With Nirvana, data can be analyzed and managed no matter where it's located. Backup or migration from a local administrative domain can be done centrally, enterprise-wide with Nirvana via the MCAT and the Global Namespace.
- Data Load: There are two mechanisms to load data into a Nirvana Federation: Registration and ingestion. Registration leaves the data untouched and only registers a pointer to the data in MCAT without transferring any of the data contents. Ingestion is similar to Registration but also transfers the data to a Nirvana Storage Resource.
- Data Retrieval: Data is as easily retrieved from a Nirvana Federation as it is loaded. Where the data is retrieved from is completely transparent because Nirvana and TCP/IP handle the routing of the data streams.
- Replication: Data replication, the process of creating and maintaining synchronized copies of a Data Object throughout the Nirvana Federation, has a number of uses and benefits: Protecting data in a disaster recovery scenario with instantaneous failover from one replica to another; caching by allowing clients to access a local replica rather than a remote one; and load balancing by distributing synchronized replicas across multiple servers.
- Persistent Migration: Using the Global Namespace as a logical view of federation-wide data, Nirvana Clients can migrate data to new Storage Resources without affecting the way users or applications "see" data organized in their Collections.
- Backup: Unlike Replication, a backup is not automatically overwritten when the original data is modified. This protects against accidental or malicious destruction of data. A backup in Nirvana makes copies of Data Objects or entire Collections in designated backup Collections.
- Restore: The restore process recovers deleted or wrongfully modified objects to an earlier point in time.
- Secure Queries: A number of different distributed objects can be queried through the MCAT: Resources, configuration information, Users, Groups, Domains, Locations, Data Objects, Collections, Tickets, Schemes, etc. There are several hundred system-level and unlimited user-level attributes that allow the creation of very complex SQL-style queries. Users, Groups, or entire Domains can be restricted from writing or even seeing certain attributes.