Common Control Plane Roadmap: Clouds Ahead, With No Silos in Sight.

Here’s a 36-month outlook on how client, CPE, server and SRM technologies can evolve to support multi-service blending.

As a follow up to the September 2009 article ("IPTV: A Cable Perspective"), this article will focus on common service control as an enabler for multi-service blending. It is advised that the reader refer to the previous article for background information.

First, what problem are we trying to solve?

Services offered today are for the most part operating as a silo to each other. Clients, servers, delivery and devices are unique for each service that the consumer is paying for and using.

This siloed approach makes it difficult to innovate and cumbersome to blend services to provide a new customer experience. Blended services will create new revenue opportunities and also become a competitive differentiator for a service provider. As the need for service blending increases to satisfy new business requirements, so does the need to create a platform that can facilitate these requirements in a very cost-effective way and with a shorter time to market.

Silos and fat clients

As shown in the diagram (see pull-out wall chart between pages 16 and 17) there are multiple backend systems used to deliver services to customers where the relationship between client applications, server (backend) and customer premise equipment (CPE) are all one-to-one, that is, siloed by service.

Today’s client applications running on CPEs are highly complex, closed, fat, and extremely proprietary with small development communities that have low innovation and integration with other services. The server side that is "servicing" client requests also remains siloed with little or no information sharing among different servers and very little, if any, sharing of functionality.

The problem statement needs to be solved from a business and technical standpoint:

Business – Drive new revenue with new service offerings at minimum cost. The cost and effort to deliver a blended service should be incremental and not equal to the cost and effort of delivering the individual service that makes up the blend.

Technical – To achieve this, a technical solution is required that is flexible and extensible while maintaining a controlled quality of service (QoS) and open environment (leveraging standards) for service innovation.

With the goal of blending services to create new product offerings, for example, mobile tracking on TV sets, multi-screen, tele-presence, etc., the need to create a common control and signaling architecture becomes important. What is also important is the ability for different clients and servers to communicate with each other in a common language, that is, a common control plane (CCP) protocol.

To bridge the gap to the legacy world where clients and servers communicate over a native language, there is a need for protocol translators to communicate with the CCP. This article discusses how the CCP will play a critical role in this new architecture for service providers.

CCP defined

Any service delivered to a consumer has a control, bearer, management and billing aspect. The first three are hidden from the consumer who only really cares about the service itself, not about how it works or is delivered.

The control aspect of a service can be associated with functions of authentication, session, resource and policy management respectively while bearer plane is associated with functions of service delivery (discussed in detail in the previously referenced article).

Control functions implemented for traditional service delivery (as it is today) of voice, video, data — referred to as "core services" — have their own control systems that typically do not communicate with each other.

From a cable operator’s perspective, video services are controlled via a digital broadband delivery system (DBDS) while voice services are delivered through a call management system, and data/Internet services via an IP management system. All operate independent of each other because they each have their own communication language (protocols) for control, bearer, management and even billing.

The goal of CCP is to bring the service control aspect together (and all functions associated with it, including session management, resource and policy management) and under one roof. This function is hosted in the network but can also be extended to leverage economics and technologies associated with cloud computing by blending Internet services with core services offered to customers today.

To explain this concept in more detail, a description of the changes to the client, CPE, servers and session and resource management (SRM) over the next 36-plus months follows in the next sections. Please refer again to the chart.

Thinning clients

Today’s clients: Most of today’s applications running on devices are proprietary and fat. Clients are embedded and written for specific hardware configuration with little or no standardization.

They have a large memory footprint and require a high amount of processing ability. This reduces the pace of innovation as there is a small development community creating new applications.

As an example, the electronic program guide (EPG) (including switched digital video client) and movies or video on demand (VOD) clients provide consumers with access to linear and nonlinear video services. There is a one-to-one relationship between EPG client on the set-top box (STB) and the DBDS application server.

Similarly, the VOD client on the set-top communicates only with the VOD application server in the backend.

For every new service offering, a new client would need to be developed, tested and integrated into the DBDS environment. Memory and processing constraints make adding new services even more challenging. There are approaches where the introduction of middleware on the set-top can reduce the complexity of delivering new services by reusing the core set of libraries.

But inevitably, some development would be required.

Clients (24-36 months): With a requirement of blending of services, clients need to be user-friendly and provide a common interface for multiple services. We will need a unification of experiences that requires a unification of clients.

For example, EPG and VOD clients combined will provide a video experience that is both linear and nonlinear. Some vendors already have released such products. Client functionality that was once confined to a device will start to shift to other devices, as well. This is also known as place shifting.

An example would be remote scheduling of linear video services on a mobile device or control of VOD functions, such as fast forward, rewind, pause and play, via a mobile phone. This will largely be possible because of these technology enablers: Intercommunication between application servers through the CCP (enabled by protocol gateways), a common IP network, and a common session and resource manager.

Protocol gateways will translate native protocols from one client to the native protocol of the service requested. The unification of IP networks will allow for intercommunication among services. Finally, a common SRM will help optimize network resources. This is discussed in more detail in the following section.

"In this era (36+ months), clients are thin, plug-ins enable new services, IP networks are converged and cloud computing comes to life."

Clients (36+ months): The next-generation clients may prove to be the most interesting and innovative yet. In this era, clients will truly become universal, allowing for a unified client to run on any device, enabling service ubiquity.

This client will need to be thin (in terms of memory) and certainly "open" to allow for a faster time to market for new services. As an example, this client can adopt Web-based technologies that will create a foundation for early adoption by a large development community. New services enabled via plug-ins to this robust client could then be downloaded from an online application storefront.

Consumers would be able to pick and choose which services they want by simply selecting the appropriate plug-in that enables the service. The model can change from pay-per-service-per-month to pay-per-service-per-use. The plug-ins will communicate with the CCP natively via the CCP protocol that will broker the request and deliver the service to the customer’s device.

Converged CPE

Today’s CPE (hardware): CPE devices typically provide a single service, for instance, digital set-tops for video service, embedded multimedia terminal adapters (EMTAs) for voice service and cable modems for data services.

In some cases, there are up to five devices per consumer household to accommodate the addition of services over time. What began as analog cable evolved to digital, then with the advent of the Internet came high-speed data and voice over IP (VoIP), then managed routers and now wireless devices.

This has created complexities associated with service management that can be solved with a home gateway device.

CPE (24-36 months): In the next two to three years, CPE will also see convergence where multiple services can be seen terminating on a single residential home gateway (HG). For instance, the HG could provide voice, video, data, home networking and wireless services.

These devices may interconnect with other CPE devices in the home, allowing for the sharing of information and entertainment. The HG will likely have multiple QAM tuners (16 or more), MoCA, DLNA, UMA, etc. That would allow the interconnected CPE to be less intelligent and more cost effective, allowing the HG to handle the heavy lifting (network-facing functions).

Servers in the cloud

Today’s servers: Each core service has its own set of servers that reside in the backend or central data center. These siloed servers only communicate to one type of client (as described earlier) and do not share information or functionality with each other.

The reasons are multiple and include the following:

• Incompatible languages, protocols, interfaces, designs and architectures;

• Network firewalls preventing cross communication among services;

• Geographic disparity without available transport.

Evolving to a CCP requires overcoming these roadblocks to allow for blending of services and true service convergence.

Server (24-36 months): Removing some of the key technical roadblocks mentioned previously and expediting the introduction of the CCP requires the creation of protocol gateways. These gateways will talk natively to existing clients and translate to an open protocol for communication with the CCP.

Protocol gateways are meant to be a stopgap measure that can be removed once all clients evolve to utilize a CCP protocol and thereby can connect directly to the CCP. (See again, the diagram.)

Server (36+ months, cloud computing): In this era, clients are thin, plug-ins enable new services, IP networks are converged and cloud computing comes to life. Cloud computing is meant to abstract from a specific service and unify common functions and information sharing across services.

The term cloud computing is associated with Internet-accessible services available to anyone with a web client, such as stock tickers, news, weather, etc. However, in this article, cloud computing is meant to also allow customers of different service providers to securely share information and services with each other.

Cloud computing and the sharing of services among service providers can be easily enabled when peering relationships already exist. Then the offering of new value-added features can be realized that would have been challenging to achieve otherwise. For example, sharing video content libraries across service providers or connecting to a converged over-the-top (OTT) video service.

While these are examples of core service sharing, services such as "presence" could allow friends belonging to different service providers to share views, opinions or information about what they are watching. What was once a nonsocial service such as television becomes social. This is part of the TV 3.0 evolution. (For more background, see previous article.)

This is a shift from the current service delivery models where services and information were closed to the outside world. The new architecture has numerous benefits because it encompasses all core services and allows for the ready addition of new ones.

From a resource perspective, most servers in the backend are under-utilized, while others are over-utilized. By breaking down services to generic functions, they can benefit from the efficiencies of virtualization.

Imagine a Mother’s Day event where suddenly 50,000 offline subscribers start requesting services simultaneously. Today’s systems must be scaled to handle this worst-case scenario by overbuilding capacity that may never be used.

This is not the case in a virtualized server environment, where for example, under-utilized servers could boot up a second operating system that is preloaded with the application required to meet the demands. Later when these subscribers go off-line, these virtualized instances can be released to serve other requests. This allows for dynamic creation of resources that can scale to meet the demands of customers.

The service provider may not even own these under-utilized servers, since leasing CPU cycles from an outsourced third-party computing provider could help manage overflow capacity situations.

Toward a global SRM

Today’s SRM: There is a session management function that manages the control aspect of service delivery. Each of these services is managed through its own SRM with a one-to-one relationship between the client and the server.

As discussed earlier, this function is hidden from the end consumer and runs in the background while services are being delivered. This functionality is also siloed; however, it is on its way to becoming more open.

SRM (24-36+ months): In the next two to three years, blending of linear and nonlinear services will create a need for a unified navigation experience and for a blending of linear and nonlinear SRMs with respect to video session management, for example.

That blending is here termed global SRM (GSRM). The GSRM is the interim step to a truly universal SRM of the future that will manage sessions across all services.

Competitive future

Having a CCP to manage existing core services can expedite time to market for new and blended services.

By implementing a nonproprietary and standards-based CCP, service providers can be in a better position to integrate with vendors entering the market that are offering new and innovative services. Implementation of standards that have a large adoption base also could benefit cost structures; as competition increases, capital expenditures may decrease.

Reduction in operating expenditures may also realized by utilizing mainstream technologies for the CCP that do not require specialized technical staff to maintain. All of this is likely to be key to revenue growth potential, cost savings and competitive advantage for service providers in the future.

Sandip Singh is architect, services, network technology, and Tony Colantonio is manager, advanced video services technology at Rogers Communications. The views represented in this article are the authors’ alone and do not represent or endorse the views of Rogers Communications, Inc. or any particular vendor.