Monika Moser

Zuse-Institut Berlin, Berlín, Berlin, Germany

Are you Monika Moser?

Claim your profile

Publications (5)0 Total impact

  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Data consistency can be violated in Distributed Hash Tables (DHTs) due to inconsistent lookups. In this paper, we identify the events leading to inconsistent lookups and inconsistent responsibilities for a key. We find the inaccuracy of failure detectors as the main reason for inconsistencies. By simulations with inaccurate failure detectors, we study the probability of reaching a system configuration which may lead to inconsistent data. We analyze majority-based algorithms for operations on replicated data. To ensure that concurrent operations do not violate consistency, they have to use non-disjoint sets of replicas. We analytically derive the probability of concurrent operations including disjoint replica sets. By combining the simulation and analytical results, we show that the probability for a violation of data consistency is negligibly low for majority-based algorithms in DHTs.
    07/2008: pages 249-260;
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Structured Overlay Networks provide a promising platform for high performance applications since they are scalable, fault-tolerant and self-managing. Structured overlays pro-vide lookup services that map keys to nodes that can be used as processing or storage resources. The lookups for a key may return inconsistent results. Consequently, it is nontrivial to provide consistent data services on the top of structured overlays that are built on key-based search. In this paper, we study the frequency of occurrence of incon-sistent lookups. We show that the effect of lookup incon-sistencies can be reduced by assigning responsibility of key intervals to nodes. We present our results as a trade-off between consistency and availability of keys. Further, since many distributed applications employ quorum techniques at their core, we analyze the probability that majority-based quorum techniques will function correctly in a structured overlay with inconsistent lookups. Our analysis shows that the probability of majority-based algorithms to function cor-rectly despite lookup inconsistencies is high.
    Proceedings of the 17th International Symposium on High-Performance Distributed Computing (HPDC-17 2008), 23-27 June 2008, Boston, MA, USA; 01/2008
  • Source
    Monika Moser, Seif Haridi
    [Show abstract] [Hide abstract]
    ABSTRACT: We investigate the problem of atomic commit in transactional database systems built on top of Distributed Hash Tables. Therefore we present a framework for DHTs to provide strong data consistency and transactions on data stored in a decentralized way. To solve the atomic commit problem within distributed transactions, we propose to use an adaption of Paxos commit as a non-blocking algorithm. We exploit the symmetric replication technique existing in the DKS DHT to determine which nodes are necessary to execute the commit algorithm. By doing so, we achieve a lower number of communication rounds in contrast to applying traditional Three-Phase-Commit protocols. We also show how the proposed solution can cope with dynamism due to churn in DHTs. Our solution works correctly relying only on an inaccurate failure detection of node failure, what is necessary for systems running over the Internet.
    Towards Next Generation Grids, Proceedings of the CoreGRID Symposium 2007, August 27-28, Rennes, France; 01/2007
  • Source
    Managing Virtualization of Networks and Services, 18th IFIP/IEEE International Workshop on Distributed Systems: Operations and Management, DSOM 2007, San José, CA, USA, October 29-31, 2007, Proceedings; 01/2007
  • Source
    Monika Moser, Seif Haridi
    [Show abstract] [Hide abstract]
    ABSTRACT: We investigate the problem of atomic commit in transactional database systems built on top of Distributed Hash Tables. DHTs provide a decentralized way to store and look up data. To solve the atomic commit problem we propose to use an adaption of Paxos commit as a non-blocking algorithm. We exploit the symmetric replication technique existing in the DKS DHT to determine which nodes are necessary to execute the commit algorithm. By doing so we achieve a lower number of communication rounds and a reduction of meta-data in contrast to traditional Three-Phase-Commit protocols. We also show how the proposed solution can cope with dynamism due to churn in DHTs. Our solution works correctly relying only on an inaccurate failure detection of node failure which is necessary for systems running over the Internet.