draft-ietf-bmwg-igp-dataplane-conv-app-00.txt   draft-ietf-bmwg-igp-dataplane-conv-app-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
June 2003 October 2003
Benchmarking Applicability for Benchmarking Applicability for
IGP Data Plane Route Convergence IGP Data Plane Route Convergence
<draft-ietf-bmwg-igp-dataplane-conv-app-00.txt> <draft-ietf-bmwg-igp-dataplane-conv-app-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 39 skipping to change at page 1, line 39
progress." progress."
The list of current Internet-Drafts can be accessed at The list of current Internet-Drafts can be accessed at
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.
ABSTRACT ABSTRACT
This draft describes the applicability of IGP Route Convergence This draft describes the applicability of IGP Route Convergence
benchmarking methodology [1] and IGP Route Convergence bechmarking benchmarking methodology [1] and IGP Route Convergence benchmarking
terminology [2]. The methodology and terminology is to be used terminology [2]. The methodology and terminology is to be used
for benchmarking route convergence and can be applied to any for benchmarking route convergence and can be applied to any
link-state IGP such as ISIS [3] and OSPF [4]. The data plane is link-state IGP such as ISIS [3] and OSPF [4]. The data plane is
measured to obtain the convergence benchmarking metrics described measured to obtain the convergence benchmarking metrics described
in [1]. in [1].
Table of Contents Table of Contents
1. Introduction ...............................................2 1. Introduction ...............................................2
2. Existing definitions .......................................2 2. Existing definitions .......................................2
3. Factors for IGP Route Convergence Time......................2 3. Factors for IGP Route Convergence Time......................2
skipping to change at page 2, line 16 skipping to change at page 2, line 16
9. Author's Address............................................5 9. Author's Address............................................5
10. Full Copyright Statement...................................5 10. Full Copyright Statement...................................5
1. Introduction 1. Introduction
IGP Convergence is a critical performance parameter. Customers IGP Convergence is a critical performance parameter. Customers
of Service Providers use packet loss due to IGP Convergence as a of Service Providers use packet loss due to IGP Convergence as a
key metric of their network service quality. Service Providers key metric of their network service quality. Service Providers
use IGP Convergence time as a key metric of router design and use IGP Convergence time as a key metric of router design and
architecture. Fast network convergence can be optimally achieved architecture. Fast network convergence can be optimally achieved
through deployment of fast converging routers. The fundamental through deployment of fast converging routers. The fundamental
basis by which customers of service providers benchmark convergence basis by which network users and operators benchmark convergence
is packet loss, which is an externally observable event having is packet loss, which is an externally observable event having
direct impact on their application performance. IGP Route direct impact on their application performance.
Convergence is a Direct Measure of Quality (DMOQ) when benchmarking
the data plane. For this reason it is important to develop a standard IGP Route Convergence is a Direct Measure of Quality (DMOQ) when
router benchmarking methodology and terminology for measuring IGP benchmarking the data plane. For this reason it is important to
convergence that uses the data plane as described in [1] and [2]. develop a standard router benchmarking methodology and terminology
This document describes all of the factors that influence a for measuring IGP convergence that uses the data plane as described
convergence measurement and how a purely black box test can be in [1] and [2]. This document describes all of the factors that
designed to account for all of these factors. This enables accurate influence a convergence measurement and how a purely black box test
benchmarking and evaluation for route convergence time. can be designed to account for all of these factors. This enables
accurate benchmarking and evaluation for route convergence time.
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. Factors for IGP Route Convergence Time 3. Factors for IGP Route Convergence Time
There are four major categories of factors for the measured Router There are four major categories of factors contributing to the
IGP Convergence Time, as described in [5], [6], [7], [8] and [9]. measured Router IGP Convergence Time. As discussed in [5], [6],
These are Event Detection, SPF Processing, IGP Advertisement, and [7], [8] and [9], these categories are Event Detection, SPF
FIB Update. Each of these factors has numerous components to Processing, IGP Advertisement, and FIB Update. These have numerous
influence the convergence time. These are listed as follow: components that influence the convergence time. These are listed
as follow:
-Event Detection- -Event Detection-
SONET failure indication time SONET failure indication time
PPP failure indication time PPP failure indication time
IGP Hello Dead Interval IGP Hello Dead Interval
-SPF Processing- -SPF Processing-
SPF Delay Time SPF Delay Time
SPF Hold time SPF Hold time
SPF Execution time SPF Execution time
IGP Data Plane Route Convergence IGP Data Plane Route Convergence
-IGP Advertisement- -IGP Advertisement-
LSA/LSP Flood Packet Pacing LSA/LSP Flood Packet Pacing
LSA/LSP Retransmission Packet Pacing LSA/LSP Retransmission Packet Pacing
LSA/LSP Generation time LSA/LSP Generation time
-FIB Update- -FIB Update-
Tree Build time Tree Build time
Hardware Update time Hardware Update time
Each of the factors listed above will have a varying amount of The contribution of each of these factors listed above will vary
influence on the convergence result with each router vendors' with each router vendors' architecture and IGP implementation.
architecture and IGP implementation. It is necessary to design a It is therefore necessary to design a convergence test that
convergence test that considers not just one or a few of these considers all of these components, not just one or a few of these
components, but instead all of these components. The additional components. The additional benefit of designing a test for all
benefit of designing a test for all components is that it enables components is that it enables black-box testing in which knowledge
black-box testing in which knowledge of the routers' internal of the routers' internal implementations is not required. It is
implementations is not required. It is then possible to make then possible to make valid use of the convergence benchmarking
valid use of the benchmarking metrics when comparing routers from metrics when comparing routers from different vendors.
different vendors.
4. Network Events that Cause Convergence 4. Network Events that Cause Convergence
There are different types of network events that can cause IGP There are different types of network events that can cause IGP
convergence. These network events are administrative link convergence. These network events are administrative link
removal, unplanned link removal, and route change such as removal, unplanned link failure, line card failure, and route
withdrawal, flap, next-hop change, and cost change. When changes such as withdrawal, flap, next-hop change, and cost change.
benchmarking a router it is important to measure the convergence When benchmarking a router it is important to measure the
time for local and remote occurrence of these network events. convergence time for local and remote occurrence of these network
The convergence time measured will vary whether the network event events. The convergence time measured will vary whether the network
occurred locally or remotely due to varying combinations of event occurred locally or remotely due to varying combinations of
factors listed in the previous sections. This behavior makes it factors listed in the previous sections. This behavior makes it
possible to design purely black-box tests that isolate possible to design purely black-box tests that isolate
measurements for each of the components of convergence time. measurements for each of the components of convergence time.
5. Use of Data Plane for IGP Route Convergence Benchmarking 5. Use of Data Plane for IGP Route Convergence Benchmarking
Customers of service providers use packet loss as the metric for Customers of service providers use packet loss as the metric to
convergence time. Packet loss is an externally observable event calculate convergence time. Packet loss is an externally observable
having direct impact on customers' application performance. event having direct impact on customers' application performance.
For this reason it is important to develop a standard router For this reason it is important to develop a standard router
benchmarking methodology and terminology that is a Direct Measure benchmarking methodology and terminology that is a Direct Measure
of Quality (DMOQ)for measuring IGP convergence. Such a of Quality (DMOQ)for measuring IGP convergence. Such a
methodology uses the data plane as described in [1] and [2]. methodology uses the data plane as described in [1] and [2].
An additional benefit of using packet loss for calculation of An additional benefit of using packet loss for calculation of
IGP Route Convergence time is that it enables black-box tests to IGP Route Convergence time is that it enables black-box tests to
be designed. Data traffic can be offered at line-rate to the be designed. Data traffic can be offered to the
device under test (DUT), an emulated network event can be forced device under test (DUT), an emulated network event can be forced
to occur, and packet loss can be externally to occur, and packet loss can be externally measured to calculate
observed to measure the convergence time. Knowledge of the DUT the convergence time. Knowledge of the DUT architecture and IGP
architecture and IGP implementation is not required. There is no implementation is not required. There is no need to rely on the
need to rely on the DUT to produce the test results. There is no DUT to produce the test results. There is no need to build
need to build intrusive test harnasses for the DUT. intrusive test harnesses for the DUT.
IGP Data Plane Route Convergence IGP Data Plane Route Convergence
Use of data traffic and measurement of packet loss on the data Use of data traffic and measurement of packet loss on the data
plane also enables Route Convergence methodology test cases that plane also enables Route Convergence methodology test cases that
consider the time for the Route Controller to update the FIB on consider the time for the Route Controller to update the FIB on
the forwarding engine of the hardware. A router is not fully the forwarding engine of the hardware. A router is not fully
converged until all components are updated and traffic is converged until all components are updated and traffic is
rerouted along the correct path. As long as there is packet rerouted to the correct egress interface. As long as there is
loss, routes have not converged. It is possible to send diverse packet loss, routes have not converged. It is possible to send
traffic flows to destinations matching every route in the FIB diverse traffic flows to destinations matching every route in the
so that the time it takes for the router to converge an entire FIB so that the time it takes for the router to converge an entire
route table can be benchmarked. route table can be benchmarked.
6. Security Considerations 6. Security Considerations
Documents of this type do not directly effect the security of Documents of this type do not directly effect the security of
the Internet or of corporate networks as long as benchmarking the Internet or of 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 7. Acknowledgements
Thanks to Curtis Villamizar for sharing so much of his Thanks to Curtis Villamizar for sharing so much of his
knowledge and experience through the years. Also, special knowledge and experience through the years. Also, special
thanks to the many Network Engineers and Network Architects thanks to the many Network Engineers and Network Architects
at the Service Providers who are always eager to discuss at the Service Providers who are always eager to discuss
Route Convergence. Route Convergence.
8. References 8. References
[1] Poretsky, S., "Benchmarking Methodology for IGP Data Plane [1] Poretsky, S., "Benchmarking Methodology for IGP Data Plane
Route Convergence", draft-ietf-bmwg-igp-dataplane-conv-meth-00, Route Convergence", draft-ietf-bmwg-igp-dataplane-conv-meth-01,
work in progress, June 2003. work in progress, October 2003.
[2] Poretsky, S., "Benchmarking Terminology for IGP Data Plane [2] Poretsky, S., "Benchmarking Terminology for IGP Data Plane
Route Convergence", draft-ietf-bmwg-igp-dataplane-conv-term-00, Route Convergence", draft-ietf-bmwg-igp-dataplane-conv-term-01,
work in progress, June 2003. work in progress, October 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.
[5] Villamizar, C., "Convergence and Restoration Techniques for [5] Villamizar, C., "Convergence and Restoration Techniques for
ISP Interior Routing", NANOG 25, June 2002. ISP Interior Routing", NANOG 25, October 2002.
[6] Katz, D., "Why are we Scared of SPF? IGP Scaling and [6] Katz, D., "Why are we Scared of SPF? IGP Scaling and
Stability", NANOG 25, June 2002. Stability", NANOG 25, October 2002.
[7] Filsfils, C., "Deploying Tight-SLA Services on an Internet [7] Filsfils, C., "Deploying Tight-SLA Services on an Internet
Backbone: ISIS Fast Convergence and Differentiated Services Backbone: ISIS Fast Convergence and Differentiated Services
Design (tutorial)", NANOG 25, June 2002. Design (tutorial)", NANOG 25, October 2002.
IGP Data Plane Route Convergence IGP Data Plane Route Convergence
[8] Alaettinoglu, C. and Casner, S., "ISIS Routing on the Qwest [8] Alaettinoglu, C. and Casner, S., "ISIS Routing on the Qwest
Backbone: a Recipe for Subsecond ISIS Convergence", NANOG 24, Backbone: a Recipe for Subsecond ISIS Convergence", NANOG 24,
June 2002. October 2002.
[9] Alaettinoglu, C., Jacobson, V., and Yu, H., "Towards [9] Alaettinoglu, C., Jacobson, V., and Yu, H., "Towards
Millisecond IGP Convergence", NANOG 20, October 2000. Millisecond IGP Convergence", NANOG 20, October 2000.
9. Author's Address 9. 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
10. Full Copyright Statement 10. 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
prepared, copied, published and distributed, in whole or in prepared, copied, published and distributed, in whole or in
 End of changes. 

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