Managed Network Mashups: Monetizing Micromarkets

Ben Brauer, Microsoft Corporation

Introduction

For the past six to eight years, people have talked about the Semantic Web with hyperbole and catchphrases. We commonly hear terms like "SOA", "mashups", and "Web 2.0", which have become the marketing jargon of the next generation of stump-thumpers who make a living from explaining gee-whiz technologies. Although these terms are based on sound architectural principles with many clear benefits, they have become too technically loaded and miss many of the important business drivers that make up the true nature of the latest disruptive business models. This paper’s objective is to outline some of the primary business drivers that service providers are facing and how to utilize the newest evolution of Web technologies with a pragmatic objective in mind: make more money!

The most important business aspect to understand about the Semantic Web is that the effort is focused on top-line growth versus cost-cutting initiatives. Many people have described service-oriented architecture (SOA) as a highly cost-effective methodology for implementing enterprise application integration (EAI). Although well-governed Web services integrations do help with such cost rationalization activities, this is only one scenario—one that will have limited success and does not deserve the hype. If a business can harness the numerous network effects that occur in today’s thriving Web communities, , the potential revenue streams become something to talk about. For instance, in his book The Long Tail, Chris Anderson suggests that the Web and low-cost technologies have made it possible to finally address the underserved demand that results from scarce resources (for instance, shelf space, capital, bandwidth) and superficial barriers (such as geography, regulations, culture). Anderson’s research has determined that this underserved demand could double the size of an addressable market while actually decreasing the costs of addressing it. Several examples specific to service providers are described below.

The second most important business aspect is the fundamental economic impacts the Semantic Web will have on traditional business models that evolved from scarcity and superficial barriers. Disintermediation and margin rationalization have been hallmarks of the Internet’s near-zero marginal costs of distribution, but these are just the beginning. Context-based search capabilities and filters make it easier than ever for people to find low-cost alternatives that exactly meet their needs. As a result, advertising-based revenue models have become the norm, and sticky content and services have become critical. Service providers not only need to take into account the highly positive benefits of the next technology wave, but also to safeguard against getting blindsided by it.

The tactical issues of riding the Web 2.0 wave are almost as daunting as the strategic ones. Without some structure in the ecosystem, the abstraction and openness of the new technologies make it very difficult to monetize the benefits of flexibility. There are also several critical components that only certain custodians can cost-effectively offer, such as location/presence, security, identity management, quality of service, troubleshooting, customer care, centralized billing/metering, and service provisioning. These components will become the sweet spot of the service provider’s offering—plus, the service provider can charge for them. The term coined for this sweet spot is the "managed network mashup."

The Services Network

The service provider’s typical network configuration started as a single broadcast domain that had severe constraints limiting its ability to scale. The solution was to contextualize the network by creating subnetworks with specific criteria, such as organization, security, administration, and network boundaries. Bridges enforced this contextualization and ultimately characterized the data link layer. The next layer was meant to eliminate network boundaries by introducing LAN routers and local switches, followed closely by the next evolution, which was based mainly on more-specialized access routers. These access routers had their own administration so that the protocol could stay the same even if its purpose changed within the broader network.

The next set of network layers constitutes the Internet as we know it today. These layers characterize applications, presentation, session, and transport. The primary benefit of the Internet is that it has lost all boundaries, making it flexible, dynamic, and easy to use, while also maintaining necessary reliability and performance.

Network-based applications were historically built in such a way as to take into account this dynamism and flexibility. There applications share several important characteristics:

•         An Advanced Intelligent Network (AIN) architecture

•         Event-driven processes

•         Units of orchestration working together in a collaborative environment—for example, the service logic execution environment (SLEE)

•         Service interaction (synchronization between components that process events)

•         No master service logic that determines how the network works

•         Constraint-based programming

The most significant characteristic is that state was held at the endpoint, and events represent the possible transition of that state. This is an important notion, because there is no flow that describes state, which is very similar to the Web services paradigm. However, given that the network has evolved into a stateful, dynamic but highly performant system, very different characteristics of communications-based and enterprise applications emerge.

Terminology

In order to understand the term "managed network mashup," several other terms must first be defined. All terms relate to Web 2.0 and service-oriented concepts.

Mashup: A mashup is a hybrid Web application that has been combined from two or more Web services, and that provides an entirely difference solution from either of the two disparate services.

Contextually aware: A key difference between Web 1.0 and Web 2.0 is the semantically rich descriptions of how services should collaborate with each other based on the context in which they find themselves. For instance, if two services are behind a firewall, they will have lower trust levels than if collaborating on the open Internet. Another more advanced example would be introducing a third service, inherently changing the collaboration between the original two. The dynamism of such relationships is a key enabler of making simple requests for highly sophisticated tasks at any time and from anywhere.

Collaboration versus Orchestration: In order to implement contextually aware mashups that can dynamically adapt to new contexts, plans for collaboration and orchestration must be both specific and distinct from each other. Collaboration describes the interaction that should occur between two or more services, while orchestration describes a sequence of interactions that should occur between two or more services. Take this metaphor, for instance: You are planning an all-day meeting between you and a potential customer. First, you need to decide the nature of the interaction (or "collaboration"). Is this a first engagement that requires a presentation by you, or is it feedback about a particular response to an RFI that requires a presentation by the customer, or is it an ice-breaker at the local golf course? The second planning item is to determine the agenda (or "orchestration"). Often, the nature of the integration determines the agenda, but not always. For instance, if it is an ice-breaker, the actual conversation that occurs on the golf course does not dictate whether 9 or 18 holes should be played.

Two very important implications exist for distinguishing collaboration from orchestration: 1) you can now reuse the orchestration many times (you don’t need to build a golf course every time you want to have an ice-breaker) and thus lower the costs of building composite applications, and 2) you can use differing levels of orchestration in various mashup scenarios that may need little or a lot of structured sequencing.

Well-enabled service: Well-enabled means that a service is managed as if it were a network element; without manageability, a mashup quickly becomes unstable, rendering it useless and unscalable for a larger population. Well-enabled services provide the essential elements of management.

Services network: A services network acts as a service aggregation layer in which mashups are composed, delivered, and managed by a service provider (see the City Live example in the "Examples of Managed Network Mashups" section below). Although any service can be integrated into the services network in order to leverage core capabilities (for example, location, presence, notification, messaging, and so forth), it is always advised to ensure that each service is well enabled (see "Well-Enabled Service," above) to ensure proper behavior (see Figure 1 below).

Services Network Layer within a Network Stack

Digital marketplace: A digital marketplace is essentially the commercial version of a services network, but several important constructs characterize it differently. First is the notion that there is a hierarchy of control and influence over the services network. At the top of the pyramid are gateways that act as "on-ramps" to the services network. All gateways have two major characteristics in common: a large index to Web content and services, coupled with a large advertising platform that matches queries to ads. Gateways do not own content or sell products, so they are perceived as an unbiased source of information. Examples of gateways are Google, Yahoo!, and the MSN® and Windows Live® networks of Internet services.

In the middle layer are hubs that aggregate services and content within a specific domain. Unlike gateways, they not only index, but also own or control the information and services to which they provide access. The primary value hubs provide is to weed out garbage, allowing customers to quickly gain access to goods and services of high value to them while also instilling trust that the transaction will meet expectations. Examples include NTT Docomo, Televisa, or any online yellow pages.
In the bottom tier are the nodes that supply the goods, services, or information. The nodes attract specific customers/users with specialized offerings that effectively becomes the destination of the search process (starting from the gateway or hub). Examples include company Web sites, news sites, and e-commerce sites.

Mashups are provided and used at every layer of the pyramid, which, in turn, evolves the notion of a rigid hierarchy of a services network that has concentrated clusters of interdependencies. Thus, the introduction of mashups brings fluidity and efficiency to the digital marketplace so that all participants are allowed to function effectively within their communities at much lower costs (see Figure 2 below).

Digital Marketplace

Long tail: According to IDC, the long tail is the engine that drives the digital marketplace. All market-demand curves are concave, with a head near the y-axis and a tail along the x-axis. The head represents the disproportionate amount of demand for the big hits, while the tail represents the demand for all of the other products or services below a certain threshold. Typically, this threshold is determined by the economics of addressing this relatively obscure demand. Scarcity and geography have long been the drivers for determining the cut-off points for addressing certain levels of demand. One of the primary reasons the Internet has done so well is that it has addressed both of these drivers by providing a medium that allows nodes to offer a nearly unlimited set of products and services to anyone, anywhere. Of course, fulfillment remains an issue unless the product or service can be digitized. Now that this underserved demand can be addressed, the addressable market can nearly double (assuming fulfillment issues are resolved) by aggregating all of these niches into a sustainable and structured marketplace (see Figure 3).

Mashups Address Underserved Micromarkets

But the fun doesn’t stop there. If you begin to consider these niches as micromarkets that only require certain catalysts to move them up the demand curve (up and to the left), the addressable market becomes even bigger. Search engines and community-based constructs (such as blogs and filters) have helped tremendously to create buzz and provide exposure of previously obscure products and services. Another important catalyst is the establishment of micromarkets’ validity within the conventional distribution and marketing channels, which are usually hit-driven. Given that hits will never go away (Google Video won’t threaten the business model of $100 million action movies), the established marketing vehicles used to steer cultural norms will determine just how far up a micromarket can rise.

Mashups for the Triple Play

Service providers have been investing in their network services infrastructure in order to offer customers a combination of voice, data, and video services. Most of the infrastructure that service providers are currently implementing relate to Internet Multimedia Subsystems, such as SIP servers and other types of application servers. Although this type of infrastructure helps to deliver core voice, data, and video services, it does not necessarily provide the best environment in which to create, deploy, manage, and monetize mashups based on them. Hence, service delivery platforms and service aggregators have become important to the network layer stack.

A service delivery platform provides the services control layer that enables mashups to flourish as Web-enabled services in a network-based paradigm. Microsoft calls these scenarios "managed network mashups." This infrastructure allows service providers to readily deploy hit value-added services such as Xbox Live®, Windows Live, or Hosted Messaging as anchor services while also building a digital marketplace based on these services. These anchor services can also be mashed up with the service provider’s core network services, such as location and presence, to create even more compelling services. Service providers can also open up their development and production environments to third-party developers because of the supporting security, trust, and QoS features offered by service delivery platforms. The service provider can ‘certify’ that a third party service should be allowed on the network based on criteria that determine whether it’s a well-enabled service. This open approach allows the service provider to offer, deliver, and monetize compelling services that might only meet the demands of a micromarket—though profitably. Service providers can also more easily utilize their existing BSS/OSS systems for billing, customer care, and provisioning requirements.

Examples of Managed Network Mashups

Although the market for managed network mashups is still in its infancy, several examples give a clear indication of its potential. This paper describes two that came from very different origins with different objectives and timelines. The first example is a project known as City Live already deployed in Hasselt, Belgium . The purpose of the project was to provide an interactive experience of services that is contextually aware of the user’s surroundings in the city. The second example is the Premium Yellow Pages project that has begun to look at scenarios in which traditional network services are mashed up with node-specific services to augment the experience of business advertising.

City Live

City Live commercializes the City Service Platform (CSP) that has been developed by i-City. This CSP enables third parties to rapidly develop and deploy applications in an easy way.

Out of the box, CSP comes with various applications and services that can be implemented and used immediately. These applications and services are a result of the ISV ecosystem that i-City has set up to rapidly develop new and attractive services and applications. The following applications/services from the i-City applications/services catalog all come with CSP. More applications are still in development.

Basically, the applications and services for an end user fit into one of three categories:

1.  Communication: Messenger, mail, VoIP, Internet, and so forth

2.  Personal: Calendar, news, games, music, and other applications that the user self-selects

3.  Location-based: City guide, event guide, shopping guide, and so forth

Premium Yellow Pages

Many telecommunications providers make a large portion of their profits from their electronic and paper edition yellow pages offerings. The primary business model is to provide exposure of local businesses to the provider’s subscribers via advertisements and listings. The local businesses pay for the advertisements at different rates depending on the size of the ad, assuming that the larger ads will create more awareness.

Telecommunications providers could extend this business model by using their new triple-play infrastructure and offering targeted advertising. Each of the local businesses could create its own services and then subscribe to a bundle of the provider’s services (for example, location and presence) to make its mashups more relevant (and stickier) for the provider’s subscribers. The telecommunications provider could also offer (at a price) relevant market information about a certain population important to the local businesses by aggregating and analyzing data about that population’s behavior according to a variety of metrics (for instance, usage, location, and time of day).

This targeted advertising scenario could be taken a step further with low-cost mashups that enable local businesses to offer "limited specials" for an extra advertising fee. Telecommunications subscribers could take immediate action either by clicking a banner ad or actually purchasing in person. This scenario becomes much more powerful (and pertinent) when location, maps, and search capabilities are mashed up with the advertisement service that renders the contextually aware advertisement.

Conclusion

Managed network mashups will help to evolve a common set of applications into a digital marketplace that succeeds on more than just advertising revenue. The result will be a mix of business models and revenue streams based on the infrastructure laid by service providers. These models will most likely come in the form of subscriptions, ‘platform taxes’ similar to yellow pages, and revenue sharing. These models will complement the advertising models that are employed at both hubs and gateways.

However, a service provider’s traditional network-based mentality could thwart the advances made by managed network mashups. The traditional view of network services is that they reside in an ordered system of defined atomic motion, such that all services behave and react accordingly; a typical Web 2.0 perspective is that services are in a loosely coupled organic system that continually adapts to an ever-changing environment. The view of managed network mashups is somewhere in the middle, where a predefined set of governance rules and methodologies ensure that mashups are based on well-enabled services. Will the traditional view of network services migrate toward the middle to leverage managed network mashups?

The answer is, "it’s already happening." Several factors already indicate a cultural shift: 1) a younger culture adapts more quickly; 2) service-oriented governance models used within IT will enable quicker adaptation; 3) managed network mashups are increasingly considered a strategic imperative by executive leadership; and 4) easy-to-use platforms that enable mashups are being employed so that all constituents (marketing, development, and operations) can be integrated into the mashup’s lifecycle.