We have presented the idea of enabling the cloud storage to support traditional applications. In this way, the applications running in local host can be deployed and running on the cloud storage without any modification. An acceptable solution is to enhance current cloud storage with standard POSIX file system interface. So that it can benefit from the scalable capacity of existing cloud storage and compatible interface of traditional file systems. With this point in mind, we designed and implemented the PosixCloud, a general purpose scalable storage cloud with standard POSIX interface for traditional end users. Similar to other cloud storage systems, the PosixCloud is designed to be chunk based data layout, built on commodity computers, chunk replication for reliability and so on. However, the metadata in our system is managed in distributed database instead of the memory single master. So that the metadata scalability of our system is enhanced and large small files can be handled without introducing high overhead. The paper describes the details of the design and implementation of PosixCloud. Also, we evaluate the POSIX compliance, functionality, data and metadata performance of PosixCloud by both test suits of third parties and customized experiments. The results demonstrate that our system can achieve reasonable performance and meet the storage requirement of most traditional applications.
[1]
Jacob R. Lorch,et al.
A five-year study of file-system metadata
,
2007,
TOS.
[2]
Howard Gobioff,et al.
The Google file system
,
2003,
SOSP '03.
[3]
Michael Vrable,et al.
Cumulus: Filesystem backup to the cloud
,
2009,
TOS.
[4]
Wilson C. Hsieh,et al.
Bigtable: A Distributed Storage System for Structured Data
,
2006,
TOCS.
[5]
Randy H. Katz,et al.
Above the Clouds: A Berkeley View of Cloud Computing
,
2009
.
[6]
Ilango Sriram,et al.
SPECI, a Simulation Tool Exploring Cloud-Scale Data Centres
,
2009,
CloudCom.
[7]
Andrei Broder,et al.
Network Applications of Bloom Filters: A Survey
,
2004,
Internet Math..
[8]
Robbert van Renesse,et al.
Chain Replication for Supporting High Throughput and Availability
,
2004,
OSDI.
[9]
Werner Vogels,et al.
Dynamo: amazon's highly available key-value store
,
2007,
SOSP.
[10]
Rajkumar Buyya,et al.
Market-oriented Grids and Utility Computing: The State-of-the-art and Future Directions
,
2008,
Journal of Grid Computing.
[11]
GhemawatSanjay,et al.
The Google file system
,
2003
.
[12]
Hans-Arno Jacobsen,et al.
PNUTS: Yahoo!'s hosted data serving platform
,
2008,
Proc. VLDB Endow..
[13]
Chandra Krintz,et al.
Key-Value Datastores Comparison in AppScale
,
2010
.