Program

Monday October 6, 2014

  • 13:40 - 13:50 Welcome and opening, João Leitão and Ricardo Vilaça
  • 13:50 - 14:40 Keynote - Geo-replicated storage with scalable deferred update replication, Fernando Pedone, University of Lugano (USI), Switzerland
  • 14:40 - 15:30 Session 1
    "Declared Causality in Wide-Area Replicated Storage". Kyle Lady, Minkyong Kim and Brian Noble
    "The Case for Fast and Invariant-Preserving Geo-Replication". Valter Balegas, Nuno Preguica, Sérgio Duarte, Carla Ferreira, Rodrigo Rodrigues, Mahsa Najafzadeh and Marc Shapiro
  • 15:30 - 15:50 Break
  • 15:50 - 16:40 Keynote - SwiftCloud: Fault-Tolerant Geo-Replication Integrated all the Way to the Client Machine, Nuno Preguiça, NOVA-LINCS & NOVA University of Lisbon, Portugal
  • 16:40 - 17:30 Session 2
    "Mitigating the Spread of a Virus in the Internet". Thanh Dang Nguyen, François Bonnet and Xavier Défago
    "Latency Amplification: Characterizing the Impact of Web Page Content on Load Times". Cătălin-Alexandru Avram, Kenneth Salem and Bernard Wong

Note: Each keynote has 50 minutes. Each full paper has 25 minutes including questions and answers.

Keynotes

Geo-replicated storage with scalable deferred update replication

Fernando Pedone, University of Lugano (USI), Switzerland


Abstract: Many current online services are deployed over geographically distributed sites (i.e., datacenters). Such services call for geo-replicated storage, that is, storage distributed and replicated among many sites. Geographical distribution and replication can improve locality and availability of a service. Locality is achieved by moving data closer to the users. High availability is attained by replicating data in multiple servers and sites. In this talk I will consider a class of highly available storage systems based on deferred update replication. Deferred update replication ensures strong consistency and is at the core of several highly available storage systems. I will first show how the performance of deferred update replication can be made to scale to dozens and potentially hundreds of nodes. I will then consider the implications of the approach on user-perceived latency in geographically distributed environments and propose solutions. The talk will conclude with some open problems in the design of strongly consistent storage systems.

Bio: Fernando Pedone received his Ph.D. degree in computer science from Ecole Polytechnique Fédérale de Lausanne (EPFL), Switzerland, in 1999. Before becoming Assistant and Associate Professor at the University of Lugano, he worked as a researcher at the Hewlett-Packard Laboratories in PaloAlto, California, and as a senior researcher at EPFL. His professional interestsinclude the theory and practice of distributed systems and distributed datamanagement systems. His theoretical work has addressed algorithmic aspects of reliable distributed systems, such as defining new abstractions for buildingreliable systems, creating algorithms implementing these abstractions and proving them correct. His systems work has considered using group communication abstractions to build replicated data management systems. To a large extent, he has been interested in the inter play between theory and practice. Fernando Pedone has authored more than 70 scientific papers. He is co-editor of the book Replication: theory and practice.

SwiftCloud: Fault-Tolerant Geo-Replication Integrated all the Way to the Client Machine

Nuno Preguiça, NOVA-LINCS & NOVA University of Lisbon, Portugal


Abstract: Client-side logic and storage are increasingly used in web and mobile applications to improve response time and availability. Current approaches tend to be ad-hoc and poorly integrated with the server-side logic. We present a principled approach to integrate client- and server-side storage. We support both mergeable and strongly consistent transactions that target either client or server replicas and provide access to causally-consistent snapshots efficiently. In the presence of infrastructure faults, a client-assisted failover solution allows client execution to resume immediately and seamlessly access consistent snapshots without waiting. We implement this approach in SwiftCloud, the first transactional system to bring geo-replication all the way to the client machine. Example applications show that our programming model is useful across a range of application areas. Our experimental evaluation shows that SwiftCloud provides better fault tolerance and at the same time can improve both latency and throughput by up to an order of magnitude, compared to classical geo-replication techniques.

Bio: