Social systems utilizing IT (Information Technology) have been deployed in various situations of our lives. For example, those situations are ticket booking at convenience stores, cash dispensing on ATMs, e-mailing on mobile phones, photo mailing, internet connection, parcel delivery confirmation on the Internet and so on. The 24-hour and 365-day running service that can handle explosive concentrated access is required for these social systems.
IT has permeated all over our lives, storage data volume increases and more management efforts are likely required. In addition, immediate recovery is required if any failure should occur because important business data is stored in the storage and therefore continuous data access is mandatory. These requirements are exemplified in questionnaire responses as shown in the following chart that many storage investors attach importance to reliability of the administrative and system as criterion for their investment.
Under these circumstances, network storages, such as a SAN and a NAS that give storages independence from server nodes and consolidate data distributed on several server nodes have become to be deployed more and more, as an alternate to a direct attached storage (DAS). With DASs, storage management is required on each server node, and as a result, unused areas are often found on many storage devices. By contrast, consolidating storage devices enables you to reduce administrative costs through centralized management and promotes more effective use of storage capacity. For this reason, the network storage such as a SAN should be a propulsion method that optimizes an entire system from the aspect of storages.
However, as storage devices and servers increase, and the SAN configuration becomes more complex, higher level know-how for designing optimal configuration concerning reliability and performance is required. Also it is more difficult to pinpoint a failed part and to figure out which data in which operation is affected by that error. This implies that the correlation between storages and applications can now hardly be managed manually because of the complex coupling structure of server nodes and storage devices. In other words, a traditional management method is almost impossible to be extended in order to maintain a certain service level for the system.
