This chapter gives you an understanding of how Voice over IP (VoIP) is deployed in service provider (SP) networks. This chapter focuses on describing a use case in which the VoIP infrastructure and the transport and the access are managed by an SP. Chapter 4, "Internet Telephony," focuses on VoIP networks in which only the VoIP infrastructure is managed. Different network components and their functions are described to illustrate how various call functions are implemented to provide voice services to residential and business customers. Figure 3-1 depicts a block architecture of the SP scenarios discussed in this chapter. Here, the service provider also owns the last-mile network access. Later chapters cover scenarios where the SP does not own the access network.
Figure 3-1 Service Provider Architecture Overview
This chapter provides a high-level view of the connectivity between different components in a VoIP SP. You learn about the common VoIP networks and the corresponding components. The intention of this chapter is not to provide design guidelines or technology-specific reference material, which is outside the scope of this book, but to offer a collection of metrics from across the various VoIP architectures. As a general note, the acronym KPI (key performance indicators) is used throughout the book to refer to key protocol counters or metrics.
This chapter covers various VoIP applications in the SP market; residential application is geared toward providing primary- or secondary-line voice services to SP's residential customers. These customers include existing high-speed data subscribers and new subscribers who are looking at either replacing their current circuit-switched telephone line or adding additional phone lines to their household. This gives SPs a chance to provide bundled services to their customers.
Another application covered in this chapter is Small/Medium Business (SMB) application, which is geared toward business customers. SPs can provide high-speed data and digital voice services to their business customers using their IP infrastructure. For the SMB, using IP infrastructure can be a more cost-effective way of getting voice services as compared to a traditional circuit-switched phone line from the telephone company.
Other applications discussed in this chapter include IP trunks, which are used for traffic offload and public switched telephone network (PSTN) bypass, and Sessions Border Controllers (SBC), which are used for offloading VoIP traffic to the PSTN, network hiding, and voice transcoding.
The latter part of the chapter highlights some of the security-related issues in SP voice networks. These issues include denial of service (DoS) attacks, theft of service, and other issues that are common in existing IP networks today.
The last part of the chapter discusses common issues and problems related to voice in SP networks. Because VoIP is primarily deployed on a converged IP network, it faces many of the same challenges as other data applications, such as failures in the network, routing protocol convergence issues, oversubscription of network resources, and so on. However, because VoIP is more sensitive to things like delay and jitter, it's important to proactively monitor the health of the SP network and prevent network outages or performance degradation that can cause loss of service to its customers. These issues are discussed in more detail in Chapter 6, "Managing VoIP Networks," Chapter 7, "Performance Analysis and Fault Isolation," and Chapter 8, "Trend Analysis and Optimization."
Service Provider Voice Implementation Models
This section goes into the details of different SP voice deployment models. Various network components and their functions are discussed with illustrations. There are two different VoIP implementation models in SP networks:
- Centralized Switching Model: In this model, the call-processing functions are controlled by a central entity such as a Softswitch (Call Agent or Call Management Switch [CMS]), which passes call control information to different network elements, sets up and tears down calls, and keeps data records for the calls as Call Detail Records (CDR). The endpoints do not need to have intelligence in regard to initiating or terminating calls; they receive the information from the Softswitch and carry out the necessary call functions.
- Distributed Switching Model: In this model, the call-processing functions are distributed to different network elements. A single entity does not control the various call functions. In this model, the endpoints have call intelligence and can initiate and tear down calls without a centralized entity controlling them. The current VoIP SPs are hesitant to go this route, because it makes the end VoIP clients fatter or richer in features and they do not need to subscribe to the SP's premium services. IP Multimedia Subsystem (IMS) is the route that SPs are looking into where presence servers are used to track the end clients.
This chapter primarily focuses on the centralized switching model because most of the current SP deployments are based on this model. The other common distributed switching model is introduced briefly, but it is discussed in more detail in Chapter 4, which also covers some of the current Peer-to-Peer Distributed switching models. The next section covers how the centralized and distributed switching models are deployed in different SP networks.