. It concerns storage, retrieval and analytics over Bigdata on the cloud. Self-adaptation techniques can be used in this context to provide architecture-based adaptation to guarantee system qualities such as performance, accuracy, or cost. This will imply the implementation of techniques as adaptation tactics and strategies in Rainbow (a framework for architecture-based self-adaptation developed at CMU) that can be applied to scale the system up and down, among other things.
Plano de Trabalhos - Semestre 1
This work includes the following activities:
(a) [2014-09-01 to 2014-10-31] The review of the state-of-the-art in Big Data, Cloud and Architecture-based self-Adaptation;
(b) [2014-11-01 to 2015-01-31] Use case development;
(c) [2014-12-01 to 2015-01-31] Write Thesis Proposal;
Plano de Trabalhos - Semestre 2
(d) [2015-02-01 to 2015-04-30] Proposal of self-adaptability mechanisms for cloud data management;
(e) [2014-11-01 to 2015-04-30] Prototype implementation and experimental validation of the approaches;
(f) [2015-04-01 to 2015-05-31] Write a paper;
(g) [2015-03-01 to 2015-07-31] Write the thesis.
Condições
The work is to be executed at the laboratories of the CISUC’s Software and Systems Engineering Group and eventually at Institute of Software Research-Carnegie Mellon University (CMU). A work place will be provided as well as the required computational resources. The student will apply for a CMU-Portugal international internship of 1350€/month during 8-12 weeks: (http://www.cmuportugal.org/tiercontent.aspx?id=5204).
Observações
CMU Co-advisers:
- Bradley Schmerl (schmerl+@cs.cmu.edu) and Javier Camara Moreno (jcmoreno@cs.cmu.edu)
DEI Co-advisers:
- Jorge Bernardino (jorge@isec.pt) and Pedro Furtado (pnf@dei.uc.pt)
Orientador
Jorge Bernardino, Pedro Furtado
jorge@isec.pt 📩