In Broadcast - November 2012 - 36

I N T E R N A T I O N A L

N E W S

WWW.INBROADCAST.COM | ISSUE 14 - NOV/DEC 2012

InFILE
Demystifying Service Oriented Architecture
We are all aware that, to remain competitive, broadcasters, content companies and facilities need to become much more efficient in the way they work. Everyone is expecting more content with better production values for smaller budgets, and the only way this can work is by using as much automation as possible, and by making the underlying technology as ﬂexible and responsive as possible. This is one of the promises of ﬁle-based infrastructures. Content exists only in virtual form, so it can be viewed and manipulated by many people and processes at the same time, and delivered to multiple platforms just as in the past it would be transmitted over a single linear channel. We will be able to think in terms of workﬂows, we were told, rather than being bound by the chain of hardware we have installed. This is clearly a good thing, and sounds perfect if you talk in top level aspirations like this. In practice it is very much more difficult to implement in a way that will actually deliver the promised benefits. At one extreme you could commission a single, monolithic system from a multi-skilled vendor. You could reasonably presume that this system would then work seamlessly and reliably, and could be maintained and updated without issues. But by deﬁnition it would not be a best of breed system, so you would be forced into some compromises. More importantly, it would force people to learn new user interfaces. An editor, for example, might be extremely competent with Avid or Final Cut Pro. Asking them to learn a new user interface, which will not necessarily be as powerful and will certainly seem strange, is likely to be counter-productive. So this is never a serious solution. You have to bring a number of specialist pieces of equipment together to create a complete infrastructure, and that is likely to involve asset management from one vendor, servers from another, sales and trafﬁc software from a third, and so on. If we are going to reap the full beneﬁts of the automated facility we also have to develop our thinking away from just the broadcast hardware to the whole enterprise software. We want to be able to make business decisions and have them implemented in the hardware. Dave Headdon of consultancy Three Media Associates advocates a four-layer model. “At the top are business processes, the top-level activities like scheduling, advertising sales, commissioning and so on,” he said. “Below this are the operational services. These are the individual workﬂows, complex tasks like ingest, editing and transmission. “The third layer is in many ways the most critical,” he continued. “This is the metadata management and
Dick Hobbs

Keeping a competitive edge in the broadcast industry relies on identifying the key performance indicators, Dick Hobbs examines the vital role that the Service Oriented Architecture (SOA) plays…

Kayak offers customisable workﬂows for a multi-screen media landscape

data ﬂows. Without good metadata there is no prospect of achieving successful automation at any level beyond individual operations. Finally, the fourth layer is the technology infrastructure and the operations that people carry out using it.” There are virtually no products which cross all four layers. ForeTV from MSA Focus is probably the only one in common use, and even that is likely to be integrated with other systems. Much more likely is that the top level will be the domain of enterprise software like SAP.

Removing Complexity
So now we have both vertical and horizontal interaction between different systems. Passing information from system to system means a large number of point-topoint links, and every time another piece of equipment is added the number of interfaces rises exponentially. This is difﬁcult to implement and virtually impossible to maintain: a seemingly innocuous upgrade in one system may cause interfaces to fail, which in turn will cause other systems to crash. Which brings us to the service oriented architecture, which is an established technique in the IT industry for just such a challenge. Ed Casaccia of Grass Valley deﬁnes the SOA in terms of three key

characteristics: “First,” he said, “it is organised around outcomes not technology. We should be deﬁning the tools by the tasks we need to complete, not limiting our workﬂows by what our technological infrastructure will allow us to do. “Second, an SOA bridges a number of distinct systems. This allows it to deliver best of breed performance, which again should be something we continue to demand, while at the same time delivering to the user the functionality required to complete the task at hand: nothing more, nothing less. Third, as its name suggests, it is based on the concept of “services”. These can be delivered by compact and efﬁcient code, connecting and communicating through simple, open interfaces.” I will come back to the “open” part of Ed’s last point in a moment. What an SOA is, then, is a common data bus which moves information around between disparate systems, each of which talk to the bus through a simple and compact piece of code – an adaptor – which extracts or adds only the data that is relevant to that particular device or user. At the device level this makes the system much easier to maintain. If a developer releases a major upgrade,
CONTINUED ON PAGE 38

Grass Valley Stratus – a suite of media production and content management tools in a single application

http://www.facebook.com/InBroadcast http://www.linkedin.com/pub/dir/?first=Phillip&last=Izzard&search=Go https://www.youtube.com/watch?v=M0BC1Gqkc4E&feature=plcp http://itunes.apple.com/gb/app/inbroadcast/id468652679?mt=8 http://www.twitter.com/inbroadcast http://WWW.INBROADCAST.COM http://www.youtube.com/watch?feature=player_profilepage&v=aEM4RXZc3Go http://www.youtube.com/watch?feature=player_detailpage&v=Ekobx-UA-iQ

In Broadcast - November 2012

Table of Contents for the Digital Edition of In Broadcast - November 2012