Table Of Contents
Setting Up Basic IP UPC Services
Network-Service Considerations
Establishing a Network-Service Definition
Perform Preconfiguration Tasks
Configure Signaling on Voice Ports
Tweak Your System Configuration
Enabling QoS Features for VoIP
Fragmentation and Interleaving
Traffic Shaping for Frame Relay
Other Bandwidth-Reduction Features
Choosing an SPE-Firmware Update or Upgrade Strategy
SPE Firmware Upgrade Scenarios
Displaying SPE Firmware Versions
Upgrading SPE Firmware from the Cisco.com TFTP Server
Downloading SPE Firmware from Cisco.com
Copying SPE Firmware from a Local TFTP Server
Upgrading SPE Firmware from Diskettes
Copying SPE Firmware to Your PC
Copying SPE Firmware from Your PC to the SPEs
Using the SPE Firmware Bundled with the Cisco IOS Software
Configuring Classic-Split Mode
Classic-Split Configuration Commands
Upgrading to a High-Availability Cisco IOS Release
Configuring Handover-Split Mode
Configuring Handover-Split Mode in Extraload
Handover-Split Configuration Commands
Managing a Handover Split with SS7
Sample Handover-Split Configuration
Provisioning
This chapter describes basic service provisioning, including IP service considerations, configuration of Voice over IP, upgrade procedures, and configuration of split modes. It contains the following sections:
•
Setting Up Basic IP UPC Services
Tip
For details on safety recommendations, site requirements (including shelf specifications, space, chassis heights, rack types, mounting options, power, and plant wiring), and site logs for monitoring installation progress or recording upgrade history, refer to the Cisco AS5850 Universal Gateway Hardware Installation Guide, available online at the Cisco AS5850 product documentation website.
Setting Up Basic IP UPC Services
This section describes how to set up and provision basic IP universal-port-card services using a Cisco AS5850 universal gateway. It is tailored for network engineers who work with dialup access technologies, and assumes that the reader is Cisco certified or familiar with Cisco IOS routers and technologies.
Corporate users and Internet service providers (ISPs) install dialup services to facilitate e-mail, e-commerce, and application/database access for employees, roaming sales personnel, household consumers, and students (see Figure 5-1). As a corporate user or ISP, you can do the following:
•
•
Figure 5-1 Business Scenario
Network-Service Considerations
The network-service definition for a corporate user generally differs from that for an ISP, as shown in Table 5-1.
Establishing a Network-Service Definition
Begin your implementation of basic IP UPC services by establishing a network service definition. Use the perspectives described in Table 5-1 preceding and in the following list of design and configuration considerations as a guide. A conservative approach is to project your current deployment and design into a three-month, one-year, and five-year timeline.
Step 1
•
•
•
Step 2
Step 3
•
•
Step 4
•
•
•
•
•
Step 5
•
•
•
•
•
Step 6
Step 7
•
•
•
Step 8
•
•
•
Step 9
Step 10
Step 11
•
•
Step 12
•
•
Step 13
•
•
Step 14
Step 15
Step 16
Step 17
Configuring Voice over IP
Voice over IP (VoIP) technology enables voice-capable routers and switches to transport packetized live voice traffic such as telephone calls over IP data intranetworks or internetworks rather than public switched telephone networks (PSTN) or private TDM (PBX) networks. VoIP thus enables toll bypass, remote PBX presence over WANs, unified voice and data trunking, and plain old telephone service (POTS)-Internet telephony gateways. VoIP enables more efficient and full use of your existing IP data network, reducing both transmission costs and possibly your need to support dual (voice and data) networks.
Routers and switches such as the Cisco AS5350 and Cisco AS5400 universal gateways can handle origination, transport, and termination of VoIP traffic. They digitize analog voice signals, compress them, package them into a series of discrete packets, and transport them interleaved with data packets. They can transmit VoIP packets to both VoIP and non-VoIP destinations, and can receive both VoIP and non-VoIP calls. When data lines are busy, they can spill traffic onto the PSTN.
To ensure acceptable quality of service (QoS) for your voice users, it is important that you configure your gateway carefully and monitor its performance vigilantly—to ensure, for voice traffic, priority service with minimal loss and delay. Unlike most other types of data, voice is intolerant of almost any form of loss or delay. Users cannot wait for a destination device to reorder packets and request that the sending device retransmit any that are missing, as it does for most other data types.
To configure basic VoIP, in general you need to do the following:
•
•
You might also need to do the following:
•
•
•
•
•
This section briefly introduces the subject of configuring VoIP and overviews the first few configuration tasks. It describes, at a high level, some of the voice QoS features that you can enable. Most important, it points you to other references from which you can gain a broader and deeper look at the subject.
Section contents are as follows:
•
Note
Note
VoIP Basics
Before you configure VoIP on your gateway, you should understand at a high level what happens when you place a VoIP call. Think of each event in a call flow as occurring on one of the several "legs" of a call, as shown in the following typical scenario. Other scenarios are possible, of course, including ones where the call destination is an IP phone and the call never leaves the IP network.
•
•
•
•
Figure 5-2 Call Legs
Legs connecting a local device (typically a phone, fax machine, or PBX) to a gateway are called POTS (plain old telephone service) legs. Legs connecting a gateway to the IP network are called VoIP legs. A POTS or VoIP leg is either inbound or outbound, from the perspective of the associated gateway, as shown in Table 5-2 below.
A gateway conferences two call legs—an inbound POTS with an outbound VoIP or an inbound VoIP with an outbound POTS—to create an end-to-end call through the gateway. A call that passes through both an originating gateway and a destination gateway has four call legs.
Call Flow
Table 5-3 and Table 5-4 detail the general call flow from the perspective of an originating and destination gateway respectively.
Dial Peers
Each kind of call leg into or out of a gateway—inbound POTS, outbound VoIP, inbound VoIP, and outbound POTS—must have assigned to it a set of allowable call scenarios, called dial peers.
•
•
A dial peer is, essentially, a single static route within a routing table. A collection of dial peers constitutes a dial plan.
Syntax
A POTS dial peer has the following syntax:
dial-peer voice tag potsdestination-pattern numberport port#other configurable optionswhere tag is a numeric value of local significance only, number is the full E.164 phone number of the associated endpoint, and port# is the voice port in the gateway through which the call is transmitted once a destination pattern is matched.
A VoIP dial peer has the following syntax:
dial-peer voip tag voipdestination-pattern numbersession target data addressother configurable optionswhere tag is a numeric value of local significance only, number is the full E.164 phone number of the associated endpoint, and data address is where the gateway sends a call whose destination pattern matches the one in the peer.
Matching Rules
A gateway redirects an incoming call along the most appropriate outbound leg. It selects the most appropriate leg by first finding the POTS or VoIP (depending on call direction) dial peer whose destination pattern matches the call's dialed digits. For outbound VoIP legs, it chooses the longest matching dial peer. If more than one such match exists, it checks whether preferences have been assigned those peers and selects the peer with the lowest preference level.
Example
Let us say, for a very simple example (your implementation will be far more complex), that a company has offices in San Jose and Newark. Extensions in the San Jose office are in the range 5000 to 5999, those in the Newark office in the range 6000 to 6999. A caller at San Jose extension 5000 wants to call Newark extension 6000. The dial peers shown in Table 5-5are needed to make this connection.
When the San Jose caller at extension 5000 dials the digits 6000, the originating gateway in San Jose does the following:
1.
2.
3.
The destination gateway in Newark now does the following:
1.
2.
3.
In this west-to-east scenario, dial peers 2 and 4 are used, in that order. If Newark extension 6000 were to call San Jose extension 5000, dial peers 3 and 1 would be used, in that order.
Configuring Basic VoIP
Configuring basic VoIP involves the following:
•
•
•
Perform Preconfiguration Tasks
Before you configure your gateway for VoIP, complete the following tasks. See the earlier sections in this book and the references at the end of this section for the additional information you need to do so.
Step 1
Step 2
Step 3
Step 4
•
•
•
Step 5
Step 6
•
•
Configure Signaling on Voice Ports
Your dial feature card supports ISDN PRI, E1 R2, and T1 CAS digital signaling. Configure your voice ports according to signaling type. Set parameters as needed for input gain, output attenuation, echo cancellation, various timeouts, and translation rules. Defaults are generally adequate, but may need to be tweaked for some networks.
Note
Tip
•
http://www.cisco.com/en/US/products/sw/iosswrel/ps1835/products_configuration_guide_book09186a0080080ada.html•
http://www.cisco.com/warp/public/788/signalling/e1r2config.html
ISDN PRI Signaling
Signaling for ISDN PRI VoIP is handled by ISDN PRI group configuration. If you have ISDN PRI voice ports, be sure to complete ISDN PRI, D-channel, and NFAS configuration, as appropriate.
Ensure that multiframes are established on the serial interfaces (acting as D channel). Then set parameters as needed for input gain, output attenuation, echo cancellation, various timeouts, and translation rules.
E1 R2 Signaling
R2 is an international signaling standard for channelized E1 networks used in Europe, Asia, and South America, equivalent to channelized T1 signaling in North America. There are two elements to R2 signaling:
•
•
If you have ISDN PRI voice ports, be sure to complete E1 R2 signaling configuration. Configure signaling types and, if necessary, set parameters unique to specific countries.
T1 CAS Signaling
Channel-associated signaling (CAS) occurs in-band within the data channel, rather than on a separate signaling channel as is the case (on the D channel) with ISDN PRI. For T1 CAS, specify parameters such as frame type and line code.
Configure Dial Peers
Your next step in preparing to set up dial peers is to determine the configurable options that you want to enable.
Configurable Options
Configurable options are the attributes to be applied to calls handled using that dial peer. These typically include, at a minimum, required quality of service, codec for voice encoding, and whether voice-activity detection is to be enabled. The following attributes, for example, are typical in a VoIP dial peer:
req-qos best-effortcodec g711ulawvadYou have many options and great flexibility in configuring dial peers. Table 5-6 and Table 5-7 show the most common configurable options that you can enable in POTS and VoIP dial peers, respectively, from config or config-dial-peer mode.
Here are a few of the things that you can do with these commands in config or config-dial-peer mode:
•
Example: Use 6... to denote a 4-digit number beginning with 6.
•
Example: Use 6... to denote a 4-digit number beginning with 6; use 9t to denote a variable-length number beginning with 9.
•
Example: Prepend 9 to calls that pass through a PBX requiring 9 to access an outside line; replace prefixes that are stripped by a dial peer because they match the destination pattern.
•
Example: Expand local 5-digit extensions beginning with 7 to the full E.164 number 1-408-7xxx.
•
Example: Establish multiple dial peers, each for a different voice port, and each containing the same destination pattern; the gateway directs inbound calls to the voice ports in sequence until it reaches one that is not busy.
•
Example: Assign preference 1 to dial-peer voice 1, which directs outbound calls over the IP network; assign preference 2 to dial-peer voice 2, which directs calls over the PSTN; the gateway, looking for the longest exact match, finds both dial peers and then uses preference as a tie breaker among those matches.
Tip
•
http://www.cisco.com/univercd/cc/td/doc/product/software/ios120/120newft/120t/120t3/voip5300/
Dial-Peer Configuration Table
The next step in creating dial peers is to create a dial-peer configuration table. Under the following headings, show data for all of your gateways and associated dial peers. Table 5-8 is for the simple gateway-to-gateway scenario described earlier; your own will be far more complex.
Tip
Tweak Your System Configuration
Should you have problems with QoS, try adding the following commands to your configuration:
•
io-cache enablevoice-fastpath enable•
ip route-cacheVoice QoS Basics
Quality of service refers to the ability of a network to provide differentiated service to selected network traffic over various underlying technologies. QoS is not inherent in a network infrastructure. Rather, you institute QoS by strategically enabling appropriate QoS features throughout an intranetwork or internetwork.
Characteristics of Voice Traffic
Voice traffic differs from data traffic in a number of ways:
•
•
•
All of these mandate use of QoS strategies to give strict priority to voice traffic, ensuring reliable delivery and minimal delay for networks that carry both voice and data.
Note
Reliability and Predictability
QoS features for voice focus on two things—reliability and predictability. Reliability ensures delivery without packet loss. Predictability ensures delivery without excessive delay. Together, they serve to eliminate poor-quality voice transmission, including crackles and missing syllables that render a call unsatisfactory or even incoherent to the recipient.
Levels of Service
Voice traffic requires real-time service, with steady and predictable throughput and low delay. In the presence of bursty, delay-tolerant data traffic, you must provide for voice traffic a differentiated—that is, higher-priority—level of service. Because networking equipment and devices that carry both data and voice cannot differentiate traffic that requires high-priority service from traffic that does not, your only means for ensuring that voice traffic is expedited or that it receives constant, predictable transmission across a backbone shared by data traffic is by enabling QoS features.
Effective end-to-end QoS throughout a network must serve disparate users, applications, organizations, and technologies, all at reasonable cost and effort. QoS features enable you to balance service levels for user satisfaction, granting priority service to voice while servicing data transmission to the degree of fairness that you require. In addition, other benefits can accrue: Internet service providers (ISPs), for example, can selectively enable QoS features so as to offer their customers differentiated services with different associated costs, as well as a spectrum of new applications and additional services based on these levels of service.
Cisco IOS software provides many features for optimizing QoS. Fine-tuning your network to adequately support VoIP almost certainly involves enabling some of these features. Be sure to read the sited references as you enable features, as the details of wide-scale QoS deployment are beyond the scope of this document. Also, keep in mind that you must configure QoS throughout your network, not just on the devices running VoIP, to optimize voice performance.
Not all QoS features are appropriate for all network devices and topologies. Edge devices and backbone devices do not necessarily perform the same operations: Edge devices handle packet classification, fragmentation, queuing, bandwidth management, and policing; backbone devices handle switching and transport, congestion management, and queue management. Thus, the QoS tasks that they perform might differ. Consider the functions of both edge and backbone devices in your network, and enable QoS features for each type as appropriate.
Enabling QoS Features for VoIP
The following text briefly overviews some of the most important QoS features that you can enable, and cites references that you need to make informed decisions about the use and optimization of those features. Features discussed include the following:
–
•
•
•
–
References in Additional VoIP Resources provide more information.
Congestion Management
Weighted Fair Queuing
You need to avoid congestion on backbone gateways serving high-traffic, high-speed networks. A weighted-fair-queuing methodology called WRED (weighted random early detection) queues traffic according to priority values that you set (you set voice traffic to critical, for example), sets different packet-drop thresholds for each queue, and drops packets in lower-priority queues as necessary so that higher-priority queues can be adequately served. This ensures that low-bandwidth conversations get through, even in the presence of other high-bandwidth applications.
Tip
•
http://www.cisco.com/univercd/cc/td/doc/product/software/ios120/12cgcr/qos_c/qcpart2/
Low-Latency Queuing
If you need to give voice packets priority but cannot allow them to starve other applications, the recommended queuing methodology is LLQ (low-latency queuing), used in conjunction with IP RTP Priority. LLQ directs voice traffic into a priority queue, but allows you to place limits on the amount of traffic serviced at this and each other priority level before the next-lower priority level is serviced.
Tip
•
http://www.cisco.com/univercd/cc/td/doc/product/software/ios120/120newft/120t/120t7/
IP RTP Priority and Frame Relay IP RTP Priority
IP RTP Priority creates a strict-priority queue for VoIP calls. Only when the priority queue empties does the gateway process the other queues. The feature becomes active only when congestion exists on the interface.
Configure IP RTP Priority when you configure dial peers. Set an IP priority level to specify, in the packet header, that a voice call be accorded class-5 (critical) priority. Other queuing and traffic-management functions such as RSVP detect this information and provide priority service.
If your voice traffic passes through a Frame Relay network, the same argument holds, but the feature is called Frame Relay IP RTP Priority (described in the third reference below).
Tip
•
http://www.cisco.com/warp/public/788/voice-qos/voip-mlppp.html•
http://www.cisco.com/univercd/cc/td/doc/product/software/ios120/120newft/120t/120t5/iprtp.htm•
http://www.cisco.com/univercd/cc/td/doc/product/software/ios120/120newft/120t/120t7/
friprtp.htm
Resource Reservation
You can set things up so that your and any other similarly set-up sending or receiving system can reserve bandwidth, on a call-by-call basis, along a router path by enabling RSVP (Resource Reservation Protocol) on all WAN links that transport voice traffic.
Configure RSVP when you configure dial peers. Do not enable RSVP in conjunction with Frame Relay traffic shaping.
Tip
•
http://www.cisco.com/univercd/cc/td/doc/product/software/ios120/120newft/120t/120t3/voip5300/
Call-Admission Control
You can gracefully prevent calls from entering your Cisco AS5850 from the PSTN when certain resources—such as CPU, memory, and interfaces—are not available to process those calls. Such intervention is called call-admission control.
If your system experiences high CPU usage, large call volumes, or occasional large numbers of simultaneous calls, you need to control two specific aspects of call-admission control: call spikes and call thresholds. Doing so is especially important if you handle transactions involving debit cards, which require AAA and similar types of support.
Configure call spikes to limit the number of incoming calls over a short period of time. Configure call thresholds to define under which circumstances system resources should be enabled.
Tip
•
http://www.cisco.com/univercd/cc/td/doc/product/software/ios122/122newft/122limit/122x/122xa/122xa_2/ft_pfavb.htm
Fragmentation and Interleaving
Transmission of voice packets, usually small (60 to 240 bytes) in size, can be unduly delayed in networks that also transmit large data packets. Fragmenting large data packets into smaller ones and interleaving voice packets among the fragments reduces jitter and delay. Use fragmentation and interleaving in conjunction with a congestion-management technique such as IP RTP Priority and/or RSVP if you have a low-bandwidth (<1.5 Mbps) WAN circuit, but not if you have a high-bandwidth (>1.5 Mbps) WAN circuit. The recommended fragmentation and interleaving methodology is FRF.12 for Voice over Frame Relay, Multilink PPP for VoIP-over-PPP leased lines.
Tip
•
http://www.cisco.com/warp/public/788/vofr/fr_frag.html•
http://www.cisco.com/warp/public/788/voice-qos/voip-mlppp.html
Traffic Shaping for Frame Relay
You must regulate traffic flow so that packets arrive at their destination only as fast as the destination can handle them. You do so by buffering packets that are generated faster than a configured value, and releasing them at that value. It is especially important that you enable traffic shaping in Frame Relay networks, but not in conjunction with RSVP. Do not enable traffic shaping with PPP leased lines.
Tip
•
http://www.cisco.com/warp/public/788/voice-qos/voip-ov-fr-qos.html•
http://www-vdtl/SPUniv/Vofr/FR_traffic.htm
Note
Other Bandwidth-Reduction Features
Voice Encoding
The Cisco AS5350 and Cisco AS5400 gateways offer the following codec (coder-decoder) methodologies for encoding (digitizing and, optionally, compressing) voice:
•
•
•
Choosing a coding methodology is a matter of balancing tradeoffs among several factors, principal among them those listed for various methodologies in Table 5-9.
Table 5-9 Tradeoffs Among Codec Methodologies
(low is optimal)
(low is optimal)
(high is optimal)264 (very high)
0.125 (low)
0.34 (low)
4.1 (high)
8 (low)
10 (med)
10 (med)
3.7 (med)
6.3/5.3 (low)
30 (high)
16 (med-high)
3.9 (med)
1 High bit rate is optimal for voice quality, because the original voice signal is better represented; low bit rate is optimal for network performance, because packets are less apt to be delayed or dropped.
2 Perceived quality is measured in standardized mean-opinion-score (MOS) studies.
Note
Other factors that might enter into your decision, or that you can use to tweak performance, include the likelihood of multiple tandem encodings and how you handle packet fragmentation.
Tip
•
http://www.cisco.com/warp/public/788/voice-qos/voip-mlppp.html
RTP Packet-Header Compression
Because of the repetitive nature of subsequent IP/UDP/RTP (network/transport/session-layer) headers, you can compress them significantly. A recommended methodology is cRTP (Compressed Real-Time Transfer Protocol), which, by tracking first-order and second-order differences between headers on subsequent packets, compresses the 40-byte header to just 2 or 4 (without or with UDP checksum) bytes. Other methodologies may be preferable if cRTP's high CPU usage causes delay. Employ a compression methodology on both ends of low-bandwidth (<1.5 Mbps) WAN circuits, but not at all on high-speed (>1.5 Mbps) WANs.
Tip
•
http://www.cisco.com/warp/public/788/voice-qos/voip-mlppp.html
Serialization Delay
You can control packet (payload) size—which, in turn, controls how long one packet takes to be placed on the system interface. Set this in bytes, ideally equating to no greater than 20 ms (typically equivalent to two 10-ms voice samples per packet). Increasing serialization delay increases end-to-end delay. You want to incur no more than 150-to-200 ms of 1-way, end-to-end delay.
Note
Tip
•
http://www.cisco.com/warp/public/788/pkt-voice-general/bwidth_consume.html
Voice Activity Detection
Because telephone users generally speak in turn, a typical voice conversation contains up to 50 percent silence. A feature called VAD (voice activity detection) causes the gateway to transmit when speech starts and cease transmitting when speech stops. During silences, it generates white noise so that callers do not mistake silence for a disconnected call. By suppress
