WORKSHOP PROGRAM: (PDF workshop program)

May 20: A6-204 room – 10pm-12pm

• João Salada: Midlewares for Data Management aiming Energy Efficiency.
• Lluís Barguñó: Google Pregel and other massive graph distributed systems.
• Zhenhua Xu: Cloud for Online Games.
• Vasia Kalavri: A study on Cloud Operating Systems.

May 24: A6-204 room – 8pm-10pm

• Ying Liu: The Internet of Things: Now and Future.
• Erisa Dervishi: Cassandra – A decentralized structured storage system.
• Mariano Vallés: An overview of major private cloud systems in the market.
• Sergio López: Green Databases.

May 27: A6-204 room – 10pm-12pm

• Jawad Manzoor: Privacy and Security in Cloud Computing.
• Sari Setianingsih: Software Licensing in Cloud Computing.
• Nuno Lucas: Cloud Computing in the Smart Grid Domain.

May 31: A6-204 room – 8pm-10pm

• Nicholas Rutherford: Falling skies and stormy teacups: what happens when the Cloud fails?
• Kiarash Rezahanjani: Utility-Oriented Federation of Clouds
• Jordi Balasch: Hybrid Solutions for Analytical Data Management in the Cloud.

Sponsored by:

EEDC Seminars 2011 edition is focusing on Cloud Computing Research arena. The speakers will present current research advances in this research field. You can donwload the slides and contact the speakers.

(link with detailed PDF program)

1 Green Networking Energy-oriented Infrastructures in a Carbon Constrained World
Speaker: Sergio Ricciardi (UPC Barcelona Tech) (slides)
Date: Fri, 25/Feb/2011

2 Optimizing Datacenter Placement and Load Distribution in Internet Services
Speaker: Ricardo Bianchini (Rutgers University) (slides)
Date: Fri, 18/Mar/2011

3 High Availability in Distributed Systems: Issues and Motivations
Speaker: Javier Alonso (UPC Barcelona Tech) (slides)
Date: Fri, 25/Mar/2011

4 Virtualization, the pillar of the Cloud
Speaker: Josep Subirats (Barcelona Supercomputing Center – Centro Nacional de Supercomputación) (slides)
Date: Fri, 1/Apr/2011

5 Large-scale Data Handling and Processing
Speaker: Jordà Polo (Barcelona Supercomputing Center – Centro Nacional de Supercomputación) (slides)
Date: Fri, 8/Apr/2011

6 Energy Savings through the Use of Distributed Storage Systems
Speaker: Valéria Quadros (University of São Paulo, Brazil) (slides)
Date: Fri, 29/Apr/2011

7 Network strategies for clouds
Speakers: Jordi Guijarro & Isabel Gandia (CESCA – Centre de Supercomputació de Catalunya) (slides 1) (slides 2)
Date: Tue 3/Mai/2011

8 Software Aging & Software Rejuvenation
Speaker: Javier Alonso (UPC Barcelona Tech) (slides)
Date: Fri, 10/Mai/2011

9 Self-Adaption and Learning on DataCenters and Cloud Systems
Speaker: Josep Ll. Berral (UPC Barcelona Tech) (slides)
Date: Tue, 13/Mai/2011

EEDC Seminars 2011 edition is focusing on Cloud Computing Research arena. The speakers will present current research advances in this research field. You can donwload the slides and contact the speakers.

(link with detailed PDF program)

1 Green Networking Energy-oriented Infrastructures in a Carbon Constrained World
Speaker: Sergio Ricciardi (UPC Barcelona Tech) (slides)
Date: Fri, 25/Feb/2011

2 Optimizing Datacenter Placement and Load Distribution in Internet Services
Speaker: Ricardo Bianchini (Rutgers University) (slides)
Date: Fri, 18/Mar/2011

3 High Availability in Distributed Systems: Issues and Motivations
Speaker: Javier Alonso (UPC Barcelona Tech) (slides)
Date: Fri, 25/Mar/2011

4 Virtualization, the pillar of the Cloud
Speaker: Josep Subirats (Barcelona Supercomputing Center – Centro Nacional de Supercomputación) (slides)
Date: Fri, 1/Apr/2011

5 Large-scale Data Handling and Processing
Speaker: Jordà Polo (Barcelona Supercomputing Center – Centro Nacional de Supercomputación) (slides)
Date: Fri, 8/Apr/2011

6 Energy Savings through the Use of Distributed Storage Systems
Speaker: Valéria Quadros (University of São Paulo, Brazil) (slides)
Date: Fri, 29/Apr/2011

7 Network strategies for clouds
Speakers: Jordi Guijarro & Isabel Gandia (CESCA – Centre de Supercomputació de Catalunya) (slides 1) (slides 2)
Date: Tue 3/Mai/2011

8 Software Aging & Software Rejuvenation
Speaker: Javier Alonso (UPC Barcelona Tech) (slides)
Date: Fri, 10/Mai/2011

9 Self-Adaption and Learning on DataCenters and Cloud Systems
Speaker: Josep Ll. Berral (UPC Barcelona Tech) (slides)
Date: Tue, 13/Mai/2011

Abstract:

We are facing an ever-increasing growth in Digital Data. Studies show that by 2020 Digial Universe is expected to increase 40-35 trillion gigabytes, a rate much faster than Disk Drivers performance improvements. Thus to deal with that problem we need to build Efficient and Scalable Storage Systems. In this work we explored middlewares which aim to create or improve the Efficiency and Performance of those systems.

Presentation

Biography:

João Salada received a B.Sc. degree in Communications Networks Engineering by Instituto Superior Técnico, Portugal. He is taking the M.Sc. at the same University,  with major area in Networks Programming. Its area of interest goes from Linux Kernel Network developing and Cloud Computing. During this semester he was studying as an Eramus student at UPC-FIB.

Abstract

In April 2011 the popular Amazon AWS Cloud platform experienced the largest service disruption of its 9 year history.
This presentation aims to give an overview of what happened, who was affected, and which lessons can be learned from those involved to be prepared for similar events in future.

Slides: Falling skies and stormy teacups

Biography

Nicholas Rutherford received a first class BSc Hons. degree in Computer Science from the University of Sheffield, UK, in 2009.

He is currently an Erasmus Mundus Masters scholar in Distributed Computing at the Technical University of Catalonia (UPC), Spain, and KTH Royal Institute of Technology, Sweden.
During and between studies he has worked as a small-medium business web application software developer, targeting conventional and cloud hosting platforms.

Contact Nick

Abstract

We will analyse some of the computing needs for building the Smart Grid, and examine the current Cloud Computing infrastructure to see whether it can address these needs. Some examples, like pilot programs and tools, related to this topic could also be given. Analysis of the market and it future opportunities will also have place.

Biography

Nuno Lucas received the degree in Informatic Engineering from University of Coimbra, Portugal, in 2004. From 2005 until 2010, he worked as software engineer in consulting sector, on sectors as Government, Telecom and Energy. He is currently taking the MSc degree in Computer Architecture, Networks and Systems from the Technical University of Catalonia (UPC), Spain.

Contact

nuno.filipe.a.m.falcao@est.fib.upc.edu

Slides: here

Abstract

This presentation is about Cloud Operating Systems. I present the requirements, challenges, design issues and trends of this area. The presentation covers both operating systems based on modern browsers and operating systems designed to manage resources of large datacenters. As an example of the first case, I present in more detail Microsoft’s Gazelle, a browser that focuses on security and is designed as a multi-principal OS that manages resource protection and sharing across web site principals. As an example of the second case, I focus on VMware’s vSphere, specifically designed to manage large collections of infrastructure as a seamless, flexible and dynamic operating environment.

Bio

Vasia Kalavri received her engineering degree on Electrical and Computer Engineering from the National Technical University of Athens, in the summer of 2010. Since last September she is pursuing the European Master in Distributed Computing, at UPC, Barcelona. Apart from studies, she has worked as a web developer at E-on Integration S.A., located in Athens, where she was developing Banking, ERP and CRM applications.

Contact

vasia@ac.upc.edu

Here you can find the Presentation Slides.

Title

Hybrid Solutions for Analytical Data Management in the Cloud

Presentation Slides

Abstract

As the data volume that needs to be analyzed is exploding, many companies are moving from their propietary machines to cheaper alternatives inside clouds. Basically, there are two technologies to perform data analysis in this environment. On the one hand, we have the typical Parallel Databases, with years of experience and research focused on improving performance. On the other hand, MapReduce systems are emerging as a good alternative for data analysis due to their major scalability and fault tolerance.

However, none of the mentioned approaches meets all the requirements expected from a large-scale data analysis system, such as efficiency, fault tolerance and ability to run in heterogeneous environments and to support other commercial tools. In this presentation, we explore some hybrid solutions which try to take the best features from both approaches in order to build a solid system to perform data analysis in the cloud.

Bio

Jordi Balasch received his Bachelor Degree in Computer Science from Universitat Politècnica de Catalunya (UPC) in June 2010. In September 2008 he joined the Data Management group (DAMA-UPC) at the Computer Architecture Department in UPC, first as a collaborating student and from March 2010 as a research associate. Also in March 2010 he was granted an IBM-CAS fellowship and spent four months in the IBM Toronto Lab from June 2010 to October 2010. Currently he is enrolled in the Computer Architecture, Networks and Systems Master in UPC.

Contact

jbalasch@ac.upc.edu

Abstract

The booming of the Cloud industry triggers new opportunities for online games. This work explores the advantages brought by Cloud for online game providers, as well as the problems that could hamper moving games into Cloud.

Traditionally, online game providers try to provision for resources allocation. Despite the high expense of the infrastructure, game providers have to over provision to cater the possible peaks of customer needs. The problem with over provisioning is that cost is high and the granularity is quite coarse.

With Cloud, resource allocation is dynamic and quite flexible. Thus Cloud is an appealing choice for online game providers. However, virtualization in Cloud could causes latency overhead which might not be desirable. Also, a common phenomena called “hot spot” problem, which happens when players gather intensely in one region of the game world, might need more programming efforts rather than Cloud technology.

Nevertheless, there are pioneering efforts in the video game industry. OnLive and Gaika are moving video games completely into server side. Using a video streaming technology, they enable customers to play anywhere with a low-end client machines. The key concept in their business model is consolidation of computing resources. Customers need no console at home anymore.

You can find the slides Here.

Biography:

Zhenhua Xu is currently an EMDC(European Master in Distributed Computing) scholarship holder. He is studying in UPC (Universitat Politècnica de Catalunya), Barcelona now and will start his third semester in KTH(Royal Institute of Technology), Sweden. He received his Master’s degree from School of Software Engineering  of Tsinghua University, China, in July2010. Before that he received his Bachelor’s degree from School of Mechanical Engineering of Shanghai Jiao Tong University, China, in July 2007.

His current interests are Distributed Computing, Cloud technology and Online Games.

Email : gonadarush@gmail.com

By Erisa Dervishi – EMDC

The presentation: Cassandra_pptx_slides

Abstract

Cassandra is a distributed database which was developed by Facebook as a method for inbox searches. It was authored by Avinash Lakshman, who had previously worked on Amazon’s Dynamo database, and Prashant Malik in 2007, and written in Java. In 2008 it became an open source project and was picked up in 2009 by Apache, who made it an incubator project. Cassandra was created to be a quick, scalable and fault tolerant system and as such it has a number of important features which distinguish it from other competitors.

This presentation outlines these key features, particularly in the areas of architecture, scalability and data model. Finally a general view of the actual market is given, together with some facts about the current situation and what the future will reserve for Cassandra.

Biography:

Erisa Dervishi was graduated in Computer Science in the Faculty of Natural Sciences (University of Tirana, Albania) in June 2008. From June 2008 until July 2010 she has worked as an IT consultant and developer for Intech+ sh.p.k., an Oracle oriented, consulting and developing company in Tirana. Since September 2010 she is a master student of the European Master in Distributed Computing, and currently she is finishing her first year of studies in the Technical University of Catalonia(UPC).

Email : erisa85d@gmail.com