Marcelo Madruga’s research while affiliated with Pontifícia Universidade Católica de Goiás and other places

What is this page?


This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.

Publications (4)


Using transparent files in a fault tolerant distributed file system
  • Conference Paper

April 2009

·

59 Reads

·

1 Citation

Marcelo Madruga

·

Sergio Loest

·

The peer-to-peer model and the bandwidth availability are fostering the creation of new distributed file systems. However, files belonging to local application and distributed applications are usually handled in the same way by the local file system, so both contribute equally to consume storage space. This paper presents a peer-to-peer distributed file system which uses the ldquotransparent filerdquo concept to improve its fault tolerance and file availability. Files are kept as transparent/volatile replicas, using the free space available in each local file system. When a replica is invalidated, peers cooperate to restore it. The proposed architecture was implemented and tested; experiments showed its feasibility, and that its costs are proportional to the size of files being replicated. The occurrence of multiple simultaneous replica invalidations did not impose a significant overhead.


BackupIT: An Intrusion-Tolerant Cooperative Backup System

January 2009

·

50 Reads

·

3 Citations

Abstract Reliable storage of large amounts,of data is always a delicate issue. Availability, efficiency, data integrity, and confidentiality are some,features a data backup,system should provide. At the same time, corporate computers offer spare disk space and unused networking,resources. In this paper, we propose an intrusion-tolerant cooperative backup system that provides a reliable collaborative backup resource by leveraging these independent, distributed re- sources. This system makes,efficient use of network and storage resources through compression, encryption, and ef- ficient verification processes. It also implements a protocol to tolerate Byzantine behaviors, when nodes arbitrarily de- viate from their specifications. Experiments performed,to evaluate the proposal showed,its viability.


Figura 2. Tempo de recuperação de uma réplica invalidada em um nó com espaço livre em disco e em um nó sem espaço livre em disco.
Um Sistema de Arquivos Distribuído Peer-to-Peer Tolerante a Faltas usando Arquivos Transparentes
  • Article
  • Full-text available

January 2002

·

74 Reads

The peer-to-peer computing model motivates the creation of new dis- tributed file systems. This paper presents a peer-to-peer distributed file system which uses "transparent files" to improve its fault tolerance and file availa- bility, without imposing much impact on the local peers' resources. Files are kept as transparent replicas, using the free disk space in each peer. However, a transparent file may be invalidated if the local filesystem needs the space it uses. When such a replica is invalidated, the peers should cooperate to restore it back. The proposed architecture was implemented and tested; experiments showed its feasibility and that its costs are proportional to the size of files being replicated. Also, the occurrence of multiple simultaneous replica invalidations do not impose a heavy burden on the system. Resumo. O modelo peer-to-peer e a disponibilidade de largura de banda têm viabilizado a criação de novos sistemas de arquivos distribuídos. Este artigo apresenta um sistema de arquivos distribuído peer-to-peer que usa "arquivos transparentes" para melhorar a tolerância a faltas e a disponibilidade dos da- dos, sem impor um impacto significativo nos recursos locais de cada nó. Arqui- vos são salvos como réplicas transparentes no espaço livre em disco de cada nó. Entretanto, um arquivo transparente pode ser invalidado, se o sistema de arquivos local precisar do espaço ocupado por ele. Quando uma réplica é in- validada, os nós cooperam entre si para restaurá-la. Experimentos mostraram a aplicabilidade da proposta, e que seu custo é proporcional ao tamanho dos arquivos replicados. Além disso, invalidações múltiplas e simultâneas não cau- sam impacto significativo no sistema.

Download

Citations (1)


... The definition of peers with specific functionality in the network differs from some proposals for P2P systems, where which each peer should be able to play all the roles, thus promoting the idea of a DHT (Distributed Hash Table) [13,11]. However, the implementation using a DHT in its essence is quite costly and difficult to scale [12]. ...

Reference:

USTO.RE: A private cloud storage software system
BackupIT: An Intrusion-Tolerant Cooperative Backup System
  • Citing Conference Paper
  • January 2009