Anyone who has attended a seminar on video security over the last year cannot fail to have noticed that content redistribution over the Internet is set to overtake traditional control word sharing as the biggest single piracy threat to premium content. This has been borne out by events, with some OTT premium sports services finding that as much as 50% of their “customers” do not have a legitimate subscription. Instead they have either bypassed security controls directly or accessed a pirated stream that they may have paid for, or that may instead carry advertising.

Live sports streaming is a particular problem because the window of value only lasts an hour or so, so that it is not much use waiting to take action against the perpetrator or taking down a pirated stream after the event has finished. By then the content has only a fraction of its original value. The challenge for live streaming, primarily of premium sports because that is where the greatest value lies, is in being able to take down pirated streams in near real time, in a matter of minutes.

For premium non-live content such as movies whose window of maximum value lasts weeks or months rather than minutes, the challenge is rather different, with greater emphasis required on acting against offending sites. This has risen up the agenda for some of the big content houses and movie studios because of the recognition that like it or not the Internet will be an increasingly important medium for distribution of their premium assets. They therefore stand to lose revenue if they continue to confine OTT rights to secondary windows of lesser value. Furthermore some are also in the technology or pay TV business and harbor ambitions of becoming major linear OTT operators themselves, as in the case of Sony, which therefore has been at the forefront of efforts to combat Internet content piracy.

Although the challenges posed by live and non-live content piracy are different, there is growing consensus within the content security industry that the solution has a major common element, which is forensic watermarking. The reason for this is that only insertion of a watermark capable of identifying not just the content itself but also individual streams as they are transmitted over the Internet can combat OTT piracy. The ability to track down pirated streams to their source relies on being to identify them uniquely through recovery of embedded watermarks.

A key move came in April 2014 when Movie Labs, representing the six major Hollywood studios, mandated use of forensic watermarking for protection of Ultra HD content, specifically to protect against content redistribution(). This will help ensure that video watermarking becomes a key part of the OTT ecosystem for premium content. The same applies to live events and sports as major broadcasters such as Sky and NBC now all agree that watermarking will be an essential component of their defense against stream piracy.

So far live IP stream piracy in particular has received relatively little attention from the industry, as opposed to defenses against traditional attacks on set top box based broadcast or multicast services, which is still a $3 billion a year business. That is set to change fast, following the first dedicated technology to protect against live streaming launched in July 2014 by Birmingham, UK, based Friend MTS. Called Exposé Source, this is based on the firm’s existing global monitoring and network forensics platform designed to uncover sources of illegal streaming automatically. The new addition is forensic watermarking.

The existing platform is based on fingerprinting, which involves taking a small snapshot of video content and then storing it in a database. This enables content itself to be identified through extraction of the corresponding fingerprint and comparison against the database. But it does not enable a specific stream to be identified, since all streams would yield essentially the same fingerprint. However watermarks inserted at the server end within the content owner’s domain can be made specific to the unicast stream, so that subsequent recovery can identify in principle the receiving device. This relies on the fact OTT distribution is unicast, so there is a one to one relationship between the sending server and receiving device.

In the case of OTT stream piracy, the perpetrator may have broken the security to access a live service illegally. Alternatively pirates may pose as legitimate subscribers, but pay using a stolen credit card to hide their identity. In either case locating the source of the piracy does not lead directly to the culprit. In any case the pirate may be located in a country within which action may be hard to take or that does not uphold the same legal principles as others. The approach advocated by Friend MTS therefore is to act not against the ultimate source of piracy but points downstream from them.

The key point here is that pirates generally rely on legitimate infrastructure, such as CDNs (Content Distribution Networks), to deliver to their “clients”. Friend MTS aims to take advantage of this to remove streams from this infrastructure, through co-operation with ISPs and CDN providers, which have no desire to be delivering unauthorized content over their networks. In the event that a particular content network in a given country is uncooperative, there is the potential to go higher up and take action through major IP peering points. The objective here is not to eliminate illegal streaming altogether, which is practically impossible, but in the words of Friend MTS CTO and founder Jonathan Friend, to “reduce it to economically insignificant levels”. In practice this might mean confining it a few thousand customers rather than a few hundred thousand.

One hurdle that has to be overcome by providers of video watermarking is that in order to operate at an end to end level the technology has to be incorporated in the video encoding system. For this reason Friend MTS has been busy courting the major encoding vendors like Ericsson and Harmonic, and claims that the response has been positive. It is even possible that these encoding vendors could tout embedded forensic watermarking as a selling point.

At any rate forensic watermarking is going to become an important component of content protection, opening a new chapter in the arms race between rights holders and pirates.