Skip to main content

NAB2009 visitor saw real applications of SOA

The move from physical media like videotape to file-based production allows broadcasters to consider changing their working practices. The IT infrastructure underlying file-based production offers the possibilities for a scale of automation just not possible when content is processed as physical media.

Outside the world of the media and entertainment, whole new ways of working have evolved to take advantage of the power of software. Much of the creative side of broadcasting must be left to a human operator, but pushing carts of videotape cassettes down corridors contributes neither to the quality of programming nor to the bottom line.
The media and entertainment sector lags behind other industries in the adoption of IT processes that could deliver efficiencies and cost savings. One area vaunted as the next step forward is the service-oriented architecture (SOA).

To quote from Footen and Faust in their book “The Service-Oriented Media Enterprise” (Focal Press), “SOA is an architecture of independent wrapped services communication via published interfaces over a common middleware layer.” A typical service could be transcoding. The advantages claimed from the use of an SOA include business agility, business visibility and a streamlined organization.

Why now?

Although broadcasters have been moving into file-based workflow for about 10 years, there have been many obstacles to constructing a truly advanced media workflow. Even today it can prove challenging to exchange “standard” formats such as MXF files between vendors. Remember how easy it was to plug analog composite equipment together? It was possible to sum up RS-170A in two pages; MXF standards run to hundreds of pages. It’s the sheer complexity that makes it difficult to achieve the vendor independence that broadcasters look for.

Before we can move to an SOA or some other IT-based architecture, all these nitty-gritty standards and interfacing issues must be solved.

The sectors that have moved forward with this architecture are often simply exchanging files. These are generally text documents; it could be information, quotes, orders, etc. For the media sector, the product is also a file, but that is where the complexity starts. The files include wrappers, containers, essence, metadata and codecs. To meet the needs of production, these cannot be reduced to a small set. A digital intermediate file will be very different from a desktop proxy recorded for compliance purposes.

2009 has brought into focus the need to improve efficiencies. To survive, the media and entertainment sector can no longer afford to continue with the luxury of working practices that belong more to a craft or cottage industry. To the parlous state of the economy is added the need to publish in multiple formats for multiple distribution channels. Plus, the established players have to compete with entrants from the new media sector.


To ask broadcasters to implement a change in the way they operate will naturally run into skepticism. This year is not the year to rock the boat, or is it? It may be rocking already. The SOA is not new; it’s just new to the media and entertainment sector. It’s a proven architecture; the risks are in the implementation.

One step on the way to deploying an SOA is the provision of a Web service interface as a means of communicating with the middleware running the system. “Broadcast” interfaces have traditionally been highly proprietary, examples being the Louth (Harris) VDCP used to run video servers and the Sony VTR control interface.

Encoding and QC are typical services. Bruce Devlin of AmberFin says, “We are working closely with MAM vendors to make sure we have good interfaces to their systems. Those looking for a new way of doing things and taking cost out of a system are looking at Web services. Those that integrate with our products using the traditional interfaces will take a week to get an interface working; using Web services typically takes one or two days.”

Multinational media companies also are moving to an SOA, and media service vendors are providing the necessary interfaces. But just like the rollout of MXF, I can see many pitfalls and competitive issues causing problems on the way. The use of experts groups — just as JPEG, MPEG and MXF evolved standards — is the way forward. The large vendors may want to provide their own way forward, but broadcasters have an affinity for internationally agreed standards, a lack of vendor lock-in and the freedom to choose best-of-breed solutions.

There is much work to be done. You cannot simply drop an SOA onto existing media services such as transcoding, file movers, QC, etc. Devlin says, “No one has drawn a box around what a transcode service should look like in the abstract, so two or three transcoder vendors could deliver the same functionality. It still requires competitors to agree on the way transcoders work, and then compete on features.”

To satisfy all interested parties, a standard often includes every possibility. This happened with ATSC (progressive, interlace) and with MXF — an example being time code stored in five different locations. As always, the devil is in the details.

Some of the biggest names in IT showed their approach to the media business at NAB2009. Both Microsoft and IBM have SOA solutions that are tailored to this sector. The middleware for an SOA falls into one of two camps: .NET and Java. The system that Microsoft showed at NAB was based on .NET, with BizTalk SOA; Silverlight and SharePoint collaboration and presentation layers; and Office Communications messaging. SQL Server provided business analytics.

Gabriele Di Piazza, managing director for media and entertainment at the communications sector of Microsoft, told Broadcast Engineering, “One of the biggest differentiators of Microsoft in an SOA architecture is not just the backend, the ability to manage processes and messaging, it is about the richness of our presentation layer. If you look at Silverlight or SharePoint, there is a wealth of collaboration possibilities.”

Microsoft partnered with OpenText (Artesia DAM), Signiant and Rhozet to show an end-to-end workflow of ingest transcoding and delivery.

IBM is taking an open-standards approach with its recently launched Abstract Service Definition (ASD), a core proposition for an SOA based on its Media Hub Solution Framework. ASD is designed for ease of use with services like transcoding, water markers and media analytics and is intended to reduce the skills necessary to integrate the services that would be needed by a broadcast infrastructure.

Vendors, broadcasters and media companies are all moving toward the SOA with its attendant advantages. The most interest has been from large corporations such as the Ascent Media Group and BBC, but there is no doubt that the technology will trickle down to smaller players as the advantages in efficiency and business agility become self-evident.