September 1, 2004

Interoperability and Video-on-Demand:

Will Multivendor Platforms Really ‘Snap Together’?

By Pragash Pillai, Charter Communications; Joe Matarese, nCUBE Corp.; and Cliff Mercer, Kasenna

Recently operators have been pursuing with great fervor the development and deployment of next generation video-on-demand (VOD) architectures. These "next-gen" architectures seek to provide fertile soil for growing a diverse vendor community in which each vendor can focus on its core competencies. This must be accomplished while keeping in check the operational costs associated with rolling out VOD. In addition, through the involvement of more vendors, next-gen interoperability promises to allow for more revenue-generating service offerings.

While a laudable goal, the practical issues surrounding system integration present difficulties in combining multiple vendors' products into an end-to-end solution. In fact, one can argue that few attempts at interoperability in the VOD business have succeeded substantially. For example, some operators have succeeded in using set-top boxes from different vendors on the same headend, but this has happened with only limited success owing to the constraints of headend interoperability, including conditional access systems and program guides.

Typically, VOD deployments include a single brand of set-top box with a closely coupled interactive program guide (IPG), a single brand of video server with closely coupled VOD application and back office system, a single brand of edge device, and a single asset distribution system. Moreover, once deployed, operational inertia kicks in, and the local system is loathe to swap existing, legacy equipment for newer, more advanced components. Why repeat the process of integration all over again?

In addition to the difficulties of system integration, interoperability incurs development costs among vendors who must standardize interfaces and implement those interfaces. Interoperability costs money. Vendors must agree to exact specifications of interfaces between their products, and these interfaces must be developed in a thoughtful manner so as not to handicap the development of either vendor. The interfaces must be thoroughly integrated, often with each vendor replicating some portion of the other vendor's equipment under the guidance of a team of dedicated personnel including engineers and program managers. Finally, the whole system must be validated to ensure proper operability prior to commercial launch. For the operator's part, interoperability often means waiting longer for the solution as various vendors coordinate their product release cycles.

Systems integration expenses tend to be magnified once VOD systems deploy into the field. As any technology matures past the initial rollout stage and into a period of broadening commercial availability, operational costs come to dominate initial capital expenditures. True, operators continue to spend capital launching VOD into new markets and scaling existing deployments to provide greater service availability and more content. However, consider the expense of sustaining what has been placed in the field already. Personnel must oversee and troubleshoot content management systems to make sure that VOD metadata and Moving Picture Experts Group (MPEG) files transfer reliably:

Other personnel must watch over the hardware and software components involved in the VOD session setup lifecycle, including set-top box client applications. Yet other personnel must monitor hardware components such as servers, power and cooling systems, hybrid fiber/coax (HFC) plant equipment, and network gear to replace parts where necessary and to keep them in working order.

Discouraged? Don't be. The alternative—proprietary solutions—is far more disheartening. While proprietary solutions promise to keep systems integration costs low and promise a stable platform for introducing new functionality, proprietary solutions increase operators' exposure to limitations in a particular vendor's technology, management capability or support.

For all the difficulty in delivering interoperable architectures, as long as vendors are provided reasonable financial incentive and vendor participation is not stifled by an unfavorable business climate, interoperability is the ticket to unleashing focused vendor creativity. Rather than having a single vendor try to cover all aspects of the VOD solution, some vendors can focus on video servers, others can focus on VOD applications, others can focus on asset management systems and yet others can focus on operations management. This certainly does not preclude any vendor from developing expertise in multiple areas, if the business climate supports such activity.

Interoperability initiatives

A number of operators have pursued, or are pursuing, VOD architecture initiatives aimed at interoperability. Time Warner Cable's Interactive Services Architecture (ISA)1 is a well-known example. Comcast has a similar initiative underway, as presented in a session at this year's NCTA show.2

Charter Communications has foregone the typically lengthy specifications portion of the process, opting instead to work with its vendors to move quickly to deploy interoperability. While a number of projects are underway, Charter's deployment of the integration of one manufacturer's VOD management system with another vendor's VOD server system, which rolled out in May in Malibu, Calif., is of particular interest because it is among the first of its kind.

The basic concept was to introduce VOD servers that utilize industry-standard, off-the-shelf server hardware and advanced server clustering technology into the constellation of Charter deployments by leveraging another vendor's expertise in delivering end-to-end VOD systems based upon a modular and open VOD back office. This architecture provides Charter the opportunity to deploy multiple hardware platforms using an existing management platform.

Delivering interoperability successfully

From conception to launch, the server/management system integration progressed rapidly, within months. The project incorporated five basic phases: planning, development, integration, validation and delivery. The planning phase involved reviewing the VOD architecture, scoping work effort, establishing milestones and agreeing to intermediate deliverables such as test plans. Because Charter had already deployed the VOD management system elsewhere in the Los Angeles market, the local system was familiar with much of the VOD infrastructure, namely the VOD management application and on-demand application for Scientific-Atlanta set-top boxes, as well as edge devices from Harmonic. The system architecture and deployment architecture are shown in Figure 1 and Figure 2, respectively.

Given the modular nature of the VOD architecture and the well-defined and open application programming interfaces (APIs) designed into the VOD servers, the development and integration phases were straightforward. Utilizing an existing implementation of the Internet Engineering Task Force's (IETF) real-time streaming protocol (RFC 2326 RTSP), the server vendor's development team made the necessary modifications to support the management vendor's RTSP extensions for its streaming server resource interface so that the session and resource manager (SRM) could manage the allocation of server resources in support of subscribers' VOD sessions.

For stream control, the server vendor leveraged its existing implementation of Time-Warner's DSM-CC-based lightweight stream control protocol (LSCP) functionality to support the system management vendor's LSCP-based VOD client. For the propagation to, and cataloging of, assets on the video server, the server vendor used its media services API, based on simple object access protocol/eXtensible markup language (SOAP/XML) over hypertext transfer protocol (HTTP) to implement the management vendor's file services interface (FSI), also based on XML over HTTP.

In terms of the project timeline, it was agreed that the server vendor would own the integration process, and the management system vendor would be accountable to Charter for system validation. Owing to the standards-based nature of the interfaces, development proved to be speedy and efficient. We found that interfaces based on Web services such as RTSP and XML over HTTP, coupled with robust exception handling, can greatly simplify the integration process.

On the management side, implementing these Web services interfaces yields high performance and scalability by minimizing transaction counts and messaging overhead. While on the server side, focusing on standards-based VOD streaming that uses Internet protocol (IP) networking technology for streaming, RTSP for stream control and Web services-based integration API can ease integration. Given the careful planning and cooperative working relationships, the server and management system vendors completed the final delivery phase of the project in a matter of days, installing equipment and bringing up the systems on schedule.

Conclusions

Deploying interoperability demands careful planning based upon a number of prerequisites:

In summary, deployable interoperability, notably that between vendors with overlapping areas of expertise, has arrived. But will operators continue to emphasize the deployment of interoperable systems? Will economic conditions allow vendors to continue focusing on areas of core competency without forcing consolidation?

Charter Communications strongly believes that the answers to these questions are in the affirmative. Interoperability will grow in sustainable fashion, which will introduce the promise of many new and improved services from a diversity of vendors whose ideas can finally gain entry to the marketplace. Best of the breed technology will prevail to create a healthy competitive environment.

References

1 http://interactiveservices.org.

2 Weidong Mao, senior director, advanced technology, Comcast Corp. The National Show 2004, "The Road to VOD: Network Design for Efficient On-Demand Provisioning and Transport."

Pragash Pillai is director of digital engineering, Charter Communications. Email him at Ppillai@chartercom.com. Joe Matarese is CTO of nCUBE Corp. Email him at matarese@ncube.com. Cliff Mercer is director of technology, Kasenna. Email him at cliff@kasenna.com.

Did this article help you? Email comments to lhamilton@accessintel.com.

Bottom Line

Smoothing VOD Integration

When planning a multivendor VOD architecture, be sure to keep these things in mind: