AVATAR-m Project Details
AVATAR-m is a research and development project that has brought together experts in archiving, broadcast technology, data storage, algorithms and modelling, to develop a family of technologies that can dramatically improve the utility of digital file storage for A/V archiving.
Planning
From the outset the project has understood the critical importance for archives of being able to plan their archival storage needs. AVATAR-m has built storage planning tools that for the first time give detailed interactive plots of the long term needs of archives to those who need it most- archival managers. Our tools allow hypothetical and realistic models of storage use to be developed, incorporating migration, redundancy and technologies from across the storage spectrum. We understand the daunting challenge of planning investment over years or decades, and our tools allow archive managers to make those decisions with the best possible information to hand.
Management
Controlling the day to day configuration of a large, heterogenous storage facility is a full time job, and we understand that the volumes of storage required to maintain audio visual assets in sufficient quality are challenging to say the least. We have developed software tools that can automatically undertake much of that management based upon clear, simple, archiving business rules, managed and accessed through highly effective graphical user interfaces. Our technology is platform agnostic, and has been shown to work across a wide range of storage devices. In simulation we have been able to explore the integration of technologies considered exotic in today’s archives, but potentially a key part of future infrastructures, and in live prototypes we’ve been able to demonstrate the integration of today’s generation of cloud storage services.
Platforms
Our engineering development has produced key storage technology components that will allow critical data manipulation deep within archive disk storage, bringing a key advantage to the management of assets across distributed environments.
Formats
To date all A/V codecs have had, at their inception, the need to transmit signals over some form of broadcast or point to point media, either analogue or digital. We wanted to get back to the drawing board, and see if we can build a codec specifically for archival usage. Using the DIRAC codec we have been exploring the challenges of signal-channel coding for long term, low cost, managed risk storage. We want to determine whether archival costs can be pushed down by making the very encoding of assets robust and able to cope with the special environment of the archive. With an accepted latency on delivery from deep store, and a need to have graceful recovery from damaged media, we plan to share the open source findings of our work at IBC 2009 in Amsterdam.