“FREE”

BellSouth IP Telephony Snapshot

A Sales Tool for Network Sales Engineers

The FREE BellSouth IP Telephony Snapshot is a sales tool to get account teams in the door to begin discussing IP Telephony. The tool can be used for Cisco and Nortel IP Telephony Solutions as well as BellSouth Network-Based IP Telephony solutions. The purpose of the tool is to generate customer interest, demonstrate to customers BellSouth knowledge of IP Telephony, initiate VoIP pre-qualification, and to further prove the need for an advanced BCS IP Telephony Readiness Assessment.

Angel Cushma, Product Manager

4/6/04

Description of Service

In order to equip BellSouth Field Sales Engineers with tools to assist with pre-qualification and to demonstrate the need for more in-depth discovery via IP Telephony Readiness Assessments, BellSouth has developed a tool that will allow Sales Teams to demonstrate to customers how their current data network would accommodate voice traffic. This tool use’s Viola’s NetAlly VOIP assessment program and is executed from the DataNetworking Helpdesk. NetAlly is a Web-Base voice over/network reliability assessment tool. By browsing to the BellSouth DNH web site and starting a temporary agent on a customer’s workstation or server, you can determine how voice would behave on the customer data network.

Overview

After discussing the specific VOIP needs with the customer, the Engineer browses to the website and starts an agent on at least 2 workstations at the host and at some of the remotes. The agents have VOIP conversations with each other and report quality statistics back to the host. These statistics include MOS score, packet loss, jitter, and latency. These statistics are then put into a report and emailed to the requesting Engineer. The Engineer will analyze the report to determine if the reported statistics will support the customer needs. Based on the results, the Engineer will determine if additional assessment services are needed to support the requirements the customer previously outlined surrounding VOIP.

The IP Telephony Snapshot allows the Sales Team to better position themselves against the competition - many of whom are also providing similar services for free. It is important to note, however, that this is just the first step in properly qualifying and designing a successful IP Telephony network. Similar to the “pre-assessments” provided by the competition, in almost all cases this assessment is not the end-all to IP Telephony network design. The IP Telephony Snapshot simply demonstrates to the customer how a voice application would operate in the customer’s current network environment, as it exists today. In most cases, the IP Telephony Snapshot will show that the existing network requires further tuning in order to provide toll-quality voice.

As shown by a recent Gartner Dataquest survey (see table above), all networks will require some sort of upgrade, replacement, or reconfiguration initiative in order to effectively support a VoIP application.

Glossary of Terms

MOS Score

Voice Quality testing is typically measured using Mean Opinioned Score (MOS) testing. A MOS score ranges from 1 for an unacceptable call to 5 for an excellent call. A typical range for Voice over IP would be from 3.5 to 4.2. A MOS of 4.0 is considered toll quality.

Packet Loss

Packet Loss causes degradation in voice quality. Packet Loss can occur for a variety of reasons including link failure, high levels of congestion that lead to buffer overflow in routers, Random Early Detection (RED), Ethernet problems, and the occasional misrouted packet.

Example audio file - 5% packet loss:

Example audio file - 10% packet loss:

Example audio file - 20% packet loss:

Example audio file - 40% packet loss:

These examples use G.711

Jitter

High levels of jitter cause large numbers of packets to be discarded by the jitter buffer in the receiving IP phone or gateway. This may result in severe degradation in call quality or large increases in delay. Excessive jitter can result from congestion on LANs, Access Links, low bandwidth WAN links/trunks or the use of load sharing.

Example audio file - 5% packet discard rate:

Example audio file - 10% packet discard rate:

Example audio file - 25% packet discard rate:

Latency

In the presence of high levels of delay the normal "protocol" of conversation breaks down. In addition, delay can make echo problems more obvious and annoying. High levels of delay (generally over 200 milliseconds round trip) can cause problems with conversational interaction. This may be due to the routing of the IP stream, mis-configuration of the jitter buffer (i.e. too large) at either end of the connection or high levels of jitter which are causing an adaptive jitter buffer to grow excessively large. In addition, delay can be caused by low bandwidth WAN links/trunks or congestion.

Methods & Procedures

1. SnapshotAssessment Design

The Engineer must identify the customer’s VOIP needs. This should include:

  • What sites will be participating in VOIP?
  • How many calls will be going from site to site?
  • Will call be allowed from one remote site to another?
  • Will faxing from site to site be required?
  • What IP compression rate will be used?

Before starting test, verify with customer that they are not blocking port 8182 outbound.

2. Scheduling the SnapshotAssessment

The Engineer will call the BellSouth Data Helpdesk at 1-888-611-6464 at least 24hours in advance to schedule the analysis. In order to obtain optimal results, the Snapshot should be scheduled at a time when there is a steady flow of data traffic on the customer network.

3. Obtaining Customer Approval to perform the Snapshot

The Engineer will be responsible for obtaining customer approval to perform the Snapshot on their network. The one page Statement of Work is posted on Sales InfoCenter and Product Encyclopedia. The Engineer will also sign the document and provide the customer a signed copy. The Engineer will keep a copy on file with the actual customer results.

43. Performing the SnapshotAssessment

The analysis can be performed from a single location and workstation or multiple locations and workstations. It is recommended that at least two workstations and as many remotes as possible be included in the analysis. While on site, the Engineer will need to access the internet from a customer workstation. The following steps are needed:

  1. Browse to from each of the workstations being used in the analysis. The login is “engineer” password is “engineer”.

  1. If the customer has Java Runtime Enviroment version 1.3.1_06 the following screen will appear.

3.Click on the “Click once to start NetRegard Agent” to start the agent. When the agent is loaded and communicating, the following will appear:

4.

  1. If Java runtime is not on the workstation, the following will appear:
  1. Click where it says “Click here” to install required version.
  1. Java will then load (choose the defaults on all questions).
  1. When Java Runtime is loaded the following should appear. Click on the “Click once to start NetRegard Agent” to start the agent. When agent is loaded and communicating the following will appear.

Note: If you receive a security warning during the agent connection, choose “Grant Always.”

  1. When the “Viola Ready” Window appears, the workstation is ready.
  1. When this exercise is completed on all the workstations being tested, the Engineer will call the BellSouth Data Networking Helpdesk at 1-888-611-6464 and tell them that you are ready to start the test. They will work with you to identify the workstations most appropriate for the report. They will set the number of conversations between workstations to match the call pattern information the customer shared with you.
  1. Upon completion of test, results will be emailed to the Engineer. The Engineer will provide the account team with the results and keep a copy of results on file as well.

5. Preparing the Customer Deliverable

Once the test results are received, the Account Team can deliver the results to the customer in a format that best suits the customer needs. The “Snapshot” will provide valuable information to a customer about how their existing Data Network would support voice traffic in its current environment. Based on MOS Score, Packet Loss, Jitter, and Latency results the Engineer can determine if a more in-depth IP Telephony Readiness Assessment is needed. It is recommended that the audio Wave files (included in this document) be incorporated into the customized report so that that a customer can clearly hear the impact of how these results would impact QoS on his network. A sample customer report is attached. In addition, a sample Power Point customer presentation is posted on Sales InfoCenter and Product Encyclopedia.

“BellSouth FREE Snapshot IPT AssessmentIP Telephony Snapshot”

Preliminary Evaluation Report

Prepared For Customer Name

Date

Executive Summary

(ENGINEER TO INTERPRET DATA AND SUMMARIZE IN 3-4 SENTENCES. INCLUDING RECOMMENDATION OF ADDITIONAL DEEP DIVE IPT ASSESSMENT IF NEEDED)

Summary of Results

The Preliminary Evaluation procedure is characterized by short test procedures whose primary objective is to identify the maximum number of calls that can be supported per test point pair. Accordingly, the test emulates an increasing number of concurrent VoIP calls during each period and uses an algorithm to calculate the maximum number of calls that still produce a MOS value above 3.6.

Test Point Pairs / Recommended
Number Of Calls / MOS / Codec
dn09.dnhelpdesk.com
DN02.dnhelpdesk.com / 5 / 4.1 / G.711

Glossary of Terms

MOS Score

Voice Quality testing is typically measured using Mean Opinioned Score (MOS) testing. A MOS score ranges from 1 for an unacceptable call to 5 for an excellent call. A typical range for Voice over IP would be from 3.5 to 4.2. A MOS of 4.0 is considered toll quality.

Packet Loss

Packet Loss causes degradation in voice quality. Packet Loss can occur for a variety of reasons including link failure, high levels of congestion that lead to buffer overflow in routers, Random Early Detection (RED), Ethernet problems, and the occasional misrouted packet.

Jitter

High levels of jitter cause large numbers of packets to be discarded by the jitter buffer in the receiving IP phone or gateway. This may result in severe degradation in call quality or large increases in delay. Excessive jitter can result from congestion on LANs, Access Links, low bandwidth WAN links/trunks or the use of load sharing.

Latency

In the presence of high levels of delay the normal "protocol" of conversation breaks down. In addition, delay can make echo problems more obvious and annoying. High levels of delay (generally over 200 milliseconds round trip) can cause problems with conversational interaction. This may be due to the routing of the IP stream, mis-configuration of the jitter buffer (i.e. too large) at either end of the connection or high levels of jitter, which are causing an adaptive jitter buffer to grow excessively large. In addition, delay can be caused by low bandwidth WAN links/trunks or congestion.

Procedural Details

  • Start Date and Time: Tue Nov 18 15:18:05 EST 2003
  • During the procedure, full duplex VoIP tests with variable number of calls were carried out among the designated test point pairs.
  • The detailed results for each pair appear in the following pages.

Detailed Results

Test Point Pair: dn09.dnhelpdesk.com - DN02.dnhelpdesk.com

Codec: G.711

Recommended: 5

Number of calls

The following chart illustrates the average quality for the number of calls tested, along with the low- and the high-designated MOS thresholds. Observe that up to and including 5 calls, the average MOS value is above 3.6.

The following chart illustrates the relative effects of the Codec, delay, jitter, and loss impairments on the call quality. For each pair of columns, the left column corresponds to the direction dn09.dnhelpdesk.com to DN02.dnhelpdesk.com and the right column corresponds to the reverse direction DN02.dnhelpdesk.com to dn09.dnhelpdesk.com

The following three charts illustrate the average delay, the average jitter and the average loss values for the number of calls tested along with the low and the high corresponding thresholds.

Additional Information

The following tables contain designation parameters for each test point pair.

Test Point Pair: dn09.dnhelpdesk.com - DN02.dnhelpdesk.com G.711

Name / Value / Units
Number of Measurements / 1
Base RTP Port / 0
Jitter Buffer / FALSE
Jitter Buffer Length / 4 / packets
Initial Playout Delay / 2 / packets
Quality of Service / 0
Frame Packing / 20 msec (2 samples)
G711 Payload Type / PCMU 64000
Use PLC / FALSE
Silence Suppression / FALSE
Min Number of Calls: / 1
Max Number of Calls: / 5
Step: / 1

Next Steps

(ACCOUNT TEAM SPECIFIC RECOMMENDATIONS)