Online Video Services and DRM Technology

This post is also available in: Russian

DRM

Over the last few years, the multimedia technology has been growing exponentially. Twenty years ago, the analog over-the-air broadcast was the top experience available. Now, online video services make an increasing impact on the TV industry. You will find useful DRM technologies overview in this article (I hope). It was published in the latest issue of Russian magazine Telesputnik.

Video Platform Technology Review
Major contemporary trends in online video services:

  • Multiscreen: anytime access from any device.
  • A variety of business models: many options to access video content: online viewing, pay-per-view, download, in-stream ads.
  • Technological complexity – a variety of video platform modules and features (content load, processing, protection and delivery; video load balancing, video playback, etc.).
  • Providing customers with access to the premium content and the related need in security tools: the major movie studios impose heavy demand on security as they offer very expensive content (multi-million-dollar budgets are common for today’s movies).

A simplified design of a modern video platform for Internet broadcast involves content preparation, protection and distribution to the end user devices (Figure 1). Usually, the source content first enters a transcoding system (ingestion) that generates a set of videos for different end-user devices and bitrates for each device (to enable dynamic bit rate). The prepared content is sent to the operator’s in-house video delivery network or the CDN provider external infrastructure.

SImplified DRM architecture
Figure 1: A simplified online video platform design

There are several most popular options for delivery of content to the user:

  • Downloading a video file to be viewed in a special application (HTTP Download).
  • Downloading a file in fragments, gluing of the fragments together in the video player and viewing of the video content in the process of download (HTTP Streaming).
  • Live streaming using ad-hoc protocols and playback of the video stream 
(RTSP, RTMP, MPEG2-TS).

If the content needs high-quality protection, prior to transferring it to the delivery network it has to be encrypted with a DRM solution. On the attempt of incoming content playback, the video player can detect the encrypted data and request authentication and a decryption key from the license server. At the level of a video player run on a client device, video analytics is very often implemented. It includes collection of detailed viewing data, monitoring of quality parameters, etc.

Specifics of Content Protection
Almost all DRM solutions are built on a unified architecture and consist of two parts: the server part (business logic) and the client part (player). The server part also consists of two modules: the encoder which prepares (encrypts) the content and the license server which issues content playback licenses to the users (players).

In terms of content delivery networks, there is no difference between the secure and open content. DRM solutions usually use asymmetric encryption algorithms, in this case, the content is encrypted with the master key, but users have to apply their unique session-specific keys. In the case of LIVE streaming, to enable a more robust content protection, key rotation mechanisms are typically used. A key is changed at a predefined time interval. In case of VOD, different keys are used for different content units and, if the copyright holder has taken care of this, license caching is also performed. That is, the key is stored on the local machine for a certain amount of time or a given number of views, so that the application does not have to contact the license server at each viewing. All these parameters usually mentioned as a Content Protection Policy which could be different for different major copyright holders and even for different content units.

Major DRM Vendors
There are many Digital Rights Management systems on the market. Each of them has its strengths and weaknesses. The situation is exacerbated by a high degree of segmentation on the end-user device market: Windows (with multiple browsers), MacOS, iOS, Android, TV sets and set-top boxes with their relevant operating systems and technologies. With such a variety of market players, it is pretty difficult to agree on common standards and technologies. Therefore, multiple groups of technical solutions emerge, deferring creation of the universal technology for a certain while.

The most popular content protection tools used by the online video services are the following:

  • Adobe Flash Access. Due to its popularity (most videos on the Web are based on Adobe Flash). This is the perfect solution to protect content in Web browsers, on PCs, and on the Android devices.
  • Microsoft PlayReady is a successor of Windows Media DRM, one of the first content protection solutions. That’s why the Microsoft PlayReady SDK has got installed on most of Connected TV devices (LG, Samsung, etc.) and, of course, on the Windows operating systems (Media Player).
  • Widevine DRM is a recently acquired by Google developer of content protection solutions. Its popularity is due to the support of a broad range of consumer devices (TVs, game consoles, iOS devices).
  • Marlin DRM is one of the few DRM solutions supported in Philips TVs, Sony PS3 and PSP devices. It is notable by its reliance on open standards and SaaS model support (paying per transaction for obtaining licenses).

Adobe Protected Streaming deserves special attention. It implements RTMPE (RTMP Encrypted) to enable secure streaming. This is an alternative to DRM. In contrast to DRM, security is implemented at the transport level rather than at the content level. It includes encryption of multimedia content between Flash Media Server and Flash Player. However this is not the most reliable solution, but still it is approved by most major copyright holders. RTMPE involves additional mechanisms for content protection at the application level, such as: tokenization of links to content (i.e., preventing of direct content access bypassing the video service logic), and user authentication. Compared to full-scale DRM, this approach combines a several times lower cost and ease of launch (a minimum configuring complexity). In addition, no user actions to install the DRM component are required.However, as new distribution technologies are spreading (such as the HTTP Streaming), operators use to give up this approach in favor of DRM that is more universal. In addition, there are certain tools capable of cracking RTMPE.

DRM Solution Architecture
Almost any DRM solution consists of three parts:
The service of content encryption (server) pre-encrypts the content for distribution via the open Web channels in the protected format only. Most of solutions use AES-128. Many vendors have supported HSM modules to optimize encryption and offload the main CPU. Licensing service (server) is a solution making a decision whether to issue a video content decryption key, in accordance with the business logic (content distribution rules). This includes identification of a pending payment or checking of viewer balance sufficiency. In most cases, this is the application server (for instance, it can be Java-based).
SDK (i.e., a set of software libraries) to handle protected content in the video player. It integrates the logic of interacting with the licensing service into the video player, and decodes the video before playback. Also, the SDK tools implement additional security features: control of access to the video player memory, protection of analog and digital outputs, and detection of hacker attacks. SDK validates all technical and logical restrictions set on the server for the content and a particular user, such as the ability to record to a disk or mobile device.

So, at the DRM input, the content is received and access restriction business rules are applied to it. The output video can then be played back on a wide range of consumer devices.
As it has already been noted above, DRM content delivery is independent of its protection method and is determined primarily by the business model. Hence, the same protected content can be made either downloadable (e.g., via the P2P networks, to minimize the cost) or presented for online streaming or viewing. However, when selecting a DRM system, you should check which delivery models and access restriction models are enabled by a given solution.

Streaming video DRM workflow
Figure 2: DRM algorithm for streaming video


Downloadable video DRM workflow
Figure 3: DRM algorithm for downloadable video

The above figures show that, in terms of DRM there is no difference between streaming and video on demand. As a rule, it all comes down to using of different alternative authentication algorithms (identifying whether the download or just online viewing has been paid).
The article shows the workflow diagrams for handling protected content on a video platform, in the streaming and pre-load modes.

DRM Vendor Selection
A critical issue in choosing a content protection solution is a range of consumer devices supported. Each online video service has its own priority of consumer platforms. Here we would like to discuss just the very basics of running content security solutions on various platforms:

  • Web browsers on Windows / MacOS / Linux
  • Desktop applications
  • Web browsers on Android
  • Android applications
  • Web browsers on iOS
  • iOS-applications
  • Connected TV

The key differences between the various content protection solutions consist in supporting the above technological platforms. For instance, Adobe Flash Access offers user-friendly tools of content protection that do not require installation of additional plugins, as Flash Player is already installed almost anywhere (according to Adobe, on 99% of computers and 80% of mobile devices). The need to install additional plugins can filter off those potential service users that possess no administrator rights or have a low level of computer literacy. In addition, there is a great community of Flash developers leveraging the Adobe technology. They have multiple years of experience in creating of cross-platform applications with rich interactive interfaces (which is pretty important for online video services). Other technologies rely on less abundant players and feel an acute deficit of technical experts. Still, it is worth noting that the Widevine based solutions provide efficient content protection for Samsung and LG TV sets and iOS devices. Historically, the solutions based on Windows Media DRM have been supported on an even greater number of Connected TVs. Also, Verimatrix is unequaled among the STB vendors.

Recently, Adobe has shook the industry by announcing discontinuation of Flash plugin in mobile browsers (in fact, this means Android mostly), explaining it by their focus on mobile AIR application development and HTML5. To clarify its stand, just the other day Adobe announced that it is now working on HTML5 DRM. It looks like we are in for something fascinating to come. The rapid spread of the new HTML5 video format is prevented by absence of uniform standards of video codecs (such as MP4, WebM, Ogg Theora) across a variety of browsers and lack of a single reliable DRM system to meet the copyright holder requirements. Different DRM solutions are used to support various video containers (mp4, wmv, ogg, mkv). Now, most of video on the Web is using mp4, so all the DRM vendors are concerned with supporting mp4. But with HTML5, mp4 is far from ubiquitous support.

In the case of Desktop applications the situation is much simpler: the vendors offer various SDKs to integrate DRM into video applications.

Content protection for iOS devices is a separate big topic. The issue is that Apple went its own way and preferred Flash to HTML5 with H.264 video (as this technology is most popular today). For video service developers, only one method of content protection is available: blockwise stream encryption with AES-128. This scheme is supported, for example, in the latest version of Adobe Flash Media Server 4.5. However, the developers have to implement their own license management mechanisms or use an SDK from Widevine or other vendors. In case of browser playback, license management has to be implemented in JavaScript, which posits a substantial vulnerability. In case of playback from an application, Objective C SDK has to be used to enable a much higher level of protection.

In the world of music, DRM solutions showed no viability, especially after Apple iTunes abandoned FairPlay DRM. In our opinion, this is due to a significantly lower cost of audio content as compared to Hollywood blockbusters. As a result, copyright holders take a substantially harder stand with videos. The well-known western movie studios present a long list of technical requirements to companies seeking to distribute their digital content. Practically all solutions of IPTV operators and the major Runet websites with premium content are the examples of content protection systems.

All DRM solutions are scalable and consume roughly the same resources, as the encryption algorithms used are uniform. When choosing the DRM solutions, we recommend you to make a list of user devices based on their priorities. Supported devices (as well as containers and codecs), price, user experience, business models and technologies for content delivery: all these make up the applicable selection criteria.

Leave a Reply