In the information society in recent years, the volume of data requested to process increases explosively. Information required from users should be extracted from it instantly. In order to process huge volumes of data, we have constructed a PC cluster consolidated with IP-SAN that integrates the front-end and the back-end networks into the same IP network. However, the detailed analysis is not yet performed how the communication between nodes of the cluster and the execution of I/O influence the behavior of the system performance, when data-intensive applications are executed on it. Thus, in this paper, a PC cluster consolidated with IPSAN is evaluated when the network and target of the cluster are burdened with a heavy load, by accessing a single target with multiple initiators. As an example of data-intensive applications, parallel association rule mining is executed. As the system is monitored and analized, detailed behavior of the PC cluster consolidated with IP-SAN is clarified.
[1]
K. Gopinath,et al.
Performance Evaluation of Multiple TCP connections in iSCSI
,
2007,
24th IEEE Conference on Mass Storage Systems and Technologies (MSST 2007).
[2]
Angelos Bilas,et al.
Performance evaluation of commodity iSCSI-based storage systems
,
2005,
22nd IEEE / 13th NASA Goddard Conference on Mass Storage Systems and Technologies (MSST'05).
[3]
Masaru Kitsuregawa,et al.
Tree Structure Based Parallel Frequent Pattern Mining on PC Cluster
,
2003,
DEXA.
[4]
Jian Pei,et al.
Mining frequent patterns without candidate generation
,
2000,
SIGMOD '00.
[5]
Abhijeet Joglekar,et al.
A scalable and high performance software iSCSI implementation
,
2005,
FAST'05.