Measure Twice, Cut Once
Download
Report
Transcript Measure Twice, Cut Once
Measure Twice, Cut Once
The Goal - One network for everything
Today
Tomorrow
Internet
Telephone
network
Mobile radio
network
IP-Network
Multimedia Access - Advantages:
• easy to handle
• reliable
• mobile
Key drivers of convergence
Short Term objective:
Create new revenue possibilities
Removal of boundaries between voice, data and video opens the
way to new types of services
Can be realized relatively quickly with limited investment
Long Term objective:
Realize cost savings
Simpler network
More efficient network
Cheaper network components
Full benefit only realized when all separate networks have fully
migrated towards the target solution
Varied sensitivities of different types of traffic
Sensitivities
Traffic Type
Bandwidth
Loss
Delay
Jitter
Voice
Very low
Medium
High
High
E-commerce
Low
High
High
Low
Transactions
Low
High
High
Low
E-mail
Low
High
Low
Low
Telnet
Low
High
Medium
Low
Casual Browsing
Low
Medium
Medium
Low
Serious Browsing
Medium
High
High
Low
File Transfers
High
Medium
Low
Low
Video Conferencing
High
Medium
High
High
Multicasting
High
High
High
High
QoS issues for different types of
traffic
Voice traffic is smooth, bandwidth-benign, drop-sensitive, and delay-sensitive,
and is typically UDP-based. Bandwidth per call depends on the particular codes
adopted, sampling rate, and Layer 2 media employed. Voice quality is directly
affected by all three QoS quality factors (loss, delay, and delay variation).
Data traffic is much more varied. It can be smooth or bursty, bandwidthbenign or bandwidth-greedy, or drop- and delay-insensitive, and involves
Transmission Control Protocol (TCP) for send/receive acknowledgment and
retransmit. Traffic patterns vary by application, and data classes must support
several different priorities or application categories.
Video traffic is bursty, bandwidth-greedy, drop-sensitive, and delay-sensitive.
IP-based videoconferencing has some of the same sensitivities as voice traffic.
Complex architecture
Testing Lifecycle - Successful Approach
Step 1
Step 2
Step 3
Step 4
PreDeployment
Assess,
Design,
Baseline
Install
&
Cutover
Manage:
Identify,
Diagnose,
Repair
Optimize
Lack of a proper lifecycle will:
Drive Costs Up
Reduce VoIP Reliability / Availability
Risk Complete Failure of Deployment
What types of tests should be conducted
Conformance test
Functional test
To verify that the device does everything it is supposed to do (i.e.
protocol support, encoding signalling, management, etc.)
Performance test
To verify the device’s behavior in correspondence with the
standards, conventions and rules.
How the device behaves under maximum load conditions
Voice quality test
Measuring the voice quality using standardized, automated scoring
Devices that can effect a users
experience
Require testing:
IP PBXs
IP Phones
& VoIP Endpoints
Media Gateways
IVR / Voice portals
SBCs (Border Controllers)
Media Servers
Firewalls/ALGs
Messaging Servers
Conference Bridges
What is required - signaling protocols
testing
Signaling Protocols – that need to be tested
H.323
SIP
MGCP
Megaco/H.248
SIP-T
SIGTRAN
BICC
Correlation testing (SS7/SIP/MGCP…)
Multiple call legs
What is required - performance testing
Performance - under different levels of stress:
Signaling and media quality evaluation
Call handling
Media degradation
Interoperability
Different codec models
PSTN and VoIP devices
Reliability
Quality evaluation
Stability testing
What are we looking for
Predicting the element behavior in real life:
How does the UUT (network element / service)
behave in real life (=stress…)
Where is the “breaking point”
What happens at the “breaking point”
How are new / existing calls affected?
Is the media degraded?
What are we looking for
Analysis of traffic – measurement/correlation
Jitter
Packet lost
Packet burst
Delay
Codec (G.711/G.729 etc..)
DTMF
Signaling (SIP, SIP-T, H.323, MGCP, Megaco, ISUP etc..
And more …
The Business Drivers for Convergence
Testing
Technology is not the driver
Services and delivery is the “key”
Access
Finding new ways to reach the customer
Increasing accessibility of services to the customer
• Mobility of voice
• Access through multiple devices
• Bypassing ownership of physical plant
Services
Enhanced ability to use existing services
• Voice
• IM/Presence
• Video (streaming/IPTV)
In other words…
Reducing operations costs
Eliminating redundant systems and infrastructure
Reduce equipment capex
Generate New Revenue
More minutes/bits of use
New billable services
Compete for integrated communications package
Offer new services such as video to increase subscriber
revenues
Triple Play or Quadruple Play?
• Up sell services
• Data, voice, video, mobility (WIFI)
• Delivery of QoS services will increase revenues
Testing Lifecycle - Successful Approach
Step 1
Step 2
Step 3
Step 4
PreDeployment
Assess,
Design,
Baseline
Install
&
Cutover
Manage:
Identify,
Diagnose,
Repair
Optimize
Successful predeployment testing ensures
optimized deployment of network and services
But it is not enough!
Post Deployment
Monitoring and testing of deployed network and
services must be constantly
Tested in a production network
Monitored 24/7 to:
• Identify issues related to performance, media and signaling integrity
and end-user experience
• Troubleshoot production network to identify subscriber issues
• Analyze both signaling and media traffic
Centralized Database management
Distributed Monitoring System
Benefits of Distributed Monitoring
Detailed performance measurements
Alarming on performance issues
Service Level Verification
Combine measurements from network elements and
monitoring devices – network-wide-view
Automated reports, threshold alerts and baselining
Real-time reporting and performance on subscriber
delivered services
Correlation of signaling and media across the network
(PSTN/IP) for problem resolution
Protects Investment
From pre-deployment tests
In
In
In
In
customer churn
corporate image
signing new subscribers
offering new services with QoS
Thank You