www.networkbulls.com
Best Institute for CCNA CCNP CCSP CCIP CCIE Training in India
M-44, Old Dlf, Sector-14 Gurgaon, Haryana, India
Call: +91-9654672192
The introduction of packet-voice technology allows the convergence of data and voice networks. This lets companies save toll charges on voice telephone calls. It also reduces companies' total cost of ownership by not having to build and operate separate networks for voice, video, and data.
In multiservice networks, digitized (coded) voice is packaged into packets, cells, or frames; sent as data throughout the networks; and converted back to analog voice. The underlying protocols used for these converged services are
Initially, VoFR and VoATM were used but lost ground to VoIP solutions. VoIP is also referred to as IP telephony (IPT) when it is integrated with IP-based signaling and call control. IPT is how almost all new deployments are being implemented.
VoFR
VoFR permits enterprise customers with existing Frame Relay networks to implement packetized voice. Access devices or cards access the Frame Relay network. PBX vendors provide VoFR cards for their switches to support call routing over the Frame Relay network. Figure 15-6 shows three PBXs connected with trunks using VoFR. The PSTN is used for backup if the Frame Relay circuit goes down. The disadvantage of VoFR is that it provides only convergence in the WAN; it still requires local dedicated telephony equipment and networks. It cannot provide convergence to LANs without a network protocol that can span the data link technologies, such as IP.
One standard for VoFR is Frame Relay Forum (FRF) 11.1. It establishes specifications for call setup, coding types, and packet formats for VoFR service. It provides the basis for interoperability between vendors.
A number of mechanisms can minimize delay and variable delay (jitter) on a Frame Relay network. The presence of long data frames on a low-speed Frame Relay link can cause unacceptable delays for time-sensitive voice frames. To reduce this problem, some vendors implement smaller frame sizes to help reduce delay and delay variation. FRF.12 is an industry-standard approach to doing this, so products from different vendors can interoperate and consumers will know what type of voice quality to expect. To ensure voice quality, you should set the committed information rate (CIR) of each permanent virtual circuit (PVC) to ensure that voice frames are not discarded.
VoATM
VoATM permits enterprise customers to use their existing ATM networks for voice traffic. ATM inherently provides guaranteed QoS for voice traffic that IP protocols alone cannot provide. ATM can provide the service levels and functionality required to support voice traffic for the WAN. For enterprise networks that have ATM, VoATM provides a mechanism to connect enterprise PBXs via ATM and other VoATM applications.
With ATM, constant bit rate (CBR) or variable bit rate–real time (VBR-rt) classes of service (CoS) provide levels of bandwidth and delay guarantees for voice. Chapter 5, "WAN Technologies," covers ATM.
PBX vendors provide VoATM cards for their switches to support call routing over the Frame Relay network. Figure 15-7 shows three PBXs that are connected via trunks using VoATM. The PSTN is used for backup if the ATM circuit goes down. As with VoFR, the disadvantage of VoATM is that it provides only convergence in the WAN. It cannot provide convergence within the LAN without a network protocol that can span the data link technologies, such as IP.
VoIP
VoIP provides transport of voice over the IP protocol family. IP makes voice globally available regardless of the data link protocol in use (Ethernet, ATM, Frame Relay). With VoIP, enterprises do not have to build separate voice and data networks. Integrating voice and data into a single converged network reduces the costs of owning and managing separate networks.
Figure 15-8 shows a company that has separate voice and data networks. Phones connect to local PBXs, and the PBXs are connected using TDM trunks. Off-net calls are routed to the PSTN. The data network uses LAN switches connected to WAN routers. The WAN for data uses Frame Relay. Separate operations and management systems are required for these networks. Each system has its corresponding monthly WAN charges and personnel, resulting in additional costs.
With IP telephony, you can reduce the number of systems, circuits, and support personnel. Figure 15-9 shows a multiservice IP telephony network that employs Ethernet-based phones with server-based call processing with gateway routers. Survivable Remote Site Telephony (SRST) is used for failover or backup to the PSTN if WAN failure occurs. On-net calls travel through the Frame Relay network, and off-net calls are forwarded to the PSTN. The PSTN link is also used if voice overflow or congestion occurs on the WAN network. Calls are then routed to the PSTN.
IPT Components
Cisco's IPT architecture divides voice system architectures into four major functional areas, as shown in Figure 15-10:
- Client endpoints
- Call processing
- Service applications
- Voice Enabled Infrastructure
Client endpoints include the IP phones, analog and digital gateways, and digital signal processor (DSP) farms. Included here is Cisco's IP Communicator, which is the software-based IP phone that runs on a PC or laptop. Gateways are used to access PBXs, analog phones, other IP telephony deployments, or the PSTN.
The Cisco Unified CallManager (CM) fulfills the role of call processing. The CM servers are the "brains" of the voice dial plan and are used to establish IPT calls between IP phones.
Service applications include IVR, Auto Attendant, and Unity Unified Messaging System for voice mail. Cisco IP Contact Center (IPCC) is used for enterprise call center applications. In addition, a standards-based Telephony Application Programming Interface (TAPI) allows third-party companies to develop applications for the Cisco Unified CallManager.
The voice-enabled infrastructure includes QoS-enabled devices such as LAN switches and routers. These devices are configured to be IPT-aware and provide service guarantees to the VoIP traffic. For example, LAN switches are configured with voice VLANs and Power over Ethernet (PoE) to service the IP phones. Also, WAN routers are configured with queuing techniques to prioritize VoIP streams over other traffic types.
Design Goals of IP Telephony
The overall goal of IP telephony is to replace traditional TDM-based telephony by deploying IPT components on existing IP networks. IPT should be highly available and as reliable as existing voice networks. IPT should provide greater flexibility and productivity while providing lower cost of ownership by using a converged network. IPT also allows third-party software providers to develop new applications for IP phones.
IPT Deployment Models
This section covers the Cisco IPT call-processing deployment models:
Single-Site Deployment
The single-site deployment model, shown in Figure 15-11, is a solution for enterprises located in a single large building or campus area with no voice on the WAN links. There are no remote sites.
A single CM cluster is deployed for redundancy in the server farm, and Unity is used for voice mail, with or without unified messaging. IP phones are deployed on PoE LAN switches. The Cisco Unified CM supports up to 30,000 IP devices in a cluster. Gateway routers are procured with PRI cards to connect to the enterprise PBX (during migration) or the PSTN.
Multisite Centralized WAN Call-Processing Model
The centralized WAN call-processing model is a solution for medium enterprises with one large location and many remote sites. Figure 15-12 shows the centralized call-processing model. A CM cluster with multiple servers is deployed for redundancy at the large site. Call processing and voice mail servers are located in only the main site. Remote-site IP phones register to the CM cluster located in the main site. PoE switches are used to power all IP phones. Remote sites use voice-enabled gateway routers with SRST for redundancy.
On the WAN, QoS features are configured to prioritize the VoIP packets over other packet types. In the event of WAN failure, SRST configured routers forward calls through the PSTN. The PSTN circuit can be used for local inbound and outbound calls at the remote site. In this model, call admission control (CAC) is configured to impose a limit on the number of on-net calls permitted between sites.
Multisite Distributed WAN Call-Processing Model
The multisite distributed WAN call-processing model is a solution for large enterprises with several large locations. Figure 15-13 shows the distributed WAN model. Up to 30,000 users are supported per CM cluster. Several CM clusters are deployed at the large sites for redundancy, and Unity servers are used for messaging. Intercluster trucks are created to establish communication between clusters. IP phones are deployed on PoE LAN switches.This model also supports remote sites to be distributed off the large sites. CAC between the CM and Cisco IOS gateway with gatekeeper (GK) is supported. Also, this model supports multiple WAN codecs. Compression of VoIP is done between sites.
Unified CallManager Express Deployments
Cisco provides Express versions of its CallManager, Unity, and IPCC solutions that are installed in a router. CallManager Express (CME) provides the call processing capabilities of CM on a router. Unity Express and IPCC Express also provide the same services on the router. CME deployments support up to 240 Cisco IP phones. It is a lower-cost solution for small branch offices.
Codecs
Because speech is an analog signal, it must be converted into digital signals for transmission over digital systems. The first basic modulation and coding technique was Pulse Code Modulation (PCM). The international standard for PCM is G.711. With PCM, analog speech is sampled 8000 times a second. Each speech sample is mapped onto 8 bits. Thus, PCM produces (8000 samples per second) * (8 bits per sample) = 64,000 bits per second = 64-kbps coded bit rate. Other coding schemes have been developed to further compress the data representation of speech. Most voice compression codes, such as G.729, begin with a G.711-coded voice stream.
Analog-to-Digital Signal Conversion
The steps involved in converting from analog-to-digital signaling are filtering, sampling, and digitizing. First, signals over 4000 Hz are filtered out of the analog signal. Second, the signal is sampled at 8000 times per second using Pulse Amplitude Modulation (PAM). Third, the amplitude samples are converted to a binary code.
Codec Standards
Codecs transform analog signals into a digital bit stream and digital signals back into analog signals. Figure 15-14 shows that an analog signal is digitized with a coder for digital transport. The decoder converts the digital signal into analog form.
Each codec provides a certain quality of speech. A measure used to describe the quality of speech is the Mean Opinion Score (MOS). With MOS, a large group of listeners judges the quality of speech from 5 (best) to 1 (bad). The scores are then averaged to provide the MOS for each sample. For example, G.711 has a MOS of 4.1, and G.729 has a MOS of 3.92. The default codec setting for VoIP dial peers in Cisco IOS Software is G.729 (g729r8). Other codec standards are shown in Table 15-5. An explanation of the compression techniques is beyond the scope of the CCDA test.
VoIP Control and Transport Protocols
You use a number of protocols to set up IP telephony clients and calls and to transport voice packets. Some of the most significant protocols are
- Dynamic Host Configuration Protocol (DHCP)— To establish IP configuration parameters
- Real-time Transport Control Protocol (RTCP)— For call control
- Media Gateway Control Protocol (MGCP)— For call establishment with gateways
- H.323— For call establishment with gateways from the ITU
DHCP, DNS, and TFTP
IP phones use DHCP to obtain their IP addressing information: IP address, subnet mask, and default gateway. DHCP also provides the IP address of the DNS servers and the name or IP address of the TFTP server. You use TFTP to download the IP phone operating system and configuration. Both DHCP and TFTP run over UDP.
SSCP
SSCP is a Cisco-proprietary client/server signaling protocol for call setup and control. SSCP runs over TCP. SSCP is called a "skinny" protocol because it uses less overhead than the call-setup protocols used by H.323. IP phones use SSCP to register with CallManager and to establish calls. SSCP is used for VoIP call signaling and for features such as Message Waiting Indicators. This protocol is not used in the voice media streams between IP phones.
RTP and RTCP
In VoIP, RTP transports audio streams. RTP is a transport layer protocol that carries digitized voice in its payload. RTP is defined in RFC 1889. RTP runs over UDP, which has lower delay than TCP. Because of the time sensitivity of voice traffic and the delay incurred in retransmissions, UDP is used instead of TCP. Real-time traffic is carried over UDP ports ranging from 16,384 to 16,624. The only requirement is that the RTP data be transported on an even port and that the RTCP data be carried on the next odd port. RTCP is also defined in RFC 1889. RTCP is a session layer protocol that monitors the delivery of data and provides control and identification functions. Figure 15-15 shows a VoIP packet with the IP, UDP, and RTP headers. Notice that the sum of the header lengths is 20 + 8 + 12 = 40 bytes.
WAN links use RTP header compression to reduce the size of voice packets. This is also called Compressed RTP (CRTP). As shown in Figure 15-16, CRTP reduces the IP/UDP/RTP header from 40 bytes to 2 or 4 bytes—a significant decrease in overhead. CRTP happens on a hop-by-hop basis, with compression and decompression occurring on every link. It must be configured on both ends of the link.
MGCP
MGCP is a client/server signaling protocol used to control gateways in VoIP networks. MGCP is defined in RFC 3435. MGCP's primary function is to control and supervise connection attempts between different media gateways. MGCP gateways handle translation between audio signals and the IP network.
MGCP defines call agents and endpoints. Call agents control the gateways. An endpoint is any gateway interface, such as a PRI trunk or analog interface.
H.323
H.323 is a standard published by the ITU that works as a framework document for multimedia protocols including voice, video, and data conferencing for use over packet-switched networks. H.323 describes terminals and other entities (such as gatekeepers) to provide multimedia applications. Cisco IOS gateways use H.323 to communicate with Cisco CallManager.
- Terminals— Telephones, video phones, and voice mail systems—devices that provide real-time two-way voice.
- Multipoint Control Units (MCU)— Responsible for managing multipoint conferences.
- H.245
- Q.931
- H.225
- RTP/RTCP
H.245 specifies messages for opening and closing channels for media streams and other commands, requests, and indications. It is a conferencing control protocol.
H.225 specifies messages for call control, including signaling between endpoints, registration and admissions, and packetization and synchronization of media streams. It performs registration, admission, and status (RAS) signaling for H.323 sessions.
RTP is the transport layer protocol used to transport VoIP packets. RTCP is a session layer protocol.
H.323 includes a series of protocols for multimedia, as shown in Table 15-6.
| Video | Audio | Data | Transport | |
|---|---|---|---|---|
| H.323 protocol | H.261 | G.711 | T.122 | RTP |
| H.263 | G.722 | T.124 | H.225 | |
| G.723.1 | T.125 | H.235 | ||
| G.728 | T.126 | H.245 | ||
| G.729 | T.127 | H.450.1 | ||
| H.450.2 | ||||
| H.450.3 | ||||
| X.224.0 |
Gatekeeper Use for Scalability
As a network grows, multiple gateways are placed to communicate with multiple endpoints. Each gateway in a zone needs to be configured with a complete dialing plan. The number of logical connections is calculated with the following formula:
L = (N * (N – 1))/2
where N is the number of gateways in the network.
For example, a network with 7 gateways would have 21 logical connections. With a gatekeeper, simple dial plans are configured on each gateway, and complete dialing is configured on the gatekeeper. This makes network operations and maintenance easier.
SIP
SIP is a protocol defined by the IETF and specified in RFC 2543. It is an alternative multimedia framework to H.323, developed specifically for IP telephony. It is meant to be a replacement to H.323. Cisco now supports SIP on its phones and gateways.
SIP is an application layer control (signaling) protocol for creating, modifying, and terminating Internet multimedia conferences, Internet telephone calls, and multimedia distribution. Communication between members of a session can be via a multicast, a unicast mesh, or a combination.
SIP is designed as part of the overall IETF multimedia data and control architecture that incorporates protocols such as the following:
- Resource Reservation Protocol (RSVP) (RFC 2205) for reserving network resources
- RTP (RFC 1889) for transporting real-time data and providing QoS feedback
- Real-Time Streaming Protocol (RTSP) (RFC 2326) for controlling delivery of streaming media
- Session Announcement Protocol (SAP) (RFC 2974) for advertising multimedia sessions via multicast
- Session Description Protocol (SDP) (RFC 2327) for describing multimedia sessions
SIP supports user mobility by using proxy and redirect servers to redirect requests to the user's current location. Users can register their current locations, and SIP location services provide the location of user agents.
- SIP redirect server— Call-control device used to provide routing information to user agents
No comments:
Post a Comment