Network Working Group R. Papneja
Internet Draft Isocore
Expires: March 2007 S.Vapiwala
J.Karthik
Cisco Systems
S. Poretsky
Reef Point
S. Rao
Qwest Communications
Jean-Louis Le Roux
France Telecom
October 06
Methodology for benchmarking MPLS Protection mechanisms
<draft-ietf-bmwg-protection-meth-00.txt>
Status of this Memo
By submitting this Internet-Draft, 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.
This document may only be posted in an Internet-Draft.
Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF), its areas, and its working groups. Note that
other groups may also distribute working documents as Internet-
Drafts.
Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference
material or to cite them other than as "work in progress."
The list of current Internet-Drafts can be accessed at
http://www.ietf.org/ietf/1id-abstracts.txt
The list of Internet-Draft Shadow Directories can be accessed at
http://www.ietf.org/shadow.html
Papneja,Vapiwala,Karthik, Expires April 13, 2007 [Page 1]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
Abstract
This draft provides the methodology for benchmarking MPLS Protection
mechanisms especially the failover time of local protection (MPLS Fast
Reroute as defined in RFC-4090). The failover to a backup tunnel could
happen at the headend of the primary tunnel or a midpoint and the backup
could offer link or node protection. It becomes vital to benchmark the
failover time for all the cases and combinations. The failover time
could also greatly differ based on the design and implementation and by
factors like the number of prefixes carried by the tunnel, the routing
protocols that installed these prefixes (IGP, BGP...), the number of
primary tunnels affected by the event that caused the failover, number
of primary tunnels the backup protects and type of failure, the physical
media type on which the failover occurs etc. All the required
benchmarking criteria and benchmarking topology required for measuring
failover time of local protection is described Conventions used in this
document
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in RFC 2119 [RFC2119].
Table of Contents
1. Introduction...................................................3
2. Existing definitions...........................................6
3. Test Considerations............................................6
3.1. Failover Events...........................................6
3.2. Failure Detection [TERMID]................................7
3.3. Use of Data Traffic for MPLS Protection Benchmarking......7
3.4. LSP and Route Scaling.....................................8
3.5. Selection of IGP..........................................8
3.6. Reversion [TERMID]........................................8
3.7. Traffic generation........................................9
3.8. Motivation for topologies.................................9
4. Test Setup.....................................................9
4.1. Link Protection with 1 hop primary (from PLR) and 1 hop
backup........................................................10
TE tunnels....................................................10
4.2. Link Protection with 1 hop primary (from PLR) and 2 hop
backup TE tunnels.............................................11
4.3. Link Protection with 2+ hop (from PLR) primary and 1 hop
backup TE tunnels.............................................11
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 2]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
4.4. Link Protection with 2+ hop (from PLR) primary and 2 hop
backup TE tunnels.............................................12
4.5. Node Protection with 2 hop primary (from PLR) and 1 hop
backup TE tunnels.............................................13
4.6. Node Protection with 2 hop primary (from PLR) and 2 hop
backup TE tunnels.............................................13
4.7. Node Protection with 3+ hop primary (from PLR) and 1 hop
backup TE tunnels.............................................14
4.8. Node Protection with 3+ hop primary (from PLR) and 2 hop
backup TE tunnels.............................................15
4.9. Baseline MPLS Forwarding Performance Test Topology.......15
5. Test Methodology..............................................16
5.1. Headend as PLR with link failure.........................16
5.2. Mid-Point as PLR with link failure.......................17
5.3. Headend as PLR with Node failure.........................18
5.4. Mid-Point as PLR with Node failure.......................20
5.5. Baseline MPLS Forwarding Performance Test Cases..........21
5.5.1. DUT Throughput as Ingress...........................21
5.5.2. DUT Latency as Ingress..............................21
5.5.3. DUT Throughput as Egress............................22
5.5.4. DUT Latency as Egress...............................22
5.5.5. DUT Throughput as Mid-Point.........................23
5.5.6. DUT Latency as Mid-Point............................23
6. Reporting Format..............................................24
7. Security Considerations.......................................25
8. Acknowledgements..............................................25
9. References....................................................25
9.1. Normative References.....................................25
9.2. Informative References...................................26
10. Author's Address.............................................26
Appendix A: Fast Reroute Scalability Table.......................29
1. Introduction
A link or a node failure could occur at the headend or the mid point
node of a given primary tunnel. The time it takes to failover to the
backup tunnel is a key measurement since it directly affects the traffic
carried over the tunnel. The failover could occur at the headend or the
midpoint of a primary tunnel and the time it takes to failover depends
on a variety of factors like the type of physical media, method of FRR
solution (detour vs facility), number of primary tunnels, number of
prefixes carried over the tunnel etc. Given all this service providers
certainly like to see a methodology to measure the failover time under
all possible conditions.
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 3]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
The following sections describe all the different topologies and
scenarios that should be used and considered to effectively benchmark
the failover time. The failure triggers, procedures, scaling
considerations and reporting format of the results are discussed as
well.
In order to benchmark failover time, data plane traffic is used as
mentioned in [IGP-METH] since traffic loss is measured in a black-box
test and is a widely accepted way to measure convergence.
Important point to be noted when benchmarking the failover time is that
depending on whether PHP is happening (whether or not implicit null is
advertised by the tail-end), and on the number of hops of primary and
backup tunnel, we could have different situations where the packets
switched over to the backup tunnel may have one, more or 0 labels.
All the benchmarking cases mentioned in this document could apply to
facility backup as well as local protection enabled in the detour mode.
The test cases and the procedures described here should completely
benchmark the failover time of a device under test in all possible
scenarios and configuration.
The additional scenarios defined in this document, are in addition to
those considered in [FRR-METH]. All the cases enlisted in this document
could be verified in a single topology that is similar to this.
---------------------------
| ------------|---------------
| | | |
| | | |
-------- -------- -------- -------- --------
TG-| R1 |-----| R2 |----| R3 | | R4 | | R5 |-TA
| |-----| |----| |----| |---| |
-------- -------- -------- -------- --------
| | | |
| | | |
| -------- | |
---------| R6 |-------- |
| |--------------------
--------
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 4]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
Fig.1: Fast Reroute Topology.
In figure 1, TG & TA are Traffic Generator & Analyzer respectively.
A tester is set outside the node as it sends and receives IP traffic
along the working Path, run protocol emulations simulating real world
peering scenarios. The tester MUST record the number of lost packets,
duplicate packet count, reordered packet count, departure time, and
arrival time so that the metrics of Failover Time, Additive Latency, and
Reversion Time can be measured. The tester may be a single device or a
test system.
Two or more failures are considered correlated if those failures occur
more or less simultaneously. Correlated failures are often expected
where two or more logical resources, such as layer-2 links, rely on a
common physical resource, such as common transport. TDM and WDM provide
multiplexing at layer-2 and layer-1 that are often the cause of
correlated failures. Where such correlations are known, such as knowing
that two logical links share a common fiber segment, the expectation of
a common failure can be compensated for by specifying Shared Risk Link
Groups [RFC-4090]. Not all correlated failures are anticipated in
advance of their occurrence. Failures due to natural disasters or due
to certain man-made disasters or mistakes are the most notable causes.
Failures of this type occur many times a year and generally a quite
spectacular failure occurs every few years.
There are two factors impacting service availability. One is the
frequency of failure. The other is the duration of failure. FRR
improves availability by minimizing the duration of the most common
failures. Unexpected correlated failures are less common. Some routers
recover much more quickly than others and therefore benchmarking this
type of failure may also be useful. Benchmarking of unexpected
correlated failures should include measurement of restoration with and
without the availability of IP fallback. The use BGP free core may be
growing, making the latter case an important test case. This document
focuses on FRR failover benchmarking with MPLS TE. Benchmarking of
unexpected correlated failures is out of scope but may be covered by a
later document.
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 5]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
2. Existing definitions
For the sake of clarity and continuity this RFC adopts the template
for definitions set out in Section 2 of RFC 1242. Definitions are
indexed and grouped together in sections for ease of reference.
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in
this document are to be interpreted as described in RFC 2119.
The reader is assumed to be familiar with the commonly used MPLS
terminology, some of which is defined in [MPLS-RSVP], [MPLS-RSVP-TE],
and [MPLS-FRR-EXT].
3. Test Considerations
This section discusses the fundamentals of MPLS Protection testing:
-The types of network events that causes failover
-Indications for failover
-the use of data traffic
-Traffic generation
-LSP Scaling
-Reversion of LSP
-IGP Selection
3.1. Failover Events
Triggers for failover to a backup tunnel are link and node failures
seen downstream of the PLR as follows.
Link failure events
- Shutdown interface on PLR side with POS Alarm
- Shutdown interface on remote side with POS Alarm
- Shutdown interface on PLR side with RSVP hello
- Shutdown interface on remote side with RSVP hello
- Shutdown interface on PLR side with BFD
- Shutdown interface on remote side with BFD
- Fiber Pull on PLR side (Both TX & RX or just the Tx)
- Fiber Pull on remote side (Both TX & RX or just the Rx)
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 6]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
- OIR on PLR side
- OIR on remote side
- Sub-interface failure (shutting down of a VLAN)
- Shut parent interface bearing multiple sub-interfaces
Node failure events
A Reload is a graceful shutdown or a power failure. We refer to Crash
as a software failure or an assert.
- Reload protected Node, when RSVP Hello are enable
- Crash Protected Node, when RSVP Hello are enable
- Reload Protected Node, when BFD is enable
- Crash Protected Node, when BFD is enable
3.2. Failure Detection [TERMID]
Local failures can be detected via SONET/SDH failure with directly
connected LSR. Failure indication may vary with the type of alarm -
LOS, AIS, or RDI. Failures on Ethernet technology links such as
Gigabit Ethernet rely upon Layer 3 signaling indication for failure.
Different MPLS protection mechanisms and different implementations
use different failure indications such as RSVP hellos, BFD etc.
Ethernet technologies such as Gigabit Ethernet rely upon layer 3
failure indication mechanisms since there is no Layer 2 failure
indication mechanism. The failure detection time may not always be
negligible and it could impact the overall failover time.
The test procedures in this document can be used against a local
failure as well as against a remote failure to account for
completeness of benchmarking and to evaluate failover performance
independent of the implemented signaling indication mechanism.
3.3. Use of Data Traffic for MPLS Protection Benchmarking
Customers of service providers use packet loss as the metric for
failover time. Packet loss is an externally observable event having
direct impact on customers' application performance. MPLS protection
mechanism is expected to minimize the packet loss in the event of a
failure. For this reason it is important to develop a standard router
benchmarking methodology for measuring MPLS protection that uses
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 7]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
packet loss as a metric. At a known rate for forwarding, packet loss
can be measured and used to calculate the Failover time. Measurement
of control plane signaling to establish backup paths is not enough
to verify failover. Failover is best determined when packets are
actually traversing the backup path.
An additional benefit of using packet loss for calculation of
Failover time is that it enables black-box tests to be designed. Data
traffic can be offered at line-rate to the device under test (DUT),
an emulated network event as described above can be forced to occur,
and packet loss can be externally measured to calculate the
convergence time. Knowledge of DUT architecture is not required.
There is no need to rely on the understanding of the implementation
details of the DUT to get the required test results.
In addition, this methodology will consider the errored packets and
duplicate packets that could have been generated during the failover
process. In extreme cases, where measurement of errored and duplicate
packets is difficult, these packets could be attributed to lost
packets.
3.4. LSP and Route Scaling
Failover time performance may vary with the number of established
primary and backup LSPs and routes learned. However the procedure
outlined here may be used for any number of LSPs, L, and number of
routes protected by PLR, R. L and R must be recorded.
3.5. Selection of IGP
The underlying IGP could be ISIS-TE or OSPF-TE for the methodology
proposed here.
3.6. Reversion [TERMID]
Fast Reroute provides a method to return or restore a backup path to
original primary LSP upon recovery from the failure. This is referred
to as Reversion, which can be implemented as Global Reversion or
Local Reversion. In all test cases listed here Reversion should not
produce any packet loss, out of order or duplicate packets. Each of
the test cases in this methodology document provides a step to verify
that there is no packet loss.
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 8]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
3.7. Traffic generation
It is suggested that there be one or more traffic streams as long as
there is a steady and constant rate of flow for all the streams. In
order to monitor the DUT performance for recovery times a set of
route prefixes should be advertised before traffic is sent. The
traffic should be configured towards these routes.
A typical example would be configuring the traffic generator to send
the traffic to the first, middle and last of the advertised routes.
(First, middle and last could be decided by the numerically smallest,
median and the largest respectively of the advertised prefix).
Generating traffic to all of the prefixes reachable by the protected
tunnel (probably in a Round-Robin fashion, where the traffic is
destined to all the prefixes but one prefix at a time in a cyclic
manner) is not recommended. The reason why traffic generation is not
recommended in a Round-Robin fashion to all the prefixes, one at a
time is that if there are many prefixes reachable through the LSP the
time interval between 2 packets destined to one prefix may be
significantly high and may be comparable with the failover time being
measured which does not aid in getting an accurate failover
measurement.
3.8. Motivation for topologies
Given that the label stack is dependent on the following 3 entities
it is recommended that the benchmarking of failover time be performed
on all the 8 topologies enlisted in section 4
- Type of protection (Link Vs Node)
- # of remaining hops of the primary tunnel from the PLR
- # of remaining hops of the backup tunnel from the PLR
4. Test Setup
Topologies to be used for benchmarking the failover time:
This section proposes a set of topologies that covers the scenarios
for local protection. All of these 8 topologies shown (figure 2-
figure 9) can be mapped to the master FRR topology shown in figure 1.
Topologies shown in section 4.1 to 4.8 refer to the network
topologies required to benchmark failover time when DUT is configured
as a PLR either in headend or midpoint role. The number of labels
listed below are all w.r.t the PLR.
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 9]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
The label stacks shown below each figure in section 4.1 to 4.9
considers the scenario when PHP is enabled.
In the following network topologies,
HE is Head-End, TE is Tail-End, MID is Mid point, MP is Merge Point,
PLR is Point of Local Repair, PRI is Primary and BKP denotes Backup
Node
4.1. Link Protection with 1 hop primary (from PLR) and 1 hop backup
TE tunnels
------- -------- PRI --------
| R1 | | R2 | | R3 |
TG-| HE |--| MID |----| TE |-TA
| | | PLR |----| |
------- -------- BKP --------
Figure 2: Represents the setup for section 4.1
Traffic No of Labels No of labels after
before failure failure
IP TRAFFIC (P-P) 0 0
Layer3 VPN (PE-PE) 1 1
Layer3 VPN (PE-P) 2 2
Layer2 VC (PE-PE) 1 1
Layer2 VC (PE-P) 2 2
Mid-point LSPs 0 0
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 10]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
4.2. Link Protection with 1 hop primary (from PLR) and 2 hop backup TE
tunnels
------- -------- --------
| R1 | | R2 | | R3 |
TG-| HE | | MID |PRI | TE |-TA
| |----| PLR |----| |
------- -------- --------
|BKP |
| -------- |
| | R6 | |
|----| BKP |----|
| MID |
--------
Figure 3: Representing setup for section 4.2
Traffic No of Labels No of labels
before failure after failure
IP TRAFFIC (P-P) 0 1
Layer3 VPN (PE-PE) 1 2
Layer3 VPN (PE-P) 2 3
Layer2 VC (PE-PE) 1 2
Layer2 VC (PE-P) 2 3
Mid-point LSPs 0 1
4.3. Link Protection with 2+ hop (from PLR) primary and 1 hop backup TE
tunnels
-------- -------- -------- --------
| R1 | | R2 |PRI | R3 |PRI | R4 |
TG-| HE |----| MID |----| MID |------| TE |-TA
| | | PLR |----| | | |
-------- -------- BKP -------- --------
Figure 4: Representing setup for section 4.3
Traffic No of Labels No of labels
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 11]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
before failure after failure
IP TRAFFIC (P-P) 1 1
Layer3 VPN (PE-PE) 2 2
Layer3 VPN (PE-P) 3 3
Layer2 VC (PE-PE) 2 2
Layer2 VC (PE-P) 3 3
Mid-point LSPs 1 1
4.4. Link Protection with 2+ hop (from PLR) primary and 2 hop backup TE
tunnels
-------- -------- PRI -------- PRI --------
| R1 | | R2 | | R3 | | R4 |
TG-| HE |----| MID |----| MID |------| TE |-TA
| | | PLR | | | | |
-------- -------- -------- --------
BKP| |
| -------- |
| | R6 | |
---| BKP |-
| MID |
--------
Figure 5: Representing the setup for section 4.4
Traffic No of Labels No of labels
before failure after failure
IP TRAFFIC (P-P) 1 2
Layer3 VPN (PE-PE) 2 3
Layer3 VPN (PE-P) 3 4
Layer2 VC (PE-PE) 2 3
Layer2 VC (PE-P) 3 4
Mid-point LSPs 1 2
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 12]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
4.5. Node Protection with 2 hop primary (from PLR) and 1 hop backup TE
tunnels
-------- -------- -------- --------
| R1 | | R2 |PRI | R3 | PRI | R4 |
TG-| HE |----| MID |----| MID |------| TE |-TA
| | | PLR | | | | |
-------- -------- -------- --------
|BKP |
-----------------------------
Figure 6: Representing the setup for section 4.5
Traffic No of Labels No of labels
before failure after failure
IP TRAFFIC (P-P) 1 0
Layer3 VPN (PE-PE) 2 1
Layer3 VPN (PE-P) 3 2
Layer2 VC (PE-PE) 2 1
Layer2 VC (PE-P) 3 2
Mid-point LSPs 1 0
4.6. Node Protection with 2 hop primary (from PLR) and 2 hop backup TE
tunnels
-------- -------- -------- --------
| R1 | | R2 | | R3 | | R4 |
TG-| HE | | MID |PRI | MID |PRI | TE |-TA
| |----| PLR |----| |----| |
-------- -------- -------- --------
| |
BKP| -------- |
| | R6 | |
---------| BKP |---------
| MID |
--------
Figure 7: Representing setup for section 4.6
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 13]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
Traffic No of Labels No of labels
before failure after failure
IP TRAFFIC (P-P) 1 1
Layer3 VPN (PE-PE) 2 2
Layer3 VPN (PE-P) 3 3
Layer2 VC (PE-PE) 2 2
Layer2 VC (PE-P) 3 3
Mid-point LSPs 1 1
4.7. Node Protection with 3+ hop primary (from PLR) and 1 hop backup TE
tunnels
-------- -------- PRI -------- PRI -------- PRI --------
| R1 | | R2 | | R3 | | R4 | | R5 |
TG-| HE |--| MID |---| MID |---| MP |---| TE |-TA
| | | PLR | | | | | | |
-------- -------- -------- -------- --------
BKP| |
--------------------------
Figure 8: Representing setup for section 4.7
Traffic No of Labels No of labels
before failure after failure
IP TRAFFIC (P-P) 1 1
Layer3 VPN (PE-PE) 2 2
Layer3 VPN (PE-P) 3 3
Layer2 VC (PE-PE) 2 2
Layer2 VC (PE-P) 3 3
Mid-point LSPs 1 1
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 14]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
4.8. Node Protection with 3+ hop primary (from PLR) and 2 hop backup
TE tunnels
-------- -------- -------- -------- --------
| R1 | | R2 | | R3 | | R4 | | R5 |
TG-| HE | | MID |PRI| MID |PRI| MP |PRI| TE |-TA
| |-- | PLR |---| |---| |---| |
-------- -------- -------- -------- --------
BKP| |
| -------- |
| | R6 | |
---------| BKP |-------
| MID |
--------
Figure 9: Representing setup for section 4.8
Traffic No of Labels No of labels
before failure after failure
IP TRAFFIC (P-P) 1 2
Layer3 VPN (PE-PE) 2 3
Layer3 VPN (PE-P) 3 4
Layer2 VC (PE-PE) 2 3
Layer2 VC (PE-P) 3 4
Any 1 2
4.9. Baseline MPLS Forwarding Performance Test Topology
------- -------- --------
| R1 | | R2 | | R3 |
| HE |--| MID |----| TE |
| | | | | |
------- -------- --------
Figure 10: Baseline Forwarding Performance
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 15]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
5. Test Methodology
The procedure described in this section can be applied to all the 8
base test cases and the associated topologies. The backup as well as
the primary tunnel are configured to be alike in terms of bandwidth
usage. In order to benchmark failover with all possible label stack
depth applicable as seen with current deployments, it is suggested
that the methodology includes all the scenarios listed here
5.1. Headend as PLR with link failure
Objective
To benchmark the MPLS failover time due to Link failure events
described in section 3.1 experienced by the DUT which is the point
of local repair (PLR).
Test Setup
- select any one topology out of 8 from section 4
- select overlay technology for FRR test e.g IGP,VPN,or VC
- The DUT will also have 2 interfaces connected to the traffic
Generator/analyzer. (If the node downstream of the PLR is not
A simulated node, then the Ingress of the tunnel should have
one link connected to the traffic generator and the node
downstream to the PLR or the egress of the tunnel should have
a link connected to the traffic analyzer).
Test Configuration
1. Configure the number of primaries on R2 and the backups on
R2 as required by the topology selected.
2. Advertise prefixes (as per FRR Scalability table describe in
Appendix A) by the tail end.
Procedure
1. Establish the primary lsp on R2 required by the topology
selected
2. Establish the backup lsp on R2 required by the selected
topology
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 16]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
3. Verify primary and backup lsps are up and that primary is
protected
4. Verify Fast Reroute protection is enabled and ready
5. Setup traffic streams as described in section 3.7
6. Send IP traffic at maximum Forwarding Rate to DUT.
7. Verify traffic switched over Primary LSP.
8. Trigger any choice of Link failure as describe in section
3.1
9. Verify that primary tunnel and prefixes gets mapped to
backup tunnels
10. Stop traffic stream and measure the traffic loss.
11. Failover time is calculated as per defined in section 6,
Reporting format.
12. Start traffic stream again to verify reversion when
protected interface comes up. Traffic loss should be 0 due
to make before break or reversion.
13. Enable protected interface that was down (Node in the case
of NNHOP)
14. Verify head-end signals new LSP and protection should be in
place again
5.2. Mid-Point as PLR with link failure
Objective
To benchmark the MPLS failover time due to Link failure events
described in section 3.1 experienced by the device under test which
is the point of local repair (PLR).
Test Setup
- select any one topology out of 8 from section 4
- select overlay technology for FRR test as Mid-Point lsps
- The DUT will also have 2 interfaces connected to the traffic
generator.
Test Configuration
1. Configure the number of primaries on R1 and the backups on
R2 as required by the topology selected
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 17]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
2. Advertise prefixes (as per FRR Scalability table describe in
Appendix A) by the tail end.
Procedure
1. Establish the primary lsp on R1 required by the topology
selected
2. Establish the backup lsp on R2 required by the selected
topology
3. Verify primary and backup lsps are up and that primary is
protected
4. Verify Fast Reroute protection
5. Setup traffic streams as described in section 3.7
6. Send IP traffic at maximum Forwarding Rate to DUT.
7. Verify traffic switched over Primary LSP.
8. Trigger any choice of Link failure as describe in section
3.1
9. Verify that primary tunnel and prefixes gets mapped to
backup tunnels
10. Stop traffic stream and measure the traffic loss.
11. Failover time is calculated as per defined in section 6,
Reporting format.
12. Start traffic stream again to verify reversion when
protected interface comes up. Traffic loss should be 0 due
to make before break or reversion
13. Enable protected interface that was down (Node in the case
of NNHOP)
14. Verify head-end signals new LSP and protection should be in
place again
5.3. Headend as PLR with Node failure
Objective
To benchmark the MPLS failover time due to Node failure events
described in section 3.1 experienced by the device under test which
is the point of local repair (PLR).
Test Setup
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 18]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
- select any one topology from section 4.5 to 4.8
- select overlay technology for FRR test e.g IGP,VPN,or VC
- The DUT will also have 2 interfaces connected to the traffic
generator.
Test Configuration
1. Configure the number of primaries on R2 and the backups on
R2 as required by the topology selected
2. Advertise prefixes (as per FRR Scalability table describe in
Appendix A) by the tail end.
Procedure
1. Establish the primary lsp on R2 required by the topology
selected
2. Establish the backup lsp on R2 required by the selected
topology
3. Verify primary and backup lsps are up and that primary is
protected
4. Verify Fast Reroute protection
5. Setup traffic streams as described in section 3.7
6. Send IP traffic at maximum Forwarding Rate to DUT.
7. Verify traffic switched over Primary LSP.
8. Trigger any choice of Node failure as describe in section
3.1
9. Verify that primary tunnel and prefixes gets mapped to
backup tunnels
10. Stop traffic stream and measure the traffic loss.
11. Failover time is calculated as per defined in section 6,
Reporting format.
12. Start traffic stream again to verify reversion when
protected interface comes up. Traffic loss should be 0 due
to make before break or reversion
13. Boot protected Node that was down.
14. Verify head-end signals new LSP and protection should be in
place again
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 19]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
5.4. Mid-Point as PLR with Node failure
Objective
To benchmark the MPLS failover time due to Node failure events
described in section 3.1 experienced by the device under test which
is the point of local repair (PLR).
Test Setup
- select any one topology from section 4.5 to 4.8
- select overlay technology for FRR test as Mid-Point lsps
- The DUT will also have 2 interfaces connected to the traffic
generator.
Test Configuration
1. Configure the number of primaries on R1 and the backups on
R2 as required by the topology selected
2. Advertise prefixes (as per FRR Scalability table describe in
Appendix A) by the tail end.
Procedure
1. Establish the primary lsp on R1 required by the topology
selected
2. Establish the backup lsp on R2 required by the selected
topology
3. Verify primary and backup lsps are up and that primary is
protected
4. Verify Fast Reroute protection
5. Setup traffic streams as described in section 3.7
6. Send IP traffic at maximum Forwarding Rate to DUT.
7. Verify traffic switched over Primary LSP.
8. Trigger any choice of Node failure as describe in section
3.1
9. Verify that primary tunnel and prefixes gets mapped to
backup tunnels
10. Stop traffic stream and measure the traffic loss.
11. Failover time is calculated as per defined in section 6,
Reporting format.
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 20]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
12. Start traffic stream again to verify reversion when
protected interface comes up. Traffic loss should be 0 due
to make before break or reversion
13. Boot protected Node that was down
14. Verify head-end signals new LSP and protection should be in
place again
5.5. Baseline MPLS Forwarding Performance Test Cases
For the following Forwarding Performance Benchmarking cases, the
egress must not send an implicit-null label. That is PHP should
not occur.
5.5.1. DUT Throughput as Ingress
Objective
To baseline the MPLS Throughput of the DUT acting as an
Ingress.
Procedure
1. Configure the DUT as R1, Ingress and the Tester as R2/R3
Midpoint and Egress as shown in Figure 10.
2. Execute the Throughput benchmarking test, as specified in
[RFC-BENCH], paragraph 26.1.
Expected Results:
The DUT will push a single label onto the IP packet and
forward it to the Tester as an MPLS packet.
5.5.2. DUT Latency as Ingress
Objective
To baseline the MPLS Latency of the DUT acting as an
Ingress.
Procedure
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 21]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
1. Configure the DUT as R1, Ingress and the Tester as R2/R3
Midpoint and Egress as shown in Figure 10.
2. Execute the Latency benchmarking test, as specified in
[RFC-BENCH], paragraph 26.2.
Expected Results:
The DUT will push a single label onto the IP packet and
forward it to the Tester as an MPLS packet.
5.5.3. DUT Throughput as Egress
Objective
To baseline the MPLS Throughput of the DUT acting as an
Egress.
Procedure
1. Configure the DUT as R3, Egress and the Tester as R1/R2
Ingress and Midpoint as shown in Figure 10.
2. Execute the Throughput benchmarking test, as specified in
[RFC-BENCH], paragraph 26.1 using MPLS labeled IP packets for
the offered load.
Expected Results:
The DUT will pop a single label from the IP packet and
forward it to the Tester as an IP packet.
5.5.4. DUT Latency as Egress
Objective
To baseline the MPLS Latency of the DUT acting as an Egress.
Procedure
1. Configure the DUT as R3, Egress and the Tester as R1/R2
Ingress and Midpoint as shown in Figure 10.
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 22]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
2. Execute the Latency benchmarking test, as specified in
[RFC-BENCH], paragraph 26.2 using MPLS labeled IP packets for
the offered load.
Expected Results:
The DUT will pop a single label from the IP packet and
forward it to the Tester as an IP packet.
5.5.5. DUT Throughput as Mid-Point
Objective
To baseline the MPLS Throughput of the DUT acting as a Mid-
Point.
Procedure
1. Configure the DUT as R2, Mid-Point and the Tester as
R1/R3 Ingress and Egress as shown in Figure 10.
2. Execute the Throughput benchmarking test, as specified in
[RFC-BENCH], paragraph 26.1 using MPLS labeled IP packets for
the offered load.
Expected Results:
The DUT will receive the MPLS labeled packet, swap a single
MPLS label and forward it to the Tester as an MPLS labeled
packet.
5.5.6. DUT Latency as Mid-Point
Objective
To baseline the MPLS Latency of the DUT acting as a Mid-
Point.
Procedure
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 23]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
1. Configure the DUT as R2, Mid-Point and the Tester as
R1/R3 Ingress and Egress as shown in Figure 10.
2. Execute the Latency benchmarking test, as specified in
[RFC-BENCH], paragraph 26.2 using MPLS labeled IP packets for
the offered load.
Expected Results:
The DUT will receive the MPLS labeled packet, swap a single
MPLS label and forward it to the Tester as an MPLS labeled
packet.
6. Reporting Format
For each test, it is recommended that the results be reported in the
following format.
Parameter Units
IGP used for the test ISIS-TE/ OSPF-TE
Interface types Gige,POS,ATM,VLAN etc.
Packet Sizes offered to the DUT Bytes
IGP routes advertised number of IGP routes
RSVP hello timers configured (if any) milliseconds
Number of FRR tunnels configured number of tunnels
Number of VPN routes in head-end number of VPN routes
Number of VC tunnels number of VC tunnels
Number of BGP routes number of BGP routes
Number of mid-point tunnels number of tunnels
Number of Prefixes protected by Primary number of prefixes
Number of LSPs being protected number of LSPs
Topology being used Section number
Failure Event Event type
Benchmarks
Minimum failover time milliseconds
Mean failover time milliseconds
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 24]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
Maximum failover time milliseconds
Minimum reversion time milliseconds
Mean reversion time milliseconds
Maximum reversion time milliseconds
Failover time suggested above is calculated using the following
formula: (Numbers of packet drop/rate per second * 1000) milliseconds
Note: If the primary is configured to be dynamic, and if the primary
is to reroute, make before break should occur from the backup that is
in use to a new alternate primary. If there is any packet loss seen,
it should be added to failover time.
7. Security Considerations
Documents of this type do not directly affect the security of
the Internet or of corporate networks as long as benchmarking
is not performed on devices or systems connected to operating
networks.
8. Acknowledgements
We would like to thank Jean Philip Vasseur for his invaluable input
to the document and Curtis Villamizar for his contribution in suggesting
text on definition and need for benchmarking Correlated failures.
Additionally we would like to thank Arun Gandhi, Amrit Hanspal, Karu
Ratnam and for their input to the document.
9. References
9.1. Normative References
[MPLS-RSVP] R. Braden, Ed., et al, "Resource ReSerVation
protocol (RSVP) -- version 1 functional
specification," RFC2205, September 1999.
[MPLS-RSVP-TE] D. Awduche, et al, "RSVP-TE: Extensions to
RSVP for LSP Tunnels", RFC3209, December 2001.
[MPLS-FRR-EXT] Pan, P., Atlas, A., Swallow, G.,
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 25]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
"Fast Reroute Extensions to RSVP-TE for LSP
Tunnels", RFC 4090.
[MPLS-ARCH] Rosen, E., Viswanathan, A. and R. Callon,
"Multiprotocol Label Switching Architecture",
RFC 3031, January 2001.
[RFC-BENCH] Bradner, S. and McQuaid, J., "Benchmarking
Methodology for Network Interconnect Devices",
RFC 2544.
9.2. Informative References
[MPLS-LDP] Andersson, L., Doolan, P., Feldman, N.,
Fredette, A. and B. Thomas, "LDP Specification",
RFC 3036, January 2001.
[RFC-WORDS] Bradner, S., "Key words for use in RFCs to
Indicate Requirement Levels", RFC 2119,
March 1997.
[RFC-IANA] T. Narten and H. Alvestrand, "Guidelines for
Writing an IANA Considerations Section in RFCs",
RFC 2434.
[TERM-ID] Poretsky, S., Papneja, R.,
"Benchmarking Terminology for Protection
Performance", draft-poretsky-protection-term-
00.txt, work in progress.
[IGP-METH] S. Poretsky, B. Imhoff. "Benchmarking Methodology
for IGP Data Plane Route Convergence," draft-ietf-
bmwg-igp-dataplane-conv-meth-11.txt, work in
progress.
10. Author's Address
Rajiv Papneja
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 26]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
Isocore
12359 Sunrise Valley Drive, STE 100
Reston, VA 20190
USA
Phone: +1 703 860 9273
Email: rpapneja@isocore.com
Samir Vapiwala
Cisco System
300 Beaver Brook Road
Boxborough, MA 01719
USA
Phone: +1 978 936 1484
Email: svapiwal@cisco.com
Jay Karthik
Cisco System
300 Beaver Brook Road
Boxborough, MA 01719
USA
Phone: +1 978 936 0533
Email: jkarthik@cisco.com
Scott Poretsky
Reef Point Systems
8 New England Executive Park
Burlington, MA 01803
USA
Phone: + 1 781 395 5090
EMail: sporetsky@reefpoint.com
Shankar Rao
Qwest Communications,
950 17th Street
Suite 1900
Qwest Communications
Denver, CO 80210
USA
Phone: + 1 303 437 6643
Email: shankar.rao@qwest.com
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 27]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
Jean-Louis Le Roux
France Telecom
2 av Pierre Marzin
22300 Lannion
France
Phone: 00 33 2 96 05 30 20
Email: jeanlouis.leroux@orange-ft.com
Full Copyright Statement
Copyright (C) The Internet Society (2006).
This document is subject to the rights, licenses and restrictions
contained in BCP 78, and except as set forth therein, the authors
retain all their rights.
This document and the information contained herein are provided on an
"AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS
OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET
ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE
INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED
WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
Intellectual Property
The IETF takes no position regarding the validity or scope of any
Intellectual Property Rights or other rights that might be claimed to
pertain to the implementation or use of the technology described in
this document or the extent to which any license under such rights
might or might not be available; nor does it represent that it has
made any independent effort to identify any such rights. Information
on the procedures with respect to rights in RFC documents can be
found in BCP 78 and BCP 79.
Copies of IPR disclosures made to the IETF Secretariat and any
assurances of licenses to be made available, or the result of an
attempt made to obtain a general license or permission for the use of
such proprietary rights by implementers or users of this
specification can be obtained from the IETF on-line IPR repository at
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 28]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
http://www.ietf.org/ipr.
The IETF invites any interested party to bring to its attention any
copyrights, patents or patent applications, or other proprietary
rights that may cover technology that may be required to implement
this standard. Please address the information to the IETF at ietf-
ipr@ietf.org.
Acknowledgement
Funding for the RFC Editor function is currently provided by the
Internet Society.
Appendix A: Fast Reroute Scalability Table
This section provides the recommended numbers for evaluating the
scalability of fast reroute implementations. It also recommends the
typical numbers for IGP/VPNv4 Prefixes, LSP Tunnels and VC entries.
Based on the features supported by the device under test, appropriate
scaling limits can be used for the test bed.
A 1. FRR IGP Table
No of Headend IGP Prefixes
TE LSPs
1 100
1 500
1 1000
1 2000
1 5000
2(Load Balance) 100
2(Load Balance) 500
2(Load Balance) 1000
2(Load Balance) 2000
2(Load Balance) 5000
100 100
500 500
1000 1000
2000 2000
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 29]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
A 2. FRR VPN Table
No of Headend VPNv4 Prefixes
TE LSPs
1 100
1 500
1 1000
1 2000
1 5000
1 10000
1 20000
1 Max
2(Load Balance) 100
2(Load Balance) 500
2(Load Balance) 1000
2(Load Balance) 2000
2(Load Balance) 5000
2(Load Balance) 10000
2(Load Balance) 20000
2(Load Balance) Max
A 3. FRR Mid-Point LSP Table
No of Mid-point TE LSps could be configured at the following
recommended levels
100
500
1000
2000
Max supported number
A 4. FRR VC Table
No of Headend VC entries
TE LSPs
1 100
1 500
1 1000
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 30]
Poretsky, Rao, Le Roux
Internet-Draft Methodology for benchmarking MPLS October 2006
Protection Mechanisms
1 2000
1 Max
100 100
500 500
1000 1000
2000 2000
Appendix B: Abbreviations
BFD - Bidirectional Fault Detection
BGP - Border Gateway protocol
CE - Customer Edge
DUT - Device Under Test
FRR - Fast Reroute
IGP - Interior Gateway Protocol
IP - Internet Protocol
LSP - Label Switched Path
MP - Merge Point
MPLS - Multi Protocol Label Switching
N-Nhop - Next - Next Hop
Nhop - Next Hop
OIR - Online Insertion and Removal
P - Provider
PE - Provider Edge
PHP - Penultimate Hop Popping
PLR - Point of Local Repair
RSVP - Resource reSerVation Protocol
SRLG - Shared Risk Link Group
TA - Traffic Analyzer
TE - Traffic Engineering
TG - Traffic Generator
VC - Virtual Circuit
VPN - Virtual Private Network
Papneja, Vapiwala, Karthik, Expires April 13, 2007 [Page 31]
Poretsky, Rao, Le Roux
