 1/draftietfbmwgigpdataplaneconvterm15.txt 20081015 23:12:26.000000000 +0200
+++ 2/draftietfbmwgigpdataplaneconvterm16.txt 20081015 23:12:26.000000000 +0200
@@ 1,22 +1,21 @@
Network Working Group S. Poretsky
 Internet Draft NextPoint Networks
 Expires: August 2008
+ Internet Draft Allot Communications
Intended Status: Informational Brent Imhoff
Juniper Networks
 February 25, 2008
+ October 15, 2008
Terminology for Benchmarking
LinkState IGP Data Plane Route Convergence

+
Intellectual Property Rights (IPR) statement:
By submitting this InternetDraft, each author represents that any
applicable patent or other IPR claims of which he or she is aware
have been or will be disclosed, and any of which he or she becomes
aware will be disclosed, in accordance with Section 6 of BCP 79.
Status of this Memo
InternetDrafts are working documents of the Internet Engineering
@@ 49,46 +48,46 @@
Table of Contents
1. Introduction .................................................2
2. Existing definitions .........................................3
3. Term definitions..............................................4
3.1 Convergence Event.........................................4
3.2 Route Convergence.........................................4
3.3 Full Convergence..........................................5
3.4 Network Convergence.......................................5
3.5 RouteSpecific Convergence................................6
 3.6 Packet Loss...............................................6
+ 3.6 Packet Loss...............................................7
3.7 Convergence Packet Loss...................................7
 3.8 Convergence Event Instant.................................7
+ 3.8 Convergence Event Instant.................................8
3.9 Convergence Recovery Instant..............................8
 3.10 First Route Convergence Instant..........................8
+ 3.10 First Route Convergence Instant..........................9
3.11 Convergence Event Transition.............................9
 3.12 Convergence Recovery Transition..........................9
+ 3.12 Convergence Recovery Transition..........................10
3.13 RateDerived Convergence Time............................10
 3.14 LossDerived Convergence Time............................10
+ 3.14 LossDerived Convergence Time............................11
3.15 RouteSpecific Convergence Time..........................12
3.16 Sustained Convergence Validation Time....................13
 3.17 First Route Convergence Time.............................13
 3.18 Reversion Convergence Time...............................14
 3.19 Packet Sampling Interval.................................14
 3.20 Local Interface..........................................15
 3.21 Neighbor Interface.......................................15
 3.22 Remote Interface.........................................15
 3.23 Preferred Egress Interface...............................16
 3.24 NextBest Egress Interface...............................16
 3.25 Stale Forwarding.........................................17
 3.26 Nested Convergence Events................................17
 4. IANA Considerations...........................................18
 5. Security Considerations.......................................18
 6. Acknowledgements..............................................18
 7. References....................................................18
 8. Author's Address..............................................19
+ 3.17 First Route Convergence Time.............................14
+ 3.18 Reversion Convergence Time...............................15
+ 3.19 Packet Sampling Interval.................................15
+ 3.20 Local Interface..........................................16
+ 3.21 Neighbor Interface.......................................16
+ 3.22 Remote Interface.........................................17
+ 3.23 Preferred Egress Interface...............................17
+ 3.24 NextBest Egress Interface...............................17
+ 3.25 Stale Forwarding.........................................18
+ 3.26 Nested Convergence Events................................18
+ 4. IANA Considerations...........................................19
+ 5. Security Considerations.......................................19
+ 6. Acknowledgements..............................................19
+ 7. References....................................................19
+ 8. Author's Address..............................................20
1. Introduction
This draft describes the terminology for benchmarking Interior
Gateway Protocol (IGP) Route Convergence. The motivation and
applicability for this benchmarking is provided in [Po07a]. The
methodology to be used for this benchmarking is described in [Po07m].
The methodology and terminology to be used for benchmarking Route
Convergence can be applied to any linkstate IGP such as ISIS [Ca90]
and OSPF [Mo98]. The data plane is measured to obtain blackbox
(externally observable) convergence benchmarking metrics. The
@@ 100,20 +99,21 @@
An example of Route Convergence as observed and measured from the
data plane is shown in Figure 1. The graph in Figure 1 shows
Forwarding Rate versus Time. Time 0 on the Xaxis is on the far
right of the graph. The Offered Load to the ingress interface of
the DUT SHOULD equal the measured maximum Throughput [Ba99][Ma98]
of the DUT and the Forwarding Rate [Ma98] is measured at the egress
interfaces of the DUT. The components of the graph and the metrics
are defined in the Term Definitions section.
+ Full Convergence,
Convergence Convergence
Recovery Event
Instant Instant Time = 0sec
Forwarding Rate = ^ ^ ^ Offered Load =
Offered Load > \ Packet / <Max Throughput
\ Loss /<Convergence
Convergence>\ / Event Transition
Recovery Transition \ /
\_____/<Maximum Packet Loss
^
@@ 209,23 +209,29 @@
equal to the offered load.
Discussion:
When benchmarking convergence, it is useful to measure
the time to converge an entire FIB. For example,
a Convergence Event can be produced for an OSPF table of
5000 routes so that the time to converge routes 1 through
5000 is measured. Completion of Full Convergence is externally
observable from the data plane when the Throughput of the data
plane traffic on the NextBest Egress Interface equals the
 offered load. Full Convergence may or may not be sustained over
 time. The Sustained Convergence Validation Time MUST be
 applied.
+ offered load.
+
+ Full convergence MAY be measured using RateDerived Convergence
+ Time (3.13) or calculated using LossDerived Convergence Time
+ (3.14). When performing RouteSpecific Convergence (3.5)
+ measurements, Full Convergence may be obtained by measuring the
+ maximum Route Specific Convergence Time (3.15). Full
+ Convergence may or may not be sustained over time. The
+ Sustained Convergence Validation Time (3.16) MUST be applied.
Measurement Units:
N/A
Issues:
None
See Also:
Network Convergence
Route Convergence
@@ 238,30 +244,31 @@
distributed FIBs, in all routers throughout the network.
Discussion:
Network Convergence requires completion of all Route
Convergence operations for all routers in the network following
a Convergence Event. Completion of Network Convergence can be
observed by recovery of System Under Test (SUT) Throughput to
equal the offered load, with no Stale Forwarding, and no
Blenders [Ca01][Ci03].
+ LinkState IGP Data Plane Route Convergence
+
Measurement Units:
N/A
Issues:
None
See Also:
Route Convergence
Stale Forwarding
 LinkState IGP Data Plane Route Convergence
3.5 RouteSpecific Convergence
Definition:
Route Convergence for one or more specific route entries in
the FIB in which recovery from the Convergence Event is
indicated by dataplane traffic for a flow [Po06] matching that
route entry(ies) being routed to the NextBest Egress Interface.
Discussion:
@@ 276,43 +283,43 @@
Measurement Units:
N/A
Issues:
None
See Also:
Full Convergence
Route Convergence
Convergence Event
+ LinkState IGP Data Plane Route Convergence
3.6 Packet Loss
Definition:
The number of packets that should have been forwarded
by a DUT under a constant offered load that were
not forwarded due to lack of resources.
Discussion:
 Packet Lss is a modified version of the term "Frame Loss Rate"
+ Packet Loss is a modified version of the term "Frame Loss Rate"
as defined in [Ba91]. The term "Frame Loss" is intended for
Ethernet Frames while "Packet Loss" is intended for IP packets.
Packet Loss can be measured as a reduction in forwarded traffic
from the Throughput [Ba91] of the DUT.
Measurement units:
Number of offered packets that are not forwarded.
Issues: None
See Also:
Convergence Packet Loss
 LinkState IGP Data Plane Route Convergence
3.7 Convergence Packet Loss
Definition:
The number of packets lost due to a Convergence Event
until Full Convergence completes.
Discussion:
Convergence Packet Loss includes packets that were lost and
packets that were delayed due to buffering. The Convergence
@@ 324,20 +331,21 @@
Measurement Units:
number of packets
Issues: None
See Also:
Packet Loss
Route Convergence
Convergence Event
Packet Sampling Interval
+ LinkState IGP Data Plane Route Convergence
3.8 Convergence Event Instant
Definition:
The time instant that a Convergence Event becomes observable in
the data plane.
Discussion:
Convergence Event Instant is observable from the data
plane as the precise time that the device under test begins
@@ 347,21 +355,20 @@
hh:mm:ss:nnn:uuu,
where 'nnn' is milliseconds and 'uuu' is microseconds.
Issues:
None
See Also:
Convergence Event
Convergence Packet Loss
Convergence Recovery Instant
 LinkState IGP Data Plane Route Convergence
3.9 Convergence Recovery Instant
Definition:
The time instant that Full Convergence completion is
measured and then maintained for an interval of duration
equal to the Sustained Convergence Validation Time.
Discussion:
Convergence Recovery Instant is measurable from the data
@@ 372,23 +379,23 @@
hh:mm:ss:nnn:uuu,
where 'nnn' is milliseconds and 'uuu' is microseconds.
Issues:
None
See Also:
Sustained Convergence Validation Time
Convergence Packet Loss
Convergence Event Instant
+ LinkState IGP Data Plane Route Convergence
3.10 First Route Convergence Instant

Definition:
The time instant a first route entry has converged
following a Convergence Event, as observed by receipt of
the first packet from the NextBest Egress Interface.
Discussion:
The First Route Convergence Instant is an indication that the
process to achieve Full Convergence has begun. Any route may
be the first to converge for First Route Convergence Instant.
Measurement on the dataplane enables the First Route
@@ 388,34 +395,31 @@
the first packet from the NextBest Egress Interface.
Discussion:
The First Route Convergence Instant is an indication that the
process to achieve Full Convergence has begun. Any route may
be the first to converge for First Route Convergence Instant.
Measurement on the dataplane enables the First Route
Convergence Instant to be observed without any whitebox
information from the DUT.
 Measurement Units:
 N/A
+ Measurement Units: N/A
Issues:
None
See Also:
Route Convergence
Full Convergence
Stale Forwarding
 LinkState IGP Data Plane Route Convergence
3.11 Convergence Event Transition

Definition:
A time interval observed following a Convergence Event in which
Throughput gradually reduces to a minimum value.
Discussion:
The Convergence Event Transition is best observed for Full
Convergence. The egress packet rate observed during a
Convergence Event Transition may not decrease linearly and may
not decrease to zero. Both the offered load and the Packet
Sampling Interval influence the observations of the Convergence
@@ 428,20 +432,21 @@
Measurement Units:
seconds
Issues:
None
See Also:
Convergence Event
Full Convergence
Packet Sampling Interval
+ LinkState IGP Data Plane Route Convergence
3.12 Convergence Recovery Transition
Definition:
The characteristic of the DUT in which Throughput gradually
increases to equal the offered load.
Discussion:
The Convergence Recovery Transition is best observed for
Full Convergence. The egress packet rate observed during
@@ 452,21 +457,20 @@
"Packet Sampling Interval".
Measurement Units:
seconds
Issues: None
See Also:
Full Convergence
Packet Sampling Interval
 LinkState IGP Data Plane Route Convergence
3.13 RateDerived Convergence Time
Definition:
The amount of time for Convergence Packet Loss to persist upon
occurrence of a Convergence Event until Full Convergence has
completed.
RateDerived Convergence Time can be measured as the time
difference from the Convergence Event Instant to the
@@ 480,20 +484,22 @@
RateDerived Convergence Time SHOULD be measured at the maximum
Throughput of the DUT. At least one packet per route in the FIB
for all routes in the FIB MUST be offered to the DUT within the
Packet Sampling Interval.
Failure to achieve Full Convergence results in a RateDerived
Convergence Time benchmark of infinity. It is RECOMMENDED that
the RateDerived Convergence Time be measured when benchmarking
Full Convergence.
+ LinkState IGP Data Plane Route Convergence
+
Measurement Units:
seconds
Issues: None
See Also:
Convergence Packet Loss
Convergence Recovery Instant
Convergence Event Instant
Full Convergence
@@ 504,57 +510,57 @@
The amount of time it takes for Full Convergence to be
completed as calculated from the amount of Convergence
Packet Loss. LossDerived Convergence Time can be
calculated from Convergence Packet Loss as shown with
Equation 2.
Equation 2 
LossDerived Convergence Time =
Convergence Packets Loss / Offered Load
where units are packets / packets/second = seconds
 LinkState IGP Data Plane Route Convergence
Discussion:
Optimally, the Convergence Event Transition and Convergence
Recovery Transition are instantaneous so that the
RateDerived Convergence Time = LossDerived Convergence Time.
However, router implementations are less than ideal.
LossDerived Convergence Time gives a better than
actual result when converging many routes simultaneously
because it ignores the Convergence Recovery Transition.
RateDerived Convergence Time takes the Convergence Recovery
Transition into account. Equation 2 calculates the average
convergence time over all routes to which packets have been
sent. Since this average convergence time is in general
smaller than the maximum convergence time over all routes,
LossDerived Convergence Time is not the preferred metric to
indicate Full Convergence completion. For this reason the
RECOMMENDED benchmark metric for Full Convergence is the
RateDerived Convergence Time.
+ LinkState IGP Data Plane Route Convergence
+
Guidelines for reporting LossDerived Convergence Time are
provided in [Po07m].
Measurement Units:
seconds
Issues:
None
See Also:
Convergence Event
Convergence Packet Loss
RateDerived Convergence Time
RouteSpecific Convergence
Convergence Event Transition
Convergence Recovery Transition
 LinkState IGP Data Plane Route Convergence
3.15 RouteSpecific Convergence Time
Definition:
The amount of time it takes for RouteSpecific Convergence to
be completed as calculated from the amount of Convergence
Packet Loss per flow.
RouteSpecific Convergence Time can be calculated from
Convergence Packet Loss as shown with Equation 3.
@@ 574,43 +580,71 @@
RouteSpecific Convergence, Convergence Packet Loss is measured
for specific flow(s) and Equation 3 is applied for each flow.
Each flow has a single destination address matching a different
route entry. The fastest measurable convergence time is equal
to the time between two consecutive packets of a flow offered
by the Tester.
The RouteSpecific Convergence Time benchmarks enable minimum,
maximum, average, and median convergence time measurements to be
reported by comparing the results for the different route
+ LinkState IGP Data Plane Route Convergence
+
entries. It also enables benchmarking of convergence time when
configuring a priority value for route entry(ies). Since
multiple RouteSpecific Convergence Times can be measured it is
possible to have an array of results. The format for reporting
RouteSpecific Convergence Time is provided in [Po07m].
The RouteSpecific Convergence Time MAY be used to benchmark
Full Convergence when used in combination with many flows
 matching every FIB entry.
+ matching every FIB entry. In this case
+ Full Convergence = max(RouteSpecific Convergence Time).
Measurement Units:
seconds
Issues:
None
See Also:
Convergence Event
Convergence Packet Loss
RouteSpecific Convergence
+
+ 3.16 Sustained Convergence Validation Time
+
+ Definition:
+ The amount of time for which the completion of Full
+ Convergence is maintained without additional packet loss.
+
+ Discussion:
+ The purpose of the Sustained Convergence Validation Time is to
+ produce Convergence benchmarks protected against fluctuation
+ in Throughput after the completion of Full Convergence is
+ observed. The RECOMMENDED Sustained Convergence Validation
+ Time to be used is 5 seconds. The BMWG selected 5 seconds
+ based upon RFC 2544 [Ba99] which recommends waiting 2 seconds
+ for residual frames to arrive and 5 seconds for DUT
+ restabilization.
+
+ Measurement Units:
+ seconds
+
+ Issues: None
+
+ See Also:
+ Full Convergence
+ Convergence Recovery Instant
LinkState IGP Data Plane Route Convergence
 3.16 First Route Convergence Time
+ 3.17 First Route Convergence Time
Definition:
The amount of time for Convergence Packet Loss until the
convergence of a first route entry on the NextBest Egress
Interface, as indicated by the First Route Convergence
Instant.
Discussion:
The First Route Convergence Time benchmarking metric can be
measured when benchmarking either Full Convergence or
@@ 639,102 +673,76 @@
benchmark of infinity.
Measurement Units:
seconds
Issues: None
See Also:
Convergence Packet Loss
First Route Convergence Instant

 3.17 Sustained Convergence Validation Time

 Definition:
 The amount of time for which the completion of Full
 Convergence is maintained without additional packet loss.

LinkState IGP Data Plane Route Convergence
 Discussion:
 The purpose of the Sustained Convergence Validation Time is to
 produce Convergence benchmarks protected against fluctuation
 in Throughput after the completion of Full Convergence is
 observed. The RECOMMENDED Sustained Convergence Validation
 Time to be used is 5 seconds.

 Measurement Units:
 seconds

 Issues: None

 See Also:
 Full Convergence
 Convergence Recovery Instant

3.18 Reversion Convergence Time
Definition:
The amount of time for the DUT to complete Full Convergence
to the Preferred Egress Interface, instead of the NextBest
Egress Interface, upon recovery from a Convergence Event.
Discussion:
Reversion Convergence Time is the amount of time for Full
 COnvergence to the original egress interface. This is
+ Convergence to the original egress interface. This is
achieved by recovering from the Convergence Event, such as
restoring the failed link. Reversion Convergence Time is
measured using the RateDerived Convergence Time calculation
technique, as provided in Equation 1. It is possible to have
the Reversion Convergence Time differ from the RateDerived
Convergence Time.
 Measurement Units:
 seconds
+ Measurement Units: seconds
Issues: None
See Also:
Preferred Egress Interface
Convergence Event
RateDerived Convergence Time
3.19 Packet Sampling Interval
Definition:
The interval at which the tester (test equipment) polls to make
measurements for arriving packet flows.
Discussion:
 At least one packet per route in the FIB
 for all routes in the FIB MUST be offered to the DUT within the
 Packet Sampling Interval. Metrics measured at the Packet
 Sampling Interval MUST include Forwarding Rate and Convergence
 Packet Loss.

 LinkState IGP Data Plane Route Convergence

 Measurement Units:
 seconds
+ At least one packet per route in the FIB for all routes in the
+ FIB MUST be offered to the DUT within the Packet Sampling
+ Interval. Metrics measured at the Packet Sampling Interval
+ MUST include Forwarding Rate and Convergence Packet Loss.
 Issues:
Packet Sampling Interval can influence the Convergence Graph.
This is particularly true when implementations complete Full
Convergence in less than the Packet Sampling Interval. The
Convergence Event Transition and Convergence Recovery Transition
can become exaggerated when the Packet Sampling Interval is too
long. This will produce a larger than actual RateDerived
Convergence Time. The recommended value for configuration of
the Packet Sampling Interval is provided in [Po07m].
+ Measurement Units: seconds
+
+ Issues: None
+
See Also:
Convergence Packet Loss
Convergence Event Transition
Convergence Recovery Transition
+ LinkState IGP Data Plane Route Convergence
3.20 Local Interface
Definition:
An interface on the DUT.
Discussion:
A failure of the Local Interface indicates that the failure
occurred directly on the DUT.
@@ 754,28 +762,28 @@
The interface on the neighbor router or tester that is
directly linked to the DUT's Local Interface.
Discussion:
A failure of a Neighbor Interface indicates that a
failure occurred on a neighbor router's interface that
directly links the neighbor router to the DUT.
Measurement Units:
N/A
 LinkState IGP Data Plane Route Convergence
Issues:
None
See Also:
Local Interface
Remote Interface
+ LinkState IGP Data Plane Route Convergence
3.22 Remote Interface
Definition:
An interface on a neighboring router that is not directly
connected to any interface on the DUT.
Discussion:
A failure of a Remote Interface indicates that the failure
occurred on a neighbor router's interface that is not
@@ 809,35 +817,37 @@
See Also:
NextBest Egress Interface
3.24 NextBest Egress Interface
Definition:
The outbound interface from the DUT for traffic routed to the
secondbest nexthop. It is the same media type and link speed
as the Preferred Egress Interface
 LinkState IGP Data Plane Route Convergence
Discussion:
The NextBest Egress Interface becomes the egress interface
after a Convergence Event.
+ LinkState IGP Data Plane Route Convergence
+
Measurement Units:
N/A
Issues: None
See Also:
Preferred Egress Interface
3.25 Stale Forwarding
+
Definition:
Forwarding of traffic to route entries that no longer exist
or to route entries with nexthops that are no longer preferred.
Discussion:
Stale Forwarding can be caused by a Convergence Event and can
manifest as a "blackhole" or microloop that produces packet
loss. Stale Forwarding can exist until Network Convergence is
completed. Stale Forwarding cannot be observed with a single
DUT.
@@ 907,48 +918,48 @@
[Mo06] Morton, A., et al, "Packet Reordering Metrics", RFC 4737,
November 2006.
[Po06] Poretsky, S., et al., "Terminology for Benchmarking
Networklayer Traffic Control Mechanisms", RFC 4689,
November 2006.
[Po07a] Poretsky, S., "Benchmarking Applicability for LinkState
IGP Data Plane Route Convergence",
 draftietfbmwgigpdataplaneconvapp15, work in progress,
 February 2008.

 LinkState IGP Data Plane Route Convergence
+ draftietfbmwgigpdataplaneconvapp16, work in progress,
+ October 2008.
[Po07m] Poretsky, S. and Imhoff, B., "Benchmarking Methodology for
LinkState IGP Data Plane Route Convergence",
 draftietfbmwgigpdataplaneconvmeth15, work in progress,
 February 2008.
+ draftietfbmwgigpdataplaneconvmeth16, work in progress,
+ October 2008.
+
+ LinkState IGP Data Plane Route Convergence
7.2 Informative References
[Ca01] S. Casner, C. Alaettinoglu, and C. Kuan, "A FineGrained View
of High Performance Networking", NANOG 22, June 2001.
[Ci03] L. Ciavattone, A. Morton, and G. Ramachandran, "Standardized
Active Measurements on a Tier 1 IP Backbone", IEEE
Communications Magazine, pp9097, May 2003.
8. Author's Address
Scott Poretsky
 NextPoint Networks
 3 Federal Street
 Billerica, MA 01821
+ Allot Communications
+ 67 South Bedford Street, Suite 400
+ Burlington, MA 01803
USA
 Phone: + 1 508 439 9008
 EMail: sporetsky@nextpointnetworks.com
+ Phone: + 1 508 309 2179
+ Email: sporetsky@allot.com
Brent Imhoff
Juniper Networks
1194 North Mathilda Ave
Sunnyvale, CA 94089
USA
Phone: + 1 314 378 2571
EMail: bimhoff@planetspork.com
Full Copyright Statement