SRB General FAQs

General questions about Nirvana and SRB are addressed here. For detailed answers to specific questions, please contact us. Also see Technical FAQs for additional information.

What is SRB and what does it do?
What makes SRB unique?
What is an SRB Federation?
What is the SRB architecture?
What is the history of SRB?
What are some typical implementations?
What type of data and/or application is right for SRB?
What functionality is gained with SRB?
How easy is SRB to use?
What happens to data brought into an SRB Federation?
How is data stored and accessed within SRB Vaults?
Does SRB support Information Lifecycle Management (ILM)?
What are SRB Daemons?
How hard is it to integrate legacy data with SRB?
How does SRB scale with the addition of data and users?
Which SRB packaged solution is right for me?
Is SRB secure enough for my organization?
Is SRB available for evaluation?
How does SRB compare to other, similar products?
When should a strategic (grid) approach be considered?
How does SRB differ from similar data storage solutions?

What does SRB do?
Powerful yet easy to implement and use, SRB helps knowledge workers collaborate by sharing and managing high-value data through versatile and user-friendly interfaces. A proven solution with over a decade of development, SRB enables the federation of enterprise data. It is ideal in complex computing environments where groups at multiple sites require immediate access to shared information for discovery, reporting, and data mining.

SRB federates data from disparate and distributed data resources like Storage Area Networks (SAN), Network Attached Storage (NAS), and Content Addressed Storage (CAS). SRB can easily scale up to an enterprise-wide or cross-enterprise collaborative data grid implementation.

Simple and yet powerful SRB interfaces allow end-users to access an enterprise-wide, global view of information from sources including file systems, databases, and archival storage.

Once SRB is implemented – and business value realized, then IT Managers benefit from easier ILM implementation, technology migrations, and site integrations following mergers and acquisitions.

What makes SRB unique?
SRB is a unique package designed for data-intensive enterprises that combines the following features:

SRB is an information management suite that combines all the necessary features in order to perform a variety of data management tasks enterprise-wide.
SRB provides organization capabilities for all data in the enterprise, independent of its data repository.
A Global Namespace spans all enterprise-wide data, making it possible to create taxonomies (Virtual Collections) or cross-references (Data Object or Collection Links) besides the standard hierarchical organization that a file system provides.
Metadata on Data Objects, Collections, and Storage Resources enables powerful search and discovery operations and provides for the complete automation of storage management tasks. Metadata Schemes help to organize metadata into groups.
SRB Daemons automate data management tasks and allow for the management of an entire enterprise through Administrator-defined policies.
Fine-grained and role-based security layer ensures that only authorized users gain access to any of SRB's Data Objects spanning multiple storage systems and sites.
Patented SRB Containers help to reduce access times and latencies and allow for the expansion of file systems beyond their physical limits (typically around 20 Million files).
Tickets allow external users to gain controlled access to internal data on a time and access-count limited basis.

What is an SRB Federation?
A Federation is the combination of all SRB servers, clients and storage systems, implemented as a distributed software layer using client/server architecture.

What is the SRB Architecture?
The SRB Architecture is a three-tiered design:

The Presentation Tier: The SRB Client
The Logic Tier: Metadata Catalog (MCAT)
Data Tier: SRB Agents with drivers to the storage systems

What is the history of SRB?
SRB was developed at the San Diego Supercomputing Center (SDSC) at the University of California, San Diego (UCSD), was first demonstrated in 1995, and has been under continuous development since then. General Atomics founded the SDSC, managed it through 2003, owns commercial rights to SRB, and to this day cooperates with UCSD in continuing SRB development. Today, SRB is the basis for over 150 of the world's most sophisticated data grids. The Nirvana was formed to commercialize SRB, and began in 2000 by launching a commercial version of the product with improved capabilities adapted to government and commercial users, licensing, implementation and support services, maintenance and continuing product development.

What are some typical implementations?
SRB is at the core of some of the world's most complex computing environments. Data grids, media libraries, and digital archives in a wide variety of academic, supercomputing and government environments depend on SRB for transparent access to data.

SRB federates high-value data that is located among heterogeneous and distributed storage systems. SRB allows the virtual organization of data and a simplified presentation view for end-users.

SRB is currently operational at over 150 government agencies, corporations, and research organizations. With scalable capacity to manage increasing volumes of data, SRB is a proven solution for complex and demanding computing environments.

A few notable implementations:

Planetarium at a major university, visualizing the collision of galaxies
Corporation document archival and retrieval for over 50 years of records
Government agency image library expanding more than a terabyte per week
Government agency integrating case and evidence data from multiple locations, and permitting secure and controlled access by field agents

What type of data and/or application is right for SRB?
Geographically dispersed and technologically disparate computing environments benefit from this easily-integrated, non-disruptive solution. SRB provides secure, real-time data access to a federation of heterogeneous and distributed data resources.

SRB supports all classes of data, but it is particularly well suited for unstructured or fixed-content data (e.g., documents, images, and video files). SRB is extensible to recognize new, unknown data types when detected. In addition to recognizing new data types, SRB also supports all languages and characters supported by UNICODE.

Structured documents such as Microsoft Excel spreadsheets or entire databases such as ESRI personal geo databases can be directly stored and managed by SRB.

SRB also provides support to access relational databases and to customize the returned query results into any output format (i.e., XML, HTML, custom templates, etc.).

What functionality is gained with SRB?
Data Federations within SRB can be managed, accessed, organized, and secured using consistent and coherent policies across departments, the enterprise, or even across multiple enterprises. Policies are not dependent on a data source (or repository) but span storage systems and allow for persistent migration of data among systems. Data migration, replication, or archiving are also supported independent of each storage system's capabilities and are transparent to end-users or individual applications.

SRB is ideally suited for classifying data based on business value or Information Lifecycle Management (ILM) policies. With SRB, Administrators can set enterprise-wide policies to manage the different data classes. Data can be archived, migrated, replicated, synchronized, deleted, or backed-up using such policies. The ILM Daemon is specifically designed for this purpose; SRB Daemons are background processes that automate data management and synchronization throughout a Federation.

How easy is SRB to use?
SRB is initially implemented with a simple InstallShield Wizard. As a prerequisite to this, a relational database to house the Metadata Catalog (MCAT) is required. SRB is compatible with most relational databases. Most of the configuration is performed using a graphical interface, from which Administrators manage the entire SRB Federation at a central console. Once this data source configuration is complete, the SRB Federation is automatically populated with data and metadata.

Users and applications going through SRB Gateways will not even notice that they are communicating with the system since SRB Collections can be accessed as Windows network shares or UNIX/Linux directories. This eases adoption and increases collaboration and productivity immediately.

What happens to data brought into an SRB Federation?
Depending on your configuration, data can be left untouched in its current location within a file server or data repository and only be registered into SRB, meaning SRB merely creates a pointer to the existing data. Conversely, data can be ingested, meaning it is physically transferred into an SRB Vault.

Once registered or ingested, data is accessible via a single virtualization point, chosen from a selection of familiar user interfaces. SRB also stores metadata about each file or data object within the Federation and can readily supply that information to users or downstream applications with the right access privilege.

How is data stored and accessed within SRB Vaults?
SRB maintains data under management within SRB Vaults. These are directories in file systems, tables in databases, or objects in object storage systems where data is stored according to a specific format determined by an Administrator. SRB Administrators define the format of data through an abstraction called a Storage Resource.

In SRB, there are three types of Storage Resources:

Physical Resources - i.e., a tape library or a file system
Logical Resources - a logical grouping of several Resources
Cluster Resources - enabling the implementation of clustered file systems or replicated storage systems in SRB.

Unlike Logical and Cluster Resources, there is a one-to-one relationship between a Physical Resource and its underlying storage system (file system, database, archive, etc.). The Administrator determines the contents of Logical and Cluster Resources and also how data will be handled across their attached Resources. To improve data availability as well as system performance, Administrators can setup various policies for Logical Resources to handle data in a round-robin fashion or to replicate it across any number of attached Resources.

Back to top

Does SRB support Information Lifecycle Management (ILM)?
Yes. SRB is ideally suited for classifying all of an organization's data based on its business value. Administrators can then set enterprise-wide policies to manage the different data classes. Data can be archived, migrated, replicated, synchronized, deleted, or backed-up using such policies. The ILM Daemon is specifically designed for this purpose.

What are SRB Daemons?
SRB Daemons are background processes that automate data management and synchronization thoughout the SRB Federation.

How hard is it to integrate legacy data with SRB?
There are three simple steps for legacy data integration with SRB:

Install Metadata Catalog (MCAT) Server with relational database MCAT.
Install SRB Agents on servers that will be tied into the SRB federation.
Register existing data and extraction of metadata after configuration of data sources.

Populating SRB with data and metadata is automated in Step 3, using the Sync Daemon and other automated metadata extraction tools. SRB supports most legacy storage systems.Compatibility with other systems can be custom developed within SRB's rapid development environment.

How does SRB scale with the addition of data and users?
SRB is highly scalable and the data transfer rate is limited only by available bandwidth and the capabilities of the underlying storage system, not by SRB. With clustered high-performance databases such as Oracle on the back end, SRB delivers enough performance for virtually any computing environment.

Due to SRB's modular architecture and parallelization support, Administrations can add storage systems or servers of any kind (i.e., relational databases, file servers, SANs, NAS devices, object storage, CAS, web servers, FTP servers, document management systems, tapes, etc.) as needed.

Which SRB packaged solution is right for me?
SRB is available under a variety of licensing schemes designed for just about any utilization purpose, including evaluation, proof-of-concept, departmental, and full-scale enterprise implementations. For many situations, including limited geospatial and digital video management, our departmental packages are a good fit and are priced reasonably, giving you an affordable entry point to this advanced technology.

For enterprise implementations, SRB is offered as a software server license, with pricing scaled to the total amount of data under MCAT management, plus a license for each node that is to be part of the Federation, and Daemons as features. The objective is that you pay for only what is utilized, in some cases as small as 1TB of data and a few servers.

With this simple structure in place, SRB gives Administrators the ability scale up to very large and complex data federations spanning multiple sites. As more data is brought under SRB management, licenses are upgraded.

For further details, please contact us.

Is SRB secure enough for my organization?
SRB currently employs different authentication mechanisms for its clients and servers: Challenge response mechanism, GSI authentication based on the GSSAPI, and Kerberos authentication. Data can be encrypted and sent over unsecured lines with minimal overhead. The Grid Security Infrastructure (GSI) employs certificate mechanisms also used in Public Key Infrastructure (PKI) where a public and a private key pair are used to encrypt and decrypt information.

Authorization is another service provided by SRB. Access Control Lists (ACLs) are kept for every SRB Object (Data Objects, Collections, Metadata Attributes, Resources, Locations, User, etc.) so that access can be controlled at different levels. SRB Users can not see any SRB Objects that they don't have at least read access to. This allows for different SRB Users (such as 'public' and 'secret') to see very different contents for the same SRB Collection.

Furthermore, audit trails are generated for every transaction performed in SRB. Those audit trails are stored in a relational database for easy querying and filtering. They contain information like the audit time, the user performing the transaction, type of transaction, details about the transaction, etc.

Is SRB available for evaluation?
SRB is available for 30-day evaluation/proof-of-concept trials. Nirvana recommends onsite consulting, installation and configuration support for proof-of-concept implementations. If 30 days is not sufficient to meet your testing needs, SRB Express may be a better option. Please contact us to discuss short-term and long-term evaluation options.

How does SRB compare to other, similar products?
Many of SRB's functions (tasks such as backup, replication, search, or security) can be performed with individually purchased point products from various vendors. SRB seamlessly integrates with these products to help create an over-arching solution to what often begins with tactical product decisions.

Although SRB has overlapping functionality with other products in many deployment scenarios, SRB works in conjunction with them as it ties in heterogeneous systems. To complement existing products already implemented, SRB provides global namespace, metadata management, ILM, and future system compatibility/expandability.

When should a strategic (grid) approach be considered?
Consider a data grid in response to rapidly-increasing amounts of data being created and controlled at different sites. SRB creates and manages a Global Namespace spanning all data enterprise-wide while it facilitates and automates data access, management, organization, and discovery. SRB complements most similar products. For example, SRB preserves investments in existing storage hardware and increases its value by more fully utilizing that storage. The Global Namespace integrates into existing security infrastructure such as Microsoft Active Directory, Kerberos, Grid Security Infrastructure (GSI), and Public Key Infrastructure (PKI).

How does SRB differ from other data storage solutions?
In traditional data storage solutions, a file system (such as shared folders or NFS mounts) organizes data in physical directories. Typically, end users do not have the ability to pull data from several directories together into a logical grouping, or Collection.

In SRB, all data is organized by Collections independent of where the data is physically stored. This creates a much more productive work environment for both Administrators and end users. Since the logical organization of data is disconnected from the physical data storage, Administrators can perform storage and data management tasks without affecting end users.

And end users can logically group all their data without having to worry about its physical location:

Virtual Collection: Dynamic, criteria-based directories (query results) refreshed upon access or in pre-defined intervals.
Custom Organization/View: Same set of data in a Logical Collection, but organized differently by each user.

SRB keeps customized metadata about each Data Object. Other data storage file systems allow only a few pre-defined attribute, limiting search and query capabilities. The SRB philosophy is to keep data where it is produced in Federations. This runs counter to data warehousing concepts, where all data is transferred to a central location. This does, however, not exclude the usage of data warehousing and data aggregation at a single site.

In contrast to file systems, SRB creates and manages a Global Namespace that can span multiple distributed cluster file systems, storage area networks (SANs), network attached storage devices (NAS), object storage devices or content addressed storage (CAS), archival storage, relational databases, document management systems and other types of data repositories.

Back to top