This article broadly looks at the current state of play for production facilities with the use of SDI, hybrid and IT based networks within said facilities, whether those be part of broadcast and/or streaming organizations, or external production facilities providing services to those organizations.

From There To Here

Firstly, a little history of how we arrived at the current situation. Back in the early 1990’s while full resolution digital video was available for certain islands of production, such as editing systems, graphic systems, and character generators, these systems required very specialist hardware and disk storage was prohibitively expensive. Sending video and audio from point to point over IT networks was simply not possible unless you were willing to wait several days for a minute of video to arrive, and even if it did, there were probably many errors. Many IT engineers at the time had no real concept of how large video files actually were, and were generally taken aback when the bit-rate requirements were explained.

In order to address this, offline non-linear editing systems that used compressed video, and generated edit decision lists (EDL), based on timecodes, that could then be used to automatically compile the edit in an “online“ suite. At the time, this was generally done using a pile of tapes from digital tape machines.

Transferring video within a facility then became possible with the advent of SDI serial digital interface, which was standardized for the first time by SMPTE in 1989. Essentially, it allowed signal routing across baseband router infrastructure and allowed real-time parameter control and monitoring.

Quality of the signals could be verified using waveform monitors and ‘scopes. Since then, there have been a plethora of additional and updated standards created through the blood sweat and tears of SMPTE engineers, which allow video and audio, and sometimes data to be routed around facilities via sophisticated routers (aka switchers). In order to do this, broadcast controllers allowed production operations to manage the various crosspoints and routing via control panels. So SDI in its various forms, SDI / SD-SDI / HD-SDI / SDI 3G / SDI 6G / SDI 12G / SDI 24G has more than proven itself within the professional video world over its many years of usage, however it does still require some specialist hardware.

As more generic IP networks and corresponding hardware, storage, servers with fast enough I/O, and more suitable operating systems started to become available, the possibility of using standard COTS equipment became a reality, and that offered a way of reducing costs.

The opportunities that IP based networks offered in terms of flexibility and adaptation post installation soon became apparent, and this led to the increased use of hybrid systems. The greater possibility of being able to reconfigure operations in a more sophisticated way, and quickly, meant that facilities could offer many more options for use, however this also led to a need for much more sophisticated control, monitoring and planning. For larger operations, going to full IP based systems started to make sense, and the availability of large scale fiber optic connections also played a part.

Software Defined Networking

Modern SDN (Software Defined Networking) controllers can perform both managing the real-time IP networks & the media endpoints, and monitoring.  At time of writing, SDN controllers come in many shapes and sizes, there are multiple standards and switching methods, and SDN controllers supplied by different manufacturers can vary hugely. Fundamentally they connect inputs to outputs, but can suffer from the same basic issues that SDI routing could, the need to maintain constant timing. The move to PTP/master clock timing was an essential part of making this move possible.

Controllers & Orchestration Systems

In terms of monitoring, controllers that focus more towards maintaining the integrity of the network(s) may not have device specific monitoring, only seeing errors on the network, not the device itself. A number of manufacturers offer more sophisticated systems, sometimes referred to as Orchestration Systems. These platforms in general offer an over-arching view and control of both Broadcast Controllers, (either from the same manufacturer, or 3rd party), and Network Controllers, to offer a “human friendly” overall control, view, and monitoring of production, services and events.

As part of this, these platforms need to be able to connect, monitor infrastructure and manage both physical and virtual resources. The degree of sophistication involved can vary, more sophisticated versions of orchestration have full API interfacing and hierarchical control. The use of orchestration layers “over the top” of both network and device control, and in many cases also business services, represents an interesting shift in media organization policies. In the past, engineering and technical networks, particularly in areas such as transmission, were firmly separated, usually air-gapped, from other internal networks and services, for security reasons. Preventing unauthorized use of transmission PC’s for game playing, eating up the CPU being a classic misuse.

Planning For 24/7 Monitoring

When designing a full IP based facility, very detailed forward planning is needed, to fully take advantage of the benefits it can bring, and the design of any monitoring whether network or device, can make the difference between having an adaptable and quickly reconfigurable system which reports any errors immediately, and one that simply cannot respond easily to new requirements.

One consideration is whether the monitoring systems should be supplied as part of the orchestration system, or whether an independent monitoring system should be used. Certainly, this is probably the case in distribution systems, where many different partners are part of an output chain, and independently determining where errors are occurring can resolve potential disputes.

In production, in a case where there are many third party devices, and where many different formats and codecs are being used, an independent monitoring system can be useful, however if the monitoring is part of the orchestration, and it is a constant real-time monitoring system, it is possible for the monitoring to not only advise on where errors are occurring, but also perform look ahead tasks, and advise on potential faults before they happen.

Monitoring 24/7 should also be able to identify the problem and give information about the potential impact of that problem. 24/7 dynamic monitoring should offer service level monitoring, including such things as building management (fire safety, air conditioning), operating systems, including virtual machines, classic core infrastructure, servers, network, storage etc, and media flow tracking. In many cases, a combination of tools that are part of the orchestration, and 3rd party monitoring, such as QC systems monitoring video quality, connected using API’s, is probably the best current solution

Whatever the level of pre planning when designing a hybrid or fully IP based system, it is probably still a truism that goes back to the first usage of digital technology in production, particularly with storage and bandwidth, that usage will immediately expand to fill, overload, and exceed the carefully calculated needs within a very short space of time. This can lead to interesting times where, for example, several production suites are utilizing a centralized resource. A well monitored system can soon identify exactly where over use of allocated resources is occurring, with a degree of accuracy and speed that was simply not possible before IP based systems became available.

Looking Ahead

While this discussion has focused primarily on local production facilities, one of the potential benefits of IP based facilities is the possibility to link production and shared resources across other facilities in global organizations with a depth that was not possible before.  A number of organizations have begun to do so, often through the use of cloud compute based resources, and also potentially this can allow more remote working for parts of the operation.

Applying Moores law to the availability of faster networks, faster chips and hardware, gives the production community much to look forward to, as current limitations are overcome, and the software-based tools on offer become easier to set up and use.