Transcript Best Practices for Ensuring LAN Performance

Best Practices for Ensuring LAN
Performance
Transitioning Networks to Gigabit Ethernet
Agenda
 Main reasons and motivations for moving to
Gigabit Ethernet
 Ethernet technology overview
 Pros and cons of ANSI/EIA/TIA-568-B tests
 Step-by-step upgrade process to verify capability of
existing pairs to carry 1000BASE-T
2
© 2008 JDSU. All rights reserved.
Why transition the desktop to Gigabit Ethernet?
 Increasing “need for speed”  Decreasing cost
– Cost per switch port
 Streaming audio/video
–
–
–
–
VoIP
YouTube
Distance learning
Radio/TV streams
 On-Line applications
–
–
–
–
3
Presence/Collaboration
Time cards
Expenses
SAP, Oracle, etc.
© 2008 JDSU. All rights reserved.
– Cost per Network Interface
Card (NIC)
– Most existing cabling
infrastructure supports it
10BASE-T






4
Standard released in 1990
First Ethernet over twisted pair standard
Category 3 Cabling
Two pairs used – 1 transmit/1 receive
Manchester Coding
16 MHz
© 2008 JDSU. All rights reserved.
100BASE-TX






Standard released in mid-1990’s
Auto-Negotiation for speed and duplex
Category 5 Cabling
Two pairs used – 1 transmit/1 receive
Multi-Level Transition – 3 (MLT-3) Coding
31.25 MHz
+V
Actual Bit Stream
0V
-V
+V
MLT-3 Encoding
0V
-V
Bit Recovery
0
0
1
0
1
0
0
0
 The voltage level changes in succession: +V to 0V to -V to 0V to +V, etc.
 The voltage level only changes for a 1 bit
5
© 2008 JDSU. All rights reserved.
1000BASE-T





Standard released in 2000
Category 5 cabling
Four pairs used – 250 Mbps per pair full duplex
5 Level Pulse Amplitude Modulation (PAM-5)
80MHz
Clock
2-Level 1
0
Code
11
1
1
0
0
11
01
00
6
© 2008 JDSU. All rights reserved.
0
10
4-Level 10
Code
1
00
1
0
10
10BASE-T/100BASE-TX Media Dependant Interface
 Ethernet devices (NICs in PCs and routers)
transmit on pins 1&2 and receive on pins 3&6.
Referred to as MDI
 Switches and hubs are wired MDI-X (transmit on
pins 3&6, receive on pins 1&2)
T568B
7
© 2008 JDSU. All rights reserved.
T568A
Ethernet Summary
10BASE-T
100BASE-TX 1000BASE-T
Speed
10 Mbps
100 Mbps
1000 Mbps
Frequency
16 MHz
31.25 MHz
80 MHz
Pairs/Pins Used 1-2 NIC Tx 1-2 NIC Tx
3-6 NIC Rx 3-6 NIC Rx
8
© 2008 JDSU. All rights reserved.
Tx/Rx on all Pairs
1-2, 3-6, 4-5, 7-8
IEEE 802.3 vs. ANSI/EIA/TIA-568-B.1
 802.3
– Ethernet standard for rates
from 1Mbps to 10Gbps over
coax, twisted pair, and fiber
optic cabling
9
© 2008 JDSU. All rights reserved.
 568
– Specifies a generic
telecommunications cabling
system for commercial
buildings
ANSI/EIA/TIA-568-B.1 Tests – Pros and Cons
 Pros
 Cons
– Required for some
cable/connector vendor’s
warranty for new installations
– Certifies cable to meet
category 5e, 6 or 6A
– Measures a wide range of
frequency-based results:
•
•
•
•
10
NEXT
PSNEXT
ELFEXT
PSELFEXT
© 2008 JDSU. All rights reserved.
– Does not prove operation of
Ethernet over cable link
– Does not perform Ethernet Bit
Error Rate Test
– Does not measure overall
system noise
– May fail links that would work
fine for 1000BASE-T
Transitioning existing networks to 1000BASE-T
 See if existing plant can support new speed
BEFORE pulling new cables
 CAT5 (or better) cabling already installed
 No real need to run full EIA/TIA tests
 Cost-effective alternative:
– Run actual 1000BASE-T signals and analyze for errors
– Bit Error Rate Test (BERT)
11
© 2008 JDSU. All rights reserved.
Step-by-Step Process
Move to 1000BASE-T with confidence!
Process Overview
 Create a test plan
 Test
– Ensure correct cable is being disconnected
– Basic cabling (wiremap/length)
– 1000BASE-T tests (SNR, Skew, BERT)
 Verify Gigabit connectivity
– Port Discovery
– Ping
 Document results
13
© 2008 JDSU. All rights reserved.
Create a Test Plan
 What cables are you
going to test?
 Where do they run
from?
 Where do they
terminate?
14
© 2008 JDSU. All rights reserved.
Test - Correct cable being tested?
 Hub Flash
 Cisco Discovery Protocol
– Will ID the switch port
 Once SURE, disconnect the cable at the switch
and connect remote test device
15
© 2008 JDSU. All rights reserved.
Test - Perform basic cabling tests
 Remember 10/100 only needed 4 pins but gigabit
needs all 8!
 Opens/Shorts/Miswires/Split Pairs
 Length
16
© 2008 JDSU. All rights reserved.
Measure Delay Skew and SNR
 Skew
– Delay differential
 SNR
– Noise sources: NEXT/FEXT,
echo noise, ambient noise
Signal
SNR – “Distance” between Signal and Noise
Noise
17
© 2008 JDSU. All rights reserved.
Bit Error Rate Test
 Send actual 1000BASE-T signals (PAM-5)
 Make sure the data sent from one end get to the
other end without errors
18
© 2008 JDSU. All rights reserved.
Connect to new switch port
 Disconnect remote test
device
 Connect cable run to new
switch port
 Verify new port is
configured correctly
– Port Discovery
• Auto-Negotiation or manual
setting?
• If Auto-Negotiation, what is
being advertised?
19
© 2008 JDSU. All rights reserved.
Verify Network Layer (IP) connectivity
 Obtain IP host configuration
– Dynamic Host Configuration Protocol (DHCP)
– Or manual configuration
 Ping network devices
– Router
– Domain Name System (DNS)
– Any other IP device
 Connect upgraded Ethernet device!
20
© 2008 JDSU. All rights reserved.
Document Results
 Cabling plan
– Update documentation for
cabling runs
 Test Results
– Use for future
reference/troubleshooting
21
© 2008 JDSU. All rights reserved.
Summary
 For the IT manager
– Using existing infrastructure is cost effective
– Need to test cabling first
– Speed certification ensures 1000BASE-T will work
 For the cabling contractor
– New revenue opportunities
 For everyone
– JDSU has the tools and expertise to help make you
successful!
22
© 2008 JDSU. All rights reserved.
Questions and Resources?
 Q&A
 www.JDSU.com/know for:
–
–
–
–
Validator-NT product information
View Validator-NT product demo
Locate a distributor
See full line of JDSU Network and Enterprise Test solutions
 Recorded Webinar will be made available later this week
 Contact Us:
– [email protected]
23
© 2008 JDSU. All rights reserved.