Advantages

Decomposition
- No complex APIs, just HTTP and SIP ⇒ rapid development
- User can provide input to the service controller via Web servers, DTMF digit collector, voice portal (via VoiceXML), DTMF input, ... ⇒ just about any internet attached device can be used to provide input.
- Easy to scale
- New services can combine the "best of the best" (thus allowing developers to specialize)
- Servers and services can be located anywhere on the internet and operated by anyone
Decoupling
- Loosely coupled and distributed
  - Flexible location of servers
  - if properly designed, implemented, and operated ⇒ higher reliability and resilience
- Separation of businesses (leads to a rich variety of outsourcing, reseller, … models)
- Since the functions are highly independent ⇒ rapid development
Anyone can introduce a new service

However, if you want to use service components of others, then you may need to work out a suitable agreement (which will probably include an agreement about authorization) ⇒ security can be more complex.

Transcript

[slide486] The advantage is decomposition. No complex APIs, we have HTTP and SIP, we can scale it, we can combine the best of the best, it's wonderful. We get the advantage of decoupling because we now have a loosely coupled system, we can locate these servers in different places, and if were properly designed, we get very high resilience. Because we spread it all out, we have no single point of failure anymore. It means we can actually separate our businesses, so instead of buying it, we may even buy it as a service from someone else. We don't even have to run it. They're very independent so we can do rapid development, and of course anyone can introduce a new service. Sounds great, right? What's the flaw in ointment? we need to worry about how were are going to perform the authentication and authorization. And security is going to get much more complex because we no longer have one entity running all of this. There's no global view.