Bycast StorageGRID 8
May 1, 2009 12:00 PM, By Annette Saliken and David Slik
The solution provides long-term storage and preservation of digital assets.
Figure 1. Siloed storage
The transformation to file-based digital workflows in broadcasting has resulted in large, growing volumes of files that need to be stored and protected for long periods of time. These files are created and processed by various applications in the workflow chain, and it is common for organizations to store the files in siloed storage infrastructures, whereby a separate storage system is used for each application or geographic location. (See Figure 1.)
Silos require additional investments for backup and replication products to protect against data loss. They also need tiered storage to enable movement of data to the most cost-effective storage devices over time. This siloed approach creates significant complexity, and increases the capital and operating costs of storage systems. Furthermore, using silos to store large amounts of data for long periods of time results in the following issues:
- Insufficient data protection
The solution
Traditional backup/restore is not suitable for large data volumes where the time-to-backup becomes prohibitively long. In the event of disaster, restoration can take weeks, months or even years. In addition, backup does not protect against alteration or corruption.
- Downtime and unavailable data
Figure 2. Virtualized storage
Given long retention times, faults will occur over the lifetime of the data. Silos are prone to single points of failure and require manual intervention to recover from a fault, increasing downtime that results in lost productivity.
- Inadequate scalability
The technology
The lifetime of files exceeds that of the storage hardware on which they reside. Silos require manual migrations from one hardware generation to the next, limiting an organization's ability to take advantage of dropping hardware costs.
Bycast StorageGRID 8 virtualization software for large-scale digital archives allows organizations to improve operational efficiency and avoid the pitfalls of siloed systems by consolidating and virtualizing storage across sites and applications.
As shown in Figure 2, this solution manages heterogeneous storage devices, ranging from high-performance disk to tape, across multiple locations to create a unified file archive. It insulates applications from changes to the hardware, including equipment failure and obsolescence. The solution also facilitates the creation of fault-tolerant systems that can scale to tens of petabytes and hundreds of sites. Stored data are managed in accordance with user-configurable Information Lifecycle Management (ILM) policies that govern the automatic placement of content at geographic locations and storage tiers to meet the business needs of the organization. The solution simplifies the management of massive storage systems through automation, while ensuring the availability of data even during system downtime.
StorageGRID is an object-based grid storage system that stores unstructured data such as images, audio, video and documents. An object is any container of data and may have metadata associated with it. In the most common use case, an object is a file, and its metadata are the attributes of that file (type, name, path, etc).
Within this solution, objects are uniquely identifiable and stored in accordance with storage policies based on their metadata. These policies determine how many copies of the objects are stored, geographic placement of objects and storage tiers on which objects reside over time. Applications store and retrieve data from the grid via gateways that provide file or object level access over standard interfaces such as CIFS, NFS and HTTP. The solution can be deployed within one site or across multiple facilities, and objects stored at one site may be accessed at all sites.
Stored files are assigned unique digital fingerprints, which are used to verify integrity of data at the time of access, replication and movement between storage devices. Proactive checks performed in the background protect the integrity of inactive content. When a file is stored, the solution examines its metadata and enforces applicable storage policies. This allows dissimilar file types to be managed differently. For example, intermediate files associated with one project can be stored directly to tape, whereas models used in 3-D renderings may be kept on disk and replicated to a second site where a render farm resides.
| Want to use this article? Click here for options! |  |
