draft-ietf-bmwg-igp-dataplane-conv-meth-00.txt   draft-ietf-bmwg-igp-dataplane-conv-meth-01.txt 
Network Working Group Network Working Group
INTERNET-DRAFT INTERNET-DRAFT
Expires in: December 2003 Expires in: April 2004
Scott Poretsky Scott Poretsky
Avici Systems Quarry Technologies
Brent Imhoff Brent Imhoff
Wiltel Communications Wiltel Communications
June 2003 October 2003
Benchmarking Methodology for Benchmarking Methodology for
IGP Data Plane Route Convergence IGP Data Plane Route Convergence
<draft-ietf-bmwg-igp-dataplane-conv-meth-00.txt> <draft-ietf-bmwg-igp-dataplane-conv-meth-01.txt>
Status of this Memo Status of this Memo
This document is an Internet-Draft and is in full conformance with This document is an Internet-Draft and is in full conformance with
all provisions of Section 10 of RFC2026. all provisions of Section 10 of RFC2026.
Internet-Drafts are working documents of the Internet Engineering Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF), its areas, and its working groups. Note that Task Force (IETF), its areas, and its working groups. Note that
other groups may also distribute working documents as Internet- other groups may also distribute working documents as Internet-
Drafts. Drafts.
skipping to change at page 1, line 45 skipping to change at page 1, line 45
http://www.ietf.org/ietf/1id-abstracts.txt http://www.ietf.org/ietf/1id-abstracts.txt
The list of Internet-Draft Shadow Directories can be accessed at The list of Internet-Draft Shadow Directories can be accessed at
http://www.ietf.org/shadow.html. http://www.ietf.org/shadow.html.
Table of Contents Table of Contents
1. Introduction ...............................................2 1. Introduction ...............................................2
2. Existing definitions .......................................2 2. Existing definitions .......................................2
3. Test Setup..................................................2 3. Test Setup..................................................2
3.1 Test Topologies............................................2 3.1 Test Topologies............................................2
3.2 Test Considerations........................................3 3.2 Test Considerations........................................4
4. Test Cases..................................................4 3.2.1 IGP Selection............................................4
4.1 Local Events...............................................4 3.2.2 BGP Configuration........................................4
4.1.1 Convergence Due to SONET Link Failure....................4 3.2.3 IGP Route Scaling........................................5
4.1.2 Convergence Due to PPP Session Failure...................5 3.2.4 Timers...................................................5
4.1.3 Convergence Due to IGP Adjacency Failure.................5 3.2.5 Convergence Time Metrics.................................5
4.1.4 Convergence Due to Route Withdrawal......................6 3.2.6 Packet Sampling Interval.................................6
4.1.5 Convergence Due to Cost Change...........................7 3.2.7 Interface Type...........................................6
4.2 Remote Events..............................................7 3.3 Reporting Format...........................................6
4.2.1 Convergence Due to Remote SONET Link Failure.............7
IGP Data Plane Route Convergence IGP Data Plane Route Convergence
5. Measuring Convergence Times.................................8 4. Test Cases..................................................7
5.1 Measuring Peak-to-Peak Convergence Time....................8 4.1 Convergence Due to Link Failure............................7
5.2 Measuring Impact of Components for Convergence.............8 4.1.1 Convergence Due to Local Interface Failure...............7
6. Security Considerations.....................................9 4.1.2 Convergence Due to Neighbor Interface Failure............8
7. Acknowledgements............................................9 4.1.3 Convergence Due to Remote Interface Failure..............8
8. References..................................................9 4.2 Convergence Due to PPP Session Failure.....................9
9. Author's Address............................................9 4.3 Convergence Due to IGP Adjacency Failure...................10
10. Full Copyright Statement...................................10 4.4 Convergence Due to Route Withdrawal........................10
4.5 Convergence Due to Cost Change.............................11
4.6 Convergence Due to ECMP Member Interface Failure...........11
4.7 Convergence Due to Parallel Link Interface Failure.........12
5. Security Considerations.....................................13
6. References..................................................13
7. Author's Address............................................13
8. Full Copyright Statement....................................13
1. Introduction 1. Introduction
This draft describes the methodology for benchmarking IGP Route This draft describes the methodology for benchmarking IGP Route
Convergence. The applicability of this testing is described in Convergence. The applicability of this testing is described in
[1] and the new terminology that it introduces is defined in [2]. [1] and the new terminology that it introduces is defined in [2].
Service Providers use IGP Convergence time as a key metric of Service Providers use IGP Convergence time as a key metric of
router design and architecture. Customers of Service Providers router design and architecture. Customers of Service Providers
observe convergence time by packet loss. IGP Route Convergence observe convergence time by packet loss, so IGP Route Convergence
is a Direct Measure of Quality (DMOQ) when benchmarking the data is considered a Direct Measure of Quality (DMOQ). The test cases
plane and not the control plane. The test cases in this document in this document are black-box tests that emulate the network
are black-box tests that emulate the network events that cause events that cause route convergence, as described in [1]. The
route convergence, as described in [1]. Black-box test design black-box test designs benchmark the data plane accounting for
accounts for all of the factors for route convergence time, as all of the factors contributing to route convergence time, as
provided in [1]. The methodology and terminology is to be used discussed in [1]. The methodology (and terminology) for
for benchmarking route convergence and can be applied to any benchmarking route convergence can be applied to any link-state
link-state IGP such as ISIS [3] and OSPF [4]. IGP such as ISIS [3] and OSPF [4].
2. Existing definitions 2. Existing definitions
For the sake of clarity and continuity this RFC adopts the template For the sake of clarity and continuity this RFC adopts the template
for definitions set out in Section 2 of RFC 1242. Definitions are for definitions set out in Section 2 of RFC 1242. Definitions are
indexed and grouped together in sections for ease of reference. indexed and grouped together in sections for ease of reference.
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in
this document are to be interpreted as described in RFC 2119. this document are to be interpreted as described in RFC 2119.
3. Test Setup 3. Test Setup
3.1 Test Topologies 3.1 Test Topologies
--------- Ingress Traffic Path --------- Figure 1 shows the test topology to measure IGP Route Convergence due
to local Convergence Events such as SONET Link Failure, PPP Session
Failure, IGP Adjacency Failure, Route Withdrawal, and route cost
change. These test cases discussed in section 4 provide route
convergence times that account for the Event Detection time, SPF
IGP Data Plane Route Convergence
Processing time, and FIB Update time. These times are measured
by observing packet loss in the data plane.
--------- Ingress Interface ---------
| |<------------------------------| | | |<------------------------------| |
| | | | | | | |
| | Preferred Egress Path | | | | Preferred Egress Interface | |
| DUT |------------------------------>|Tester | | DUT |------------------------------>|Tester |
| | | | | | | |
| |~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~>| | | |~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~>| |
| | Backup Egress Path | | | | Next-Best Egress Interface | |
--------- --------- --------- ---------
Figure 1. IGP Route Convergence Test Topology Figure 1. IGP Route Convergence Test Topology for Local Changes
for Local Changes
IGP Data Plane Route Convergence
Figure 1 shows the test topology to measure IGP Route Convergence due
to local changes such as SONET Link Failure, PPP Session Failure, IGP
Adjacency Failure, Route Withdrawal, and Route cost change. These
test cases are described in section 4.1. These test cases provide
IGP Route Convergence times that consider the Event Detection time,
SPF Processing time, and FIB Update time. These times are measured
by observing packet loss in the data plane. Physical Links may be of
any type, such as Sonet or Ethernet, and any speed.
Figure 2 shows the test topology to measure IGP Route Convergence Figure 2 shows the test topology to measure IGP Route Convergence
time due to remote changes in the network topology. These times are time due to remote changes in the network topology. These times are
measured by observing packet loss in the data plane. Physical Links measured by observing packet loss in the data plane. In this
may be of any type, such as Sonet or Ethernet, and any speed. In this topology the three routers are considered a System Under Test (SUT).
topology, the three routers are considered a System Under Test (SUT).
Application of this topology and test cases described in section 4.2
account for the impact of IGP Advertisement on Route Convergence, as
described in [1].
----- ----------- ----- -----------
| | Preferred | | | | Preferred | |
----- |R2 |------------->| | ----- |R2 |---------------------->| |
| |---->| | Egress Path | | | |-->| | Egress Interface | |
| | ----- | | | | ----- | |
|R1 | | Tester | |R1 | | Tester |
| | ----- | | | | ----- | |
| |---->| | Backup | | | |-->| | Next-Best | |
----- |R3 |~~~~~~~~~~~~~>| | ----- |R3 |~~~~~~~~~~~~~~~~~~~~~~>| |
^ | | Egress Path | | ^ | | Egress Interface | |
| ----- ----------- | ----- -----------
| | | |
|-------------------------------- |--------------------------------------
Ingress Traffic Path Ingress Interface
Figure 2. IGP Route Convergence Test Topology Figure 2. IGP Route Convergence Test Topology
for Remote Changes for Remote Changes
Figure 3 shows the test topology to measure IGP Route Convergence
time with members of an ECMP Set. These times are measured by
observing packet loss in the data plane. In this topology, the DUT
is configured with each Egress interface as a member of an ECMP set
and the Tester emulates multiple next-hop routers (emulates one
router for each member).
IGP Data Plane Route Convergence
--------- Ingress Interface ---------
| |<--------------------------------| |
| | | |
| | ECMP Set Interface 1 | |
| DUT |-------------------------------->| Tester|
| | . | |
| | . | |
| | . | |
| |~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~>| |
| | ECMP Set Interface N | |
--------- ---------
Figure 3. IGP Route Convergence Test Topology
for ECMP Convergence
Figure 4 shows the test topology to measure IGP Route Convergence
time with members of a Parallel Link. These times are measured by
observing packet loss in the data plane. In this topology, the DUT
is configured with each Egress interface as a member of a Parallel
Link and the Tester emulates the single next-hop router.
--------- Ingress Interface ---------
| |<--------------------------------| |
| | | |
| | Parallel Link Interface 1 | |
| DUT |-------------------------------->| Tester|
| | . | |
| | . | |
| | . | |
| |~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~>| |
| | Parallel Link Interface N | |
--------- ---------
Figure 4. IGP Route Convergence Test Topology
for Parallel Link Convergence
3.2 Test Considerations 3.2 Test Considerations
3.2.1 IGP Selection 3.2.1 IGP Selection
The test cases described in section 4 can be used for ISIS or The test cases described in section 4 can be used for ISIS or
OSPF. The Route Convergence test methodology for both is OSPF. The Route Convergence test methodology for both is
identical. The IGP adjacencies are established on the Preferred identical. The IGP adjacencies are established on the Preferred
Egress Path and Backup Egress Path. Egress Interface and Next-Best Egress Interface.
3.2.2 BGP Configuration 3.2.2 BGP Configuration
The obtained results for IGP Route Convergence may vary if The obtained results for IGP Route Convergence may vary if
BGP routes are installed. For results similar to those that BGP routes are installed. It is recommended that the IGP
would be observed in an operational network it is recommended Convergence times be benchmarked without BGP routes installed.
that a BGP session be established on the Ingress Traffic Path
with routes installed.
IGP Data Plane Route Convergence IGP Data Plane Route Convergence
3.2.3 IGP Route Scaling 3.2.3 IGP Route Scaling
The number of IGP routes will impact the measured IGP Route The number of IGP routes will impact the measured IGP Route
Convergence because convergence for the entire IGP route table is Convergence because convergence for the entire IGP route table is
measured. For results similar to those that would be observed in measured. For results similar to those that would be observed in
an operational network it is recommended that the number of an operational network it is recommended that the number of
installed routes closely approximate that for routers in the installed routes closely approximate that for routers in the
network. network.
3.2.4 BGP Route Scaling 3.2.4 Timers
The number of installed BGP routes may impact the IGP Convergence
time. For results similar to those that would be observed in an
operational Network it is recommended that the number of installed
routes closely approximate that for routers in the network.
3.2.5 Timers
There are some timers that will impact the measured IGP Convergence There are some timers that will impact the measured IGP Convergence
time. The following timers should be configured to the minimum value time. The following timers should be configured to the minimum value
prior to beginning execution of the test cases: prior to beginning execution of the test cases:
SONET Failure Indication Delay
IGP Hello Timer
IGP Dead-Interval
LSA Generation Delay
LSA Flood Packet Pacing
LSA Retransmission Packet Pacing
SPF Delay
4. Test Cases Timer Recommended Value
4.1 Local Events ----- -----------------
The test cases in this section use the test topology shown in SONET Failure Indication Delay <10milliseconds
Figure 1. IGP Hello Timer 1 second
IGP Dead-Interval 3 seconds
LSA Generation Delay 0
LSA Flood Packet Pacing 0
LSA Retransmission Packet Pacing 0
SPF Delay 0
4.1.1 Convergence Due to Local SONET Link Failure 3.2.5 Convergence Time Metrics
Figure 5 shows a graph model of Convergence Time as measured
from the data plane. Refer to [2] for definitions of the terms
used. Rate-Derived Convergence Time and Loss-Derived Convergence
Time are the two metrics for convergence time. An offered Load of
maximum forwarding rate at a fixed packet size is recommended for
accurate measurement. The test duration must be greater than the
convergence time.
Ideally, Convergence Event Transition and Convergence Recovery
Transition are instantaneous so that the
Rate-Derived Convergence Time = Loss-Derived Convergence Time.
When the Convergence Event Transition and Convergence Recovery
Transition are not instantaneous so that there is a slope, as
shown in Figure 5, the accuracy of the Rate-Derived Convergence
Time and Loss-Derived Convergence Time are dependent upon the
Packet Sampling Interval.
Under this condition and the Packet Sampling Interval <= 100
millisecond, the Rate-Derived Convergence Time > Loss-Derived
Convergence Time and Rate-Derived Convergence Time is the preferred
metric. Under this condition and the Packet Sampling Interval > 100
millisecond the Rate-Derived Convergence Time < Loss-Derived
Convergence Time and Loss-Derived Convergence Time is the better
metric. For all test cases, the Rate-Derived Convergence Time
and Loss-Derived Convergence Time must be recorded.
IGP Data Plane Route Convergence
Recovery Convergence Event Time = 0sec
Maximum ^ ^ ^
Forwarding Rate--> ----\ Packet /---------------
\ Loss /<----Convergence
Convergence------->\ / Event Transition
Recovery Transition \ /
\_____/<------100% Packet Loss
X-axis = Time
Y-axis = Forwarding Rate
Figure 5. Convergence Graph
3.2.6 Packet Sampling Interval
Selection of the Packet Sampling Interval on the Test Equipment
impacts the measured Rate-Derived Convergence Time. Packet
Sampling Interval time is that is too large exaggerates the
slope of the Convergence Event Transition and Convergence
Recovery Transition producing a larger than the actual Rate-Derived
Convergence Time. This impact is greater as routers achieve
millisecond convergence times. The recommended value for the
Packet Sampling Interval is 100 millisecond. It is possible to
have commercially available test equipment with a minimum
configurable Packet Sampling Interval of 1 second.
3.2.7 Interface Types
All test cases in this methodology document may be executed with
any interface type. SONET is recommended and specifically
mentioned in the procedures because it can be configured to have
no or negligible affect on the measured convergence time.
Ethernet (10Mb, 100Mb, 1Gb, and 10Gb) is not preferred since
broadcast media are unable to detect loss of host and rely upon
IGP Hellos to detect session loss.
3.3 Reporting Format
For each test case, it is recommended that the following reporting
format be completed:
IGP Data Plane Route Convergence
Parameter Units
--------- -----
IGP (ISIS or OSPF)
Interface Type (GigE, POS, ATM, etc.)
Packet Size bytes
IGP Routes number of IGP routes
Packet Sampling Interval seconds or milliseconds
IGP Timer Values
SONET Failure Indication Delay seconds or milliseconds
IGP Hello Timer seconds or milliseconds
IGP Dead-Interval seconds or milliseconds
LSA Generation Delay seconds or milliseconds
LSA Flood Packet Pacing seconds or milliseconds
LSA Retransmission Packet Pacing seconds or milliseconds
SPF Delay seconds or milliseconds
Results
Rate-Derived Convergence Time seconds or milliseconds
Loss-Derived Convergence Time seconds or milliseconds
Restoration Convergence Time seconds or milliseconds
4. Test Cases
4.1 Convergence Due to Link Failure
4.1.1 Convergence Due to Local Interface Failure
Objective Objective
To obtain the IGP Route Convergence due to a Local SONET Link To obtain the IGP Route Convergence due to a local link
failure event. failure event at the DUT's Local Interface.
Procedure Procedure
1. Advertise IGP routes from Tester to DUT on Ingress Traffic 1. Advertise matching IGP routes from Tester to DUT on
Path. Preferred Egress Interface [2] and Next-Best Egress Interface
2. Advertise matching IGP routes from Tester to DUT on [2] using the topology shown in Figure 1. Set the cost of the
Preferred Egress Path and Backup Egress Path. Set the cost routes so that the Preferred Egress Interface is the preferred
of the routes so that the IGP routes along the Preferred next-hop.
Egress Path is the preferred next-hop. 2. Send traffic at maximum forwarding rate to destinations
3. Send traffic at maximum forwarding rate to destinations matching all IGP routes from Tester to DUT on Ingress Interface
matching all IGP routes from Tester to DUT on Ingress Traffic [2].
Path. 3. Verify traffic routed over Preferred Egress Interface.
4. Verify traffic routed over Preferred Egress Path. 4. Remove SONET on DUT's Local Interface [2] by performing an
5. Remove SONET on Tester Interface connected to Preferred Egress administrative shutdown of the interface.
Path. 5. Measure Rate-Derived Convergence Time [2] and Loss-Derived
6. Measure Peak-to-Peak Convergence Time [2] as DUT detects Convergence Time [2] as DUT detects the link down event and
the link down event and converges all IGP routes and converges all IGP routes and traffic over the Next-Best Egress
traffic over the Backup Egress Path. Interface.
6. Restore SONET on DUT's Local Interface by administratively
enabling the interface.
7. Measure Restoration Convergence Time [2] as DUT detects the link
up event and converges all IGP routes and traffic back to the
Preferred Egress Interface.
Results
The measured IGP Convergence time is influenced by the Local
SONET indication, SPF delay, SPF Holdtime, SPF Execution
Time, Tree Build Time, and Hardware Update Time.
IGP Data Plane Route Convergence IGP Data Plane Route Convergence
4.1.2 Convergence Due to Neighbor Interface Failure
Objective
To obtain the IGP Route Convergence due to a local link
failure event at the Tester's Neighbor Interface.
Procedure
1. Advertise matching IGP routes from Tester to DUT on
Preferred Egress Interface [2] and Next-Best Egress Interface
[2] using the topology shown in Figure 1. Set the cost of
the routes so that the Preferred Egress Interface is the
preferred next-hop.
2. Send traffic at maximum forwarding rate to destinations
matching all IGP routes from Tester to DUT on Ingress
Interface [2].
3. Verify traffic routed over Preferred Egress Interface.
4. Remove SONET on Tester's Neighbor Interface [2] connected to
DUT' s Preferred Egress Interface.
5. Measure Rate-Derived Convergence Time [2] and Loss-Derived
Convergence Time [2] as DUT detects the link down event and
converges all IGP routes and traffic over the Next-Best
Egress Interface.
6. Restore SONET on Tester's Neighbor Interface connected to
DUT's Preferred Egress Interface.
7. Measure Restoration Convergence Time [2] as DUT detects the
link up event and converges all IGP routes and traffic back to
the Preferred Egress Interface.
Results Results
The measured IGP Convergence time is influenced by the Local The measured IGP Convergence time is influenced by the Local
SONET indication, SPF delay, SPF Holdtime, SPF Execution SONET indication, SPF delay, SPF Holdtime, SPF Execution
Time, Tree Build Time, and Hardware Update Time. Time, Tree Build Time, and Hardware Update Time.
4.1.2 Convergence Due to PPP Session Failure 4.1.3 Convergence Due to Remote Interface Failure
Objective Objective
To obtain the IGP Route Convergence due to a Remote
Interface failure event.
Procedure
1. Advertise matching IGP routes from Tester to SUT on
Preferred Egress Interface [2] and Next-Best Egress Interface
[2] using the topology shown in Figure 2. Set the cost of the
routes so that the Preferred Egress Interface is the preferred
next-hop. NOTE: All routers in the SUT must be the same model
and identically configured.
2. Send traffic at maximum forwarding rate to destinations
matching all IGP routes from Tester to DUT on Ingress Interface
[2].
3. Verify traffic is routed over Preferred Egress Interface.
4. Remove SONET on Tester's Neighbor Interface [2] connected to
SUT' s Preferred Egress Interface.
IGP Data Plane Route Convergence
5. Measure Rate-Derived Convergence Time [2] and Loss-Derived
Convergence Time [2] as SUT detects the link down event and
converges all IGP routes and traffic over the Next-Best
Egress Interface.
6. Restore SONET on Tester's Neighbor Interface connected to
SUT's Preferred Egress Interface.
7. Measure Restoration Convergence Time [2] as SUT detects the
link up event and converges all IGP routes and traffic over
the Preferred Egress Interface.
Results
The measured IGP Convergence time is influenced by the
SONET failure indication, LSA/LSP Flood Packet Pacing,
LSA/LSP Retransmission Packet Pacing, LSA/LSP Generation
time, SPF delay, SPF Holdtime, SPF Execution Time, Tree
Build Time, and Hardware Update Time. The additional
convergence time contributed by LSP Propagation can be
obtained by subtracting the Rate-Derived Convergence Time
measured in 4.1.2 (Convergence Due to Neighbor Interface
Failure) from the Rate-Derived Convergence Time measured in
this test case.
4.2 Convergence Due to PPP Session Failure
Objective
To obtain the IGP Route Convergence due to a Local PPP Session To obtain the IGP Route Convergence due to a Local PPP Session
failure event. failure event.
Procedure Procedure
1. Advertise IGP routes from Tester to DUT on Ingress Traffic 1. Advertise matching IGP routes from Tester to DUT on
Path. Preferred Egress Interface [2] and Next-Best Egress Interface
2. Advertise matching IGP routes from Tester to DUT on [2] using the topology shown in Figure 1. Set the cost of
Preferred Egress Path and Backup Egress Path. Set the cost the routes so that the IGP routes along the Preferred Egress
of the routes so that the IGP routes along the Preferred Interface is the preferred next-hop.
Egress Path is the preferred next-hop. 2. Send traffic at maximum forwarding rate to destinations
3. Send traffic at maximum forwarding rate to destinations
matching all IGP routes from Tester to DUT on Ingress matching all IGP routes from Tester to DUT on Ingress
Traffic Path. Interface [2].
4. Verify traffic routed over Preferred Egress Path. 3. Verify traffic routed over Preferred Egress Interface.
5. Remove PPP session from Tester Interface connected to 4. Remove PPP session from Tester's Neighbor Interface [2]
Preferred Egress Path. connected to Preferred Egress Interface.
6. Measure Peak-to-Peak Convergence Time as DUT detects the 5. Measure Rate-Derived Convergence Time [2] and Loss-Derived
PPP session down event and converges all IGP routes and Convergence Time [2] as DUT detects the PPP session down event
traffic over the Backup Egress Path. and converges all IGP routes and traffic over the Next-Best
Egress Interface.
6. Restore PPP session on DUT's Preferred Egress Interface.
7. Measure Restoration Convergence Time [2] as DUT detects the
session up event and converges all IGP routes and traffic over
the Preferred Egress Interface.
Results Results
The measured IGP Convergence time is influenced by the Local The measured IGP Convergence time is influenced by the PPP
PPP failure indication, SPF delay, SPF Holdtime, SPF Execution failure indication, SPF delay, SPF Holdtime, SPF Execution
Time, Tree Build Time, and Hardware Update Time. Time, Tree Build Time, and Hardware Update Time.
4.1.3 Convergence Due to IGP Adjacency Failure IGP Data Plane Route Convergence
4.3 Convergence Due to IGP Adjacency Failure
Objective Objective
To obtain the IGP Route Convergence due to a Local IGP Adjacency To obtain the IGP Route Convergence due to a Local IGP Adjacency
failure event. failure event.
Procedure Procedure
1. Advertise IGP routes from Tester to DUT on Ingress Traffic 1. Advertise matching IGP routes from Tester to DUT on
Path. Preferred Egress Interface [2] and Next-Best Egress Interface
2. Advertise matching IGP routes from Tester to DUT on [2] using the topology shown in Figure 1. Set the cost of
Preferred Egress Path and Backup Egress Path. Set the cost the routes so that the Preferred Egress Interface is the
of the routes so that the IGP routes along the Preferred preferred next-hop.
Egress Path is the preferred next-hop. 2. Send traffic at maximum forwarding rate to destinations
3. Send traffic at maximum forwarding rate to destinations
matching all IGP routes from Tester to DUT on Ingress matching all IGP routes from Tester to DUT on Ingress
Traffic Path. Interface [2].
4. Verify traffic routed over Preferred Egress Path. 3. Verify traffic routed over Preferred Egress Interface.
5. Remove IGP adjacency from Tester interface connected to 4. Remove IGP adjacency from Tester's Neighbor Interface [2]
Preferred Egress Path. connected to Preferred Egress Interface.
5. Measure Rate-Derived Convergence Time [2] and Loss-Derived
IGP Data Plane Route Convergence Convergence Time [2] as DUT detects the IGP session failure
event and converges all IGP routes and traffic over the
6. Measure Peak-to-Peak Convergence Time as DUT detects the Next-Best Egress Interface.
IGP session failure event and converges all IGP routes and 6. Restore IGP session on DUT's Preferred Egress Interface.
traffic over the Backup Egress Path. 7. Measure Restoration Convergence Time [2] as DUT detects the
session up event and converges all IGP routes and traffic over
the Preferred Egress Interface.
Results Results
The measured IGP Convergence time is influenced by the IGP The measured IGP Convergence time is influenced by the IGP
Hello Interval, IGP Dead Interval, SPF delay, SPF Holdtime, Hello Interval, IGP Dead Interval, SPF delay, SPF Holdtime,
SPF Execution Time, Tree Build Time, and Hardware Update SPF Execution Time, Tree Build Time, and Hardware Update
Time. Time.
4.1.4 Convergence Due to Route Withdrawal 4.4 Convergence Due to Route Withdrawal
Objective Objective
To obtain the IGP Route Convergence due to Route Withdrawal. To obtain the IGP Route Convergence due to Route Withdrawal.
Procedure Procedure
1. Advertise IGP routes from Tester to DUT on Ingress Traffic 1. Advertise matching IGP routes from Tester to DUT on
Path. Preferred Egress Interface [2] and Next-Best Egress Interface
2. Advertise matching IGP routes from Tester to DUT on [2] using the topology shown in Figure 1. Set the cost of
Preferred Egress Path and Backup Egress Path. Set the cost the routes so that the Preferred Egress Interface is the
of the routes so that the IGP routes along the Preferred preferred next-hop.
Egress Path is the preferred next-hop. 2. Send traffic at maximum forwarding rate to destinations
3. Send traffic at maximum forwarding rate to destinations
matching all IGP routes from Tester to DUT on Ingress matching all IGP routes from Tester to DUT on Ingress
Traffic Path. Interface [2].
4. Verify traffic routed over Preferred Egress Path. 3. Verify traffic routed over Preferred Egress Interface.
5. Tester withdraws all IGP routes from DUT's Local Interface 4. Tester withdraws all IGP routes from DUT's Local Interface
on Preferred Egress Path. on Preferred Egress Interface.
6. Measure Peak-to-Peak Convergence Time as DUT processes the
route withdrawal event and converges all IGP routes and IGP Data Plane Route Convergence
traffic over the Backup Egress Path.
5. Measure Rate-Derived Convergence Time [2] and Loss-Derived
Convergence Time [2] as DUT processes the route withdrawal
event and converges all IGP routes and traffic over the
Next-Best Egress Interface.
6. Re-advertise IGP routes to DUT's Preferred Egress Interface.
7. Measure Restoration Convergence Time [2] as DUT converges all
IGP routes and traffic over the Preferred Egress Interface.
Results Results
The measured IGP Convergence time is the SPF Processing and FIB The measured IGP Convergence time is the SPF Processing and FIB
Update time as influenced by the SPF delay, SPF Holdtime, Update time as influenced by the SPF delay, SPF Holdtime,
SPF Execution Time, Tree Build Time, and Hardware Update Time. SPF Execution Time, Tree Build Time, and Hardware Update Time.
4.1.5 Convergence Due to Cost Change 4.5 Convergence Due to Cost Change
Objective Objective
To obtain the IGP Route Convergence due to route cost change. To obtain the IGP Route Convergence due to route cost change.
Procedure Procedure
1. Advertise IGP routes from Tester to DUT on Ingress Traffic 1. Advertise matching IGP routes from Tester to DUT on
Path. Preferred Egress Interface [2] and Next-Best Egress Interface
2. Advertise matching IGP routes from Tester to DUT on [2] using the topology shown in Figure 1. Set the cost of
Preferred Egress Path and Backup Egress Path. Set the cost the routes so that the Preferred Egress Interface is the
of the routes so that the IGP routes along the Preferred preferred next-hop.
Egress Path is the preferred next-hop. 2. Send traffic at maximum forwarding rate to destinations
3. Send traffic at maximum forwarding rate to destinations
matching all IGP routes from Tester to DUT on Ingress matching all IGP routes from Tester to DUT on Ingress
Traffic Path. Interface [2].
3. Verify traffic routed over Preferred Egress Interface.
IGP Data Plane Route Convergence 4. Tester increases cost for all IGP routes at DUT's Preferred
Egress Interface so that the Next-Best Egress Inerface
4. Verify traffic routed over Preferred Egress Path. has lower cost and becomes preferred path.
5. Tester increases cost for all IGP routes at DUT's Local 5. Measure Rate-Derived Convergence Time [2] and Loss-Derived
Interface on Preferred Egress Path so that Backup Egress Convergence Time [2] as DUT detects the cost change event
Path have lower cost and becomes preferred path. and converges all IGP routes and traffic over the Next-Best
6. Measure Reroute Convergence Time [2] as DUT detects the Egress Interface.
cost change event and converges all IGP routes and 6. Re-advertise IGP routes to DUT's Preferred Egress Interface
traffic over the Backup Egress Path. with original lower cost metric.
7. Measure Restoration Convergence Time [2] as DUT converges all
IGP routes and traffic over the Preferred Egress Interface.
Results Results
The measured IGP Convergence time is the SPF Processing and FIB There should be no measured packet loss for this case.
Update time as influenced by the SPF delay, SPF Holdtime,
SPF Execution Time, Tree Build Time, and Hardware Update Time.
There should be no packet loss for this case.
4.2 Remote Events
The test cases in this section use the test topology shown in
Figure 2.
4.2.1 Convergence Due to Remote SONET Link Failure 4.6 Convergence Due to ECMP Member Interface Failure
Objective Objective
To obtain the IGP Route Convergence due to a Remote To obtain the IGP Route Convergence due to a local link
SONET Link failure event. failure event of an ECMP Member.
IGP Data Plane Route Convergence
Procedure Procedure
1. Advertise IGP routes from Tester to DUT on Ingress Traffic 1. Configure ECMP Set as shown in Figure 3.
Path.
2. Advertise matching IGP routes from Tester to DUT on 2. Advertise matching IGP routes from Tester to DUT on
Preferred Egress Path and Backup Egress Path. Set the cost each ECMP member.
of the routes so that the IGP routes along the Preferred
Egress Path is the preferred next-hop.
3. Send traffic at maximum forwarding rate to destinations 3. Send traffic at maximum forwarding rate to destinations
matching all IGP routes from Tester to DUT on Ingress matching all IGP routes from Tester to DUT on Ingress
Traffic Path. Interface [2].
4. Verify traffic routed over Preferred Egress Path. 4. Verify traffic routed over all members of ECMP Set.
5. Remove SONET on Neighbor Interface connected to 5. Remove SONET on Tester's Neighbor Interface [2] connected to
Preferred Egress Path. one of the DUT's ECMP member interfaces.
6. Measure Peak-to-Peak Convergence time as DUT detects the 6. Measure Rate-Derived Convergence Time [2] and Loss-Derived
link down event and converges all IGP routes and traffic Convergence Time [2] as DUT detects the link down event and
over the Backup Egress Path. converges all IGP routes and traffic over the other ECMP
members.
7. Restore SONET on Tester's Neighbor Interface connected to
DUT's ECMP member interface.
8. Measure Restoration Convergence Time [2] as DUT detects the
link up event and converges IGP routes and some distribution
of traffic over the restored ECMP member.
Results Results
The measured IGP Convergence time is influenced by the The measured IGP Convergence time is influenced by the Local
SONET failure indication, LSA/LSP Flood Packet Pacing, SONET indication, Tree Build Time, and Hardware Update Time.
LSA/LSP Retransmission Packet Pacing, LSA/LSP Generation
time, SPF delay, SPF Holdtime, SPF Execution Time, Tree
Build Time, and Hardware Update Time.
IGP Data Plane Route Convergence
5. Measuring Convergence Times
5.1 Measuring Full Convergence Time
Figure 3 shows a graph model of Convergence Time as measured
from the data plane. Refer to [2] for definitions of the terms
used. IGP Route Convergence Time is the amount of time for the
Forwarding Rate to begin its downward slope upon occurrence of
a network event and then fully recover to the Maximum
Forwarding Rate.
Forwarding Rate versus Time
Time=Recovery Time=Network Event Time = 0sec
Maximum ^ ^ ^
Forwarding Rate--> ----\ /-----------
\ /<----Route Convergence
Route Convergence------->\ / Event Slope
Recovery Slope \_______/<------100% Packet Loss
X-axis = Time
Y-axis = Forwarding Rate
Figure 3. Convergence Graph
Maximum forwarding rate at a fixed packet size without packet
loss is required for accurate measurement. The test duration
must be greater than the convergence time. Full Convergence
Time is obtained directly from the graph in Figure 3 using
equation 1.
(eq 1) Convergence Time(Full)=Time(Recovery)-Time(Network Event).
Given a known constant rate of offered load in units packet per
second (pps), the Average Convergence Time can be obtained
using equation 2 or equation 3.
(eq 2) Convergence Time(Average)=Number Packets Lost/pps(Offered) 4.7 Convergence Due to Parallel Link Interface Failure
Objective
To obtain the IGP Route Convergence due to a local link
failure event for a Member of a Parallel Link.
(eq 3) Convergence Time(Average)= (Number Packets Offered - Procedure
Number of Packets Received)/pps(Offered) 1. Configure Parallel Link as shown in Figure 4.
2. Advertise matching IGP routes from Tester to DUT on
each Parallel Link member.
3. Send traffic at maximum forwarding rate to destinations
matching all IGP routes from Tester to DUT on Ingress
Interface [2].
4. Verify traffic routed over all members of Parallel Link.
5. Remove SONET on Tester's Neighbor Interface [2] connected to
one of the DUT's Parallel Link member interfaces.
6. Measure Rate-Derived Convergence Time [2] and Loss-Derived
Convergence Time [2] as DUT detects the link down event and
converges all IGP routes and traffic over the other
Parallel Link members.
7. Restore SONET on Tester's Neighbor Interface connected to
DUT's Parallel Link member interface.
8. Measure Restoration Convergence Time [2] as DUT detects the
link up event and converges IGP routes and some distribution
of traffic over the restored Parallel Link member.
As discussed in [1], Full Convergence Time is the more accurate Results
measurement. Average Convergence Time does not account for the The measured IGP Convergence time is influenced by the Local
angle of the Route Convergence Recovery Slope, so Full Convergence SONET indication, Tree Build Time, and Hardware Update Time.
Time > Average Convergence Time. Ideally, the Recovery Slope has
no angle so that it is vertical and Average Convergence Time =
Full Convergence Time.
5.2 Measuring Impact of Components for Convergence
The factors for IGP Route Convergence Time are provided in [1].
The results of the test cases in section 4 above can be used to
calculate the impact each factor has on the Convergence
IGP Data Plane Route Convergence IGP Data Plane Route Convergence
results, as follow: 5. Security Considerations
SPF Processing and FIB Update time = Result (4.1.4)
SONET failure indication time = Result(4.1.1) - Result(4.1.4)
PPP failure indication time = Result(4.1.2) - Result(4.1.4)
IGP failure indication time = Result(4.1.3) - Result(4.1.4)
IGP Advertisement time = Result(4.1.1) - Result(4.2.1)
6. Security Considerations
Documents of this type do not directly effect the security of Documents of this type do not directly affect the security of
the Internet or of corporate networks as long as benchmarking the Internet or corporate networks as long as benchmarking
is not performed on devices or systems connected to operating is not performed on devices or systems connected to operating
networks. networks.
7. Acknowledgements 6. References
Thanks to Jayant Kulkarni for doing as most Test Engineers
do - working beyond the call of duty to help advance
technology. Especially thanks to the many Network Engineers
and Network Architects at the Service Providers who are always
eager to discuss Route Convergence.
8. References
[1] Poretsky, S., "Benchmarking Applicability for IGP [1] Poretsky, S., "Benchmarking Applicability for IGP
Convergence", draft-ietf-bmwg-igp-dataplane-conv-app-00, work Convergence", draft-ietf-bmwg-igp-dataplane-conv-app-01, work
in progress, June 2003. in progress, October 2003.
[2] Poretsky, S., "Benchmarking Terminology for IGP Convergence", [2] Poretsky, S., Imhoff, B., "Benchmarking Terminology for IGP Convergence", draft-ietf-bmwg-igp-dataplane-conv-term-01, work
draft-ietf-bmwg-igp-dataplane-conv-term-00, work in progress, in progress, October 2003.
June 2003.
[3] Callon, R., "Use of OSI IS-IS for Routing in TCP/IP and Dual [3] Callon, R., "Use of OSI IS-IS for Routing in TCP/IP and Dual
Environments", RFC 1195, December 1990. Environments", RFC 1195, December 1990.
[4] Moy, J., "OSPF Version 2", RFC 2328, IETF, April 1998. [4] Moy, J., "OSPF Version 2", RFC 2328, IETF, April 1998.
9. Author's Address 7. Author's Address
Scott Poretsky Scott Poretsky
Avici Systems, Inc. Quarry Technologies
101 Billerica Avenue 8 New England Executive Park
N. Billerica, MA 01862 Burlington, MA 01803
USA USA
Phone: + 1 978 964 2287 Phone: + 1 781 395 5090
EMail: sporetsky@avici.com EMail: sporetsky@quarrytech.com
IGP Data Plane Route Convergence
Brent Imhoff Brent Imhoff
WilTel Communications WilTel Communications
3180 Rider Trail South 3180 Rider Trail South
Bridgeton, MO 63045 Bridgeton, MO 63045
USA USA
Phone: +1 314 595 6853 Phone: +1 314 595 6853
EMail: brent.imhoff@wcg.com EMail: brent.imhoff@wcg.com
10. Full Copyright Statement 8. Full Copyright Statement
Copyright (C) The Internet Society (1998). All Rights Copyright (C) The Internet Society (1998). All Rights
Reserved. Reserved.
This document and translations of it may be copied and This document and translations of it may be copied and
furnished to others, and derivative works that comment on or furnished to others, and derivative works that comment on or
otherwise explain it or assist in its implementation may be otherwise explain it or assist in its implementation may be
IGP Data Plane Route Convergence
prepared, copied, published and distributed, in whole or in prepared, copied, published and distributed, in whole or in
part, without restriction of any kind, provided that the above part, without restriction of any kind, provided that the above
copyright notice and this paragraph are included on all such copyright notice and this paragraph are included on all such
copies and derivative works. However, this document itself may copies and derivative works. However, this document itself may
not be modified in any way, such as by removing the copyright not be modified in any way, such as by removing the copyright
notice or references to the Internet Society or other Internet notice or references to the Internet Society or other Internet
organizations, except as needed for the purpose of developing organizations, except as needed for the purpose of developing
Internet standards in which case the procedures for copyrights Internet standards in which case the procedures for copyrights
defined in the Internet Standards process must be followed, or defined in the Internet Standards process must be followed, or
as required to translate it into languages other than English. as required to translate it into languages other than English.
 End of changes. 

This html diff was produced by rfcdiff 1.23, available from http://www.levkowetz.com/ietf/tools/rfcdiff/