Searching in Cloud Object Storage by Using a Metadata Model

Abstract

In a Cloud environment, an independent service or model referred to as object storage is used for the management of persistent data. This model provides a virtualized pool of storage resources and can be accessed via Application Programming Interface (API) and/or simply a browser. Object storage is used to store the persistent data from various users through put/get methods and it also store the virtual machine images as objects. Cloud storage services are used by enterprises, home users and researchers to store their files that vary in size and formats. With the huge amount of stored data which is mostly unstructured, comes the related issue of finding the exact data relevant to the requirements of a user. In current offerings such as Amazon, Eucalyptus and Google, the contents or objects are defined by a key-value pair. These objects can be listed based on the key, but searching and finding contents with any other parameters (metadata) such as type, size, owner, time stamps etc., is not supported. In this paper we propose a metadata model that address the issue of collecting, adding, and storing metadata associated with objects in Eucalyptus Cloud. Furthermore, the stored metadata is utilized to provide the feature of fast, effective and efficient searching of contents in a Cloud.