Internet Engineering Task Force (IETF) R. Papneja Request for Comments: 7747 Huawei Technologies Category: Informational B. Parise ISSN: 2070-1721 Skyport Systems S. Hares Huawei Technologies D. Lee IXIA I. Varlashkin Google April 2016
Basic BGP Convergence Benchmarking Methodology for Data-Plane Convergence
Abstract
BGP is widely deployed and used by several service providers as the default inter-AS (Autonomous System) routing protocol. It is of utmost importance to ensure that when a BGP peer or a downstream link of a BGP peer fails, the alternate paths are rapidly used and routes via these alternate paths are installed. This document provides the basic BGP benchmarking methodology using existing BGP convergence terminology as defined in RFC 4098.
Status of This Memo
This document is not an Internet Standards Track specification; it is published for informational purposes.
This document is a product of the Internet Engineering Task Force (IETF). It represents the consensus of the IETF community. It has received public review and has been approved for publication by the Internet Engineering Steering Group (IESG). Not all documents approved by the IESG are a candidate for any level of Internet Standard; see Section 2 of RFC 5741.
Information about the current status of this document, any errata, and how to provide feedback on it may be obtained at http://www.rfc-editor.org/info/rfc7747.
Papneja, et al. Informational [Page 1]
RFC 7747 BGP Convergence Methodology April 2016
Copyright Notice
Copyright (c) 2016 IETF Trust and the persons identified as the document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.
This document may contain material from IETF Documents or IETF Contributions published or made publicly available before November 10, 2008. The person(s) controlling the copyright in some of this material may not have granted the IETF Trust the right to allow modifications of such material outside the IETF Standards Process. Without obtaining an adequate license from the person(s) controlling the copyright in such materials, this document may not be modified outside the IETF Standards Process, and derivative works of it may not be created outside the IETF Standards Process, except to format it for publication as an RFC or to translate it into languages other than English.
This document defines the methodology for benchmarking data-plane Forwarding Information Base (FIB) convergence performance of BGP in routers and switches using topologies of three or four nodes. The methodology proposed in this document applies to both IPv4 and IPv6, and if a particular test is unique to one version, it is marked accordingly. For IPv6 benchmarking, the Device Under Test (DUT) will require the support of Multiprotocol BGP (MP-BGP) [RFC4760] [RFC2545]. Similarly, both Internal BGP (iBGP) and External BGP (eBGP) are covered in the tests as applicable.
The scope of this document is to provide methodology for BGP FIB convergence measurements with BGP functionality limited to IPv4 and IPv6 as defined in [RFC4271] and MP-BGP [RFC4760] [RFC2545]. Other BGP extensions to support Layer 2 and Layer 3 Virtual Private Networks (VPNs) are outside the scope of this document. Interaction with IGPs (IGP interworking) is outside the scope of this document.
The terminology used in this document is defined in [RFC4098]. One additional term is defined in this document as follows.
FIB (data-plane) convergence is defined as the completion of all FIB changes so that all forwarded traffic then takes the newly proposed route. RFC 4098 defines the terms 'BGP device', 'FIB', and 'forwarded traffic'. Data-plane convergence is different than control-plane convergence within a node.
This document defines methodology to test
o data-plane convergence on a single BGP device that supports the BGP functionality with a scope as outlined above; and
o using test topology of three or four nodes that are sufficient to recreate the convergence events used in the various tests of this document.
In the current Internet architecture, the inter-AS transit is primarily available through BGP. To maintain reliable connectivity within intra-domains or across inter-domains, fast recovery from failures remains most critical. To ensure minimal traffic losses, many service providers are requiring BGP implementations to converge the entire Internet routing table within sub-seconds at FIB level.
Papneja, et al. Informational [Page 4]
RFC 7747 BGP Convergence Methodology April 2016
Furthermore, to compare these numbers amongst various devices, service providers are also looking at ways to standardize the convergence measurement methods. This document offers test methods for simple topologies. These simple tests will provide a quick high- level check of BGP data-plane convergence across multiple implementations from different vendors.
The convergence of BGP occurs at two levels: Routing Information Base (RIB) and FIB convergence. RFC 4098 defines terms for BGP control- plane convergence. Methodologies that test control-plane convergence are out of scope for this document.
In order to ensure that the results obtained in tests are repeatable, careful setup of initial conditions and exact steps are required.
This document proposes these initial conditions, test steps, and result checking. To ensure uniformity of the results, all optional parameters SHOULD be disabled and all settings SHOULD be changed to default; these may include BGP timers as well.
"Benchmarking Terminology for Network Interconnect Devices" [RFC1242] and "Benchmarking Terminology for LAN Switching Devices" [RFC2285] SHOULD be reviewed in conjunction with this document. WLAN-specific terms and definitions are also provided in Clauses 3 and 4 of the IEEE 802.11 standard [IEEE.802.11]. Commonly used terms may also be found in RFC 1983 [RFC1983].
For the sake of clarity and continuity, this document adopts the general template for benchmarking terminology set out in Section 2 of [RFC1242]. Definitions are organized in alphabetical order and grouped into sections for ease of reference. The following terms are assumed to be taken as defined in RFC 1242 [RFC1242]: Throughput, Latency, Constant Load, Frame Loss Rate, and Overhead Behavior. In addition, the following terms are taken as defined in [RFC2285]: Forwarding Rates, Maximum Forwarding Rate, Loads, Device Under Test (DUT), and System Under Test (SUT).
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].
This section describes the test setups for use in BGP benchmarking tests measuring convergence of the FIB (data-plane) after BGP updates have been received.
These test setups have three or four nodes with the following configuration:
Figure 4: Four-Node Setup for eBGP and iBGP Convergence
4. Test Considerations
The test cases for measuring convergence for iBGP and eBGP are different. Both iBGP and eBGP use different mechanisms to advertise, install, and learn the routes. Typically, an iBGP route on the DUT is installed and exported when the next hop is valid. For eBGP, the route is installed on the DUT with the remote interface address as the next hop, with the exception of the multihop test case (as specified in the test).
"Number of Peers" is defined as the number of BGP neighbors or sessions the DUT has at the beginning of the test. The peers are established before the tests begin. The relationship could be either iBGP or eBGP peering depending upon the test case requirement.
The DUT establishes one or more BGP peer sessions with one or more emulated routers or Helper Nodes. Additional peers can be added based on the testing requirements. The number of peers enabled during the testing should be well documented in the report matrix.
"Number of Routes per Peer" is defined as the number of routes advertised or learned by the DUT per session or through a neighbor relationship with an emulator or Helper Node. The Tester, emulating as a BGP neighbor, MUST advertise at least one route per BGP peer.
Each test run must identify the route stream in terms of route packing, route mixture, and number of routes. This route stream must be well documented in the reporting stream. RFC 4098 defines these terms.
It is RECOMMENDED that the user consider advertising the entire current Internet routing table per peering session using an Internet route mixture with unique or non-unique routes. If multiple peers are used, it is important to precisely document the timing sequence between the peer sending routes (as defined in RFC 4098).
The DUT MUST run one baseline test where policy is the Minimal policy as defined in RFC 4098. Additional runs may be done with the policy that was set up before the tests began. Exact policy settings MUST be documented as part of the test.
There are configured parameters and timers that may impact the measured BGP convergence times.
The benchmark metrics MAY be measured at any fixed values for these configured parameters.
Papneja, et al. Informational [Page 9]
RFC 7747 BGP Convergence Methodology April 2016
It is RECOMMENDED these configure parameters have the following settings: a) default values specified by the respective RFC, b) platform-specific default parameters, and c) values as expected in the operational network. All optional BGP settings MUST be kept consistent across iterations of any specific tests
Examples of the configured parameters that may impact measured BGP convergence time include, but are not limited to:
The type of media dictates which test cases may be executed; each interface type has a unique mechanism for detecting link failures, and the speed at which that mechanism operates will influence the measurement results. All interfaces MUST be of the same media and throughput for all iterations of each test case.
Since observed packet loss is used to measure the route convergence time, the time between two successive packets offered to each individual route is the highest possible accuracy of any packet-loss- based measurement. When packet jitter is much less than the convergence time, it is a negligible source of error, and hence, it will be treated as within tolerance.
Other options to measure convergence are the Time-Based Loss Method (TBLM) and Timestamp-Based Method (TBM) [RFC6414].
An exterior measurement on the input media (such as Ethernet) is defined by this specification.
The benchmark measurements may vary for each trial due to the statistical nature of timer expirations, CPU scheduling, etc. It is recommended to repeat the test multiple times. Evaluation of the test data must be done with an understanding of generally accepted testing practices regarding repeatability, variance, and statistical significance of a small number of trials.
For any repeated tests that are averaged to remove variance, all parameters MUST remain the same.
Authentication in BGP is done using the TCP Authentication Option [RFC5925]. (In some legacy situations, the authentication may still be with TCP MD5). The processing of the authentication hash, particularly in devices with a large number of BGP peers and a large amount of update traffic, can have an impact on the control plane of
Papneja, et al. Informational [Page 11]
RFC 7747 BGP Convergence Methodology April 2016
the device. If authentication is enabled, it MUST be documented correctly in the reporting format.
Also, it is recommended that trials MUST be with the same Secure Inter-Domain Routing (SIDR) features [RFC7115] [BGPsec]. The best convergence tests would be with no SIDR features and then to repeat the convergence tests with the same SIDR features.
Convergence events or triggers are defined as abnormal occurrences in the network, which initiate route flapping in the network and hence forces the reconvergence of a steady state network. In a real network, a series of convergence events may cause convergence latency operators desire to test.
These convergence events must be defined in terms of the sequences defined in RFC 4098. This basic document begins all tests with a router initial setup. Additional documents will define BGP data- plane convergence based on peer initialization.
The convergence events may or may not be tied to the actual failure. A soft reset [RFC4098] does not clear the RIB or FIB tables. A hard reset clears BGP peer sessions, RIB tables, and FIB tables.
Due to the different Non-Stop-Routing (sometimes referred to High- Availability) solutions available from different vendors, it is RECOMMENDED that any redundancy available in the routing processors should be disabled during the convergence measurements. For cases where the redundancy cannot be disabled, the results are no longer comparable and the level of impact on the measurements is out of scope of this document.
5. Test Cases
All tests defined under this section assume the following:
a. BGP peers are in Established state.
b. BGP state should be cleared from Established state to Idle prior to each test. This is recommended to ensure that all tests start with BGP peers being forced back to Idle state and databases flushed.
Papneja, et al. Informational [Page 12]
RFC 7747 BGP Convergence Methodology April 2016
c. Furthermore, the traffic generation and routing should be verified in the topology to ensure there is no packet loss observed on any advertised routes.
d. The arrival timestamp of advertised routes can be measured by installing an inline monitoring device between the emulator and the DUT or by using the span port of the DUT connected with an external analyzer. The time base of such an inline monitor or external analyzer needs to be synchronized with the protocol and traffic emulator. Some modern emulators may have the capability to capture and timestamp every NLRI packet leaving and arriving at the emulator ports. The timestamps of these NLRI packets will be almost identical to the arrival time at the DUT if the cable distance between the emulator and DUT is relatively short.
This test measures the convergence time taken to receive and install a route in RIB using BGP.
Reference Test Setup:
This test uses the setup as shown in Figure 1
Procedure:
A. All variables affecting convergence should be set to a basic test state (as defined in Section 4.4).
B. Establish BGP adjacency between the DUT and one peer of the emulator, Emp1.
Papneja, et al. Informational [Page 13]
RFC 7747 BGP Convergence Methodology April 2016
C. To ensure adjacency establishment, wait for three keepalives to be received from the DUT or a configurable delay before proceeding with the rest of the test.
D. Start the traffic from the emulator tx towards the DUT targeted at a route specified in the route mixture (e.g., routeA). Initially, no traffic SHOULD be observed on the egress interface as routeA is not installed in the forwarding database of the DUT.
E. Advertise routeA from the peer (Emp1) to the DUT and record the time.
This is Tup(Emp1,Rt-A), also named XMT-Rt-time(Rt-A).
F. Record the time when routeA from Emp1 is received at the DUT.
This is Tup(DUT,Rt-A), also named RCV-Rt-time(Rt-A).
G. Record the time when the traffic targeted towards routeA is received by the emulator on the appropriate traffic egress interface.
This is TR(TDr,Rt-A), also named DUT-XMT-Data-Time(Rt-A).
H. The difference between the Tup(DUT,RT-A) and traffic received time (TR (TDr, Rt-A) is the FIB convergence time for routeA in the route mixture. A full convergence for the route update is the measurement between the first route (Rt-A) and the last route (Rt-last).
Route update convergence is
TR(TDr, Rt-last)- Tup(DUT, Rt-A), or
(DUT-XMT-Data-Time - RCV-Rt-Time)(Rt-A).
Note: It is recommended that a single test with the same route mixture be repeated several times. A report should provide the standard deviation and the average of all tests.
Running tests with a varying number of routes and route mixtures is important to get a full characterization of a single peer.
This test measures the convergence time taken by an implementation to receive, install, and advertise a route using BGP.
Reference Test Setup:
This test uses the setup as shown in Figure 2.
Procedure:
A. The Helper Node (HLP) MUST run same version of BGP as the DUT.
B. All devices MUST be synchronized using NTP or some local reference clock.
C. All configuration variables for the Helper Node, DUT, and emulator SHOULD be set to the same values. These values MAY be basic test or a unique set completely described in the test setup.
D. Establish BGP adjacency between the DUT and the emulator.
E. Establish BGP adjacency between the DUT and the Helper Node.
F. To ensure adjacency establishment, wait for three keepalives to be received from the DUT or a configurable delay before proceeding with the rest of the test.
G. Start the traffic from the emulator towards the Helper Node targeted at a specific route (e.g., routeA). Initially, no traffic SHOULD be observed on the egress interface as routeA is not installed in the forwarding database of the DUT.
H. Advertise routeA from the emulator to the DUT and note the time.
This is Tup(EMx, Rt-A), also named EM-XMT-Data-Time(Rt-A).
This is Tup(DUTr, Rt-A), also named DUT-RCV-Rt-Time(Rt-A).
J. Record the time when routeA is forwarded by the DUT towards the Helper Node.
This is Tup(DUTx, Rt-A), also named DUT-XMT-Rt-Time(Rt-A).
Papneja, et al. Informational [Page 15]
RFC 7747 BGP Convergence Methodology April 2016
K. Record the time when the traffic targeted towards routeA is received on the Route Egress Interface. This is TR(EMr, Rt-A), also named DUT-XMT-Data Time(Rt-A).
This test measures the convergence time taken by an implementation to receive, install, and advertise a route in an eBGP Multihop Scenario.
Reference Test Setup:
This test uses the setup as shown in Figure 3. The DUT is used along with a Helper Node.
Procedure:
A. The Helper Node MUST run the same version of BGP as the DUT.
B. All devices MUST be synchronized using NTP or some local reference clock.
C. All variables affecting convergence, like authentication, policies, and timers, SHOULD be set to basic settings.
D. All three devices, the DUT, emulator, and Helper Node, are configured with different ASs.
E. Loopback interfaces are configured on the DUT and Helper Node, and connectivity is established between them using any config options available on the DUT.
F. Establish BGP adjacency between the DUT and the emulator.
G. Establish BGP adjacency between the DUT and the Helper Node.
H. To ensure adjacency establishment, wait for three keepalives to be received from the DUT or a configurable delay before proceeding with the rest of the test
I. Start the traffic from the emulator towards the DUT targeted at a specific route (e.g., routeA).
J. Initially, no traffic SHOULD be observed on the egress interface as routeA is not installed in the forwarding database of the DUT.
K. Advertise routeA from the emulator to the DUT and note the time (Tup(EMx,RouteA), also named Route-Tx-time(Rt-A).
Papneja, et al. Informational [Page 17]
RFC 7747 BGP Convergence Methodology April 2016
L. Record the time when the route is received by the DUT. This is Tup(EMr,DUT), also named Route-Rcv-time(Rt-A).
M. Record the time when the traffic targeted towards routeA is received from the egress interface of the DUT on the emulator. This is Tup(EMd,DUT) named Data-Rcv-time(Rt-A)
N. Record the time when routeA is forwarded by the DUT towards the Helper Node. This is Tup(EMf,DUT), also named Route-Fwd-time(Rt- A).
Note: It is recommended that the test be repeated with a varying number of routes and route mixtures. With each set route mixture, the test should be repeated multiple times. The results should record the average, mean, standard deviation.
This test measures the route convergence time due to a local link failure event at the DUT's Local Interface.
Reference Test Setup:
This test uses the setup as shown in Figure 1. The shutdown event is defined as an administrative shutdown event on the DUT.
Procedure:
A. All variables affecting convergence, like authentication, policies, and timers, should be set to basic-test policy.
B. Establish two BGP adjacencies from the DUT to the emulator, one over the peer interface and the other using a second peer interface.
C. Advertise the same route, routeA, over both adjacencies with preferences so that the Best Egress Interface for the preferred next hop is (Emp1) interface.
Papneja, et al. Informational [Page 18]
RFC 7747 BGP Convergence Methodology April 2016
D. To ensure adjacency establishment, wait for three keepalives to be received from the DUT or a configurable delay before proceeding with the rest of the test.
E. Start the traffic from the emulator towards the DUT targeted at a specific route (e.g., routeA). Initially, traffic would be observed on the best egress route, Emp1, instead of Emp2.
F. Trigger the shutdown event of Best Egress Interface on the DUT (Dp1). This time is called Shutdown time.
G. Measure the convergence time for the event to be detected and traffic to be forwarded to Next-Best Egress Interface (Dp2).
Time = Data-detect(Emp2) - Shutdown time
H. Stop the offered load and wait for the queues to drain. Restart the data flow.
I. Bring up the link on the DUT's Best Egress Interface.
J. Measure the convergence time taken for the traffic to be rerouted from Dp2 to Best Egress Interface, Dp1.
Time = Data-detect(Emp1) - Bring Up time
K. It is recommended that the test be repeated with a varying number of routes and route mixtures or with a number of routes and route mixtures closer to what is deployed in operational networks.
5.2.2. Physical Link Failure on Remote/Emulator End
Objective:
This test measures the route convergence time due to a local link failure event at the Tester's Local Interface.
Reference Test Setup:
This test uses the setup as shown in Figure 1. The shutdown event is defined as a shutdown of the local interface of the Tester via a logical shutdown event. The procedure used in Section 5.2.1 is used for the termination.
This test measures the route convergence time due to a local link failure event at the ECMP member. The FIB configuration and BGP are set to allow two ECMP routes to be installed. However, policy directs the routes to be sent only over one of the paths.
Reference Test Setup:
This test uses the setup as shown in Figure 1, and the procedure used in Section 5.2.1.
5.3. BGP Adjacency Failure (Non-Physical Link Failure) on Emulator
Objective:
This test measures the route convergence time due to BGP Adjacency Failure on the emulator.
Reference Test Setup:
This test uses the setup as shown in Figure 1.
Procedure:
A. All variables affecting convergence, like authentication, policies, and timers, should be set to basic-policy.
B. Establish two BGP adjacencies from the DUT to the emulator: one over the Best Egress Interface and the other using the Next-Best Egress Interface.
C. Advertise the same route, routeA, over both adjacencies with preferences so that the Best Egress Interface for the preferred next hop is (Emp1) interface.
D. To ensure adjacency establishment, wait for three keepalives to be received from the DUT or a configurable delay before proceeding with the rest of the test.
E. Start the traffic from the emulator towards the DUT targeted at a specific route (e.g., routeA). Initially, traffic would be observed on the Best Egress Interface.
Papneja, et al. Informational [Page 20]
RFC 7747 BGP Convergence Methodology April 2016
F. Remove BGP adjacency via a software adjacency down on the emulator on the Best Egress Interface. This time is called BGPadj-down-time, also termed BGPpeer-down.
G. Measure the convergence time for the event to be detected and traffic to be forwarded to Next-Best Egress Interface. This time is Tr-rr2, also called TR2-traffic-on.
Convergence = TR2-traffic-on - BGPpeer-down
H. Stop the offered load and wait for the queues to drain and restart the data flow.
I. Bring up BGP adjacency on the emulator over the Best Egress Interface. This time is BGP-adj-up, also called BGPpeer-up.
J. Measure the convergence time taken for the traffic to be rerouted to the Best Egress Interface. This time is Tr-rr1, also called TR1-traffic-on.
This test measures the route convergence time due to a hard reset on the DUT.
Reference Test Setup:
This test uses the setup as shown in Figure 1.
Procedure:
A. The requirement for this test case is that the hard reset event should be non-recovering and should affect only the adjacency between the DUT and the emulator on the Best Egress Interface.
B. All variables affecting the test SHOULD be set to basic-test values.
C. Establish two BGP adjacencies from the DUT to the emulator: one over the Best Egress Interface and the other using the Next-Best Egress Interface.
Papneja, et al. Informational [Page 21]
RFC 7747 BGP Convergence Methodology April 2016
D. Advertise the same route, routeA, over both adjacencies with preferences so that the Best Egress Interface for the preferred next hop is (Emp1) interface.
E. To ensure adjacency establishment, wait for three keepalives to be received from the DUT or a configurable delay before proceeding with the rest of the test.
F. Start the traffic from the emulator towards the DUT targeted at a specific route (e.g., routeA). Initially, traffic would be observed on the Best Egress Interface.
G. Trigger the hard reset event of the Best Egress Interface on the DUT. This time is called time reset.
H. This event is detected and traffic is forwarded to the Next-Best Egress Interface. This time is called time-traffic flow.
I. Measure the convergence time for the event to be detected and traffic to be forwarded to Next-Best Egress Interface.
Time of convergence = time-traffic flow - time-reset
J. Stop the offered load and wait for the queues to drain and restart.
K. It is recommended that the test be repeated with a varying number of routes and route mixtures or with a number of routes and route mixtures closer to what is deployed in operational networks.
L. When varying number of routes are used, convergence time is measured using the Loss-Derived method [RFC6412].
M. Convergence time in this scenario is influenced by failure detection time on the Tester, BGP keepalive time and routing, and forwarding table update time.
This test measures the route convergence time taken by an implementation to service a BGP Route Refresh message and advertise a route.
Reference Test Setup:
This test uses the setup as shown in Figure 2.
Papneja, et al. Informational [Page 22]
RFC 7747 BGP Convergence Methodology April 2016
Procedure:
A. The BGP implementation on the DUT and Helper Node needs to support BGP Route Refresh Capability [RFC2918].
B. All devices MUST be synchronized using NTP or some local reference clock.
C. All variables affecting convergence, like authentication, policies, and timers, should be set to basic-test defaults.
D. The DUT and the Helper Node are configured in the same AS, whereas the emulator is configured under a different AS.
E. Establish BGP adjacency between the DUT and the emulator.
F. Establish BGP adjacency between the DUT and the Helper Node.
G. To ensure adjacency establishment, wait for three keepalives to be received from the DUT or a configurable delay before proceeding with the rest of the test.
H. Configure a policy under the BGP on the Helper Node to deny routes received from the DUT.
I. Advertise routeA from the emulator to the DUT.
J. The DUT will try to advertise the route to the Helper Node; it will be denied.
K. Wait for three keepalives.
L. Start the traffic from the emulator towards the Helper Node targeted at a specific route, say routeA. Initially, no traffic would be observed on the egress interface, as routeA is not present.
M. Remove the policy on the Helper Node and issue a route refresh request towards the DUT. Note the timestamp of this event. This is the RefreshTime.
N. Record the time when the traffic targeted towards routeA is received on the egress interface. This is RecTime.
O. The following equation represents the Route Refresh Convergence Time per route.
Route Refresh Convergence Time = (RecTime - RefreshTime)
This test measures the route convergence time taken by an implementation to service a BGP withdraw message and advertise the withdraw.
Reference Test Setup:
This test uses the setup as shown in Figure 2.
Procedure:
A. This test consists of two steps to determine the Total Withdraw Processing Time.
B. Step 1:
(1) All devices MUST be synchronized using NTP or some local reference clock.
(2) All variables should be set to basic-test parameters.
(3) The DUT and Helper Node are configured in the same AS, whereas the emulator is configured under a different AS.
(4) Establish BGP adjacency between the DUT and the emulator.
(5) To ensure adjacency establishment, wait for three keepalives to be received from the DUT or a configurable delay before proceeding with the rest of the test.
(6) Start the traffic from the emulator towards the DUT targeted at a specific route (e.g., routeA). Initially, no traffic would be observed on the egress interface as routeA is not present on the DUT.
(7) Advertise routeA from the emulator to the DUT.
(8) The traffic targeted towards routeA is received on the egress interface.
(9) Now the Tester sends a request to withdraw routeA to the DUT. TRx(Awith) is also called WdrawTime1(Rt-A).
(10) Record the time when no traffic is observed as determined by the emulator. This is the RouteRemoveTime1(Rt-A).
Papneja, et al. Informational [Page 24]
RFC 7747 BGP Convergence Methodology April 2016
(11) The difference between the RouteRemoveTime1 and WdrawTime1 is the WdrawConvTime1.
(1) Continuing from Step 1, re-advertise routeA back to the DUT from the Tester.
(2) The DUT will try to advertise routeA to the Helper Node (this assumes there exists a session between the DUT and Helper Node).
(3) Start the traffic from the emulator towards the Helper Node targeted at a specific route (e.g., routeA). Traffic would be observed on the egress interface after routeA is received by the Helper Node.
WATime=time traffic first flows
(4) Now the Tester sends a request to withdraw routeA to DUT. This is the WdrawTime2(Rt-A).
WAWtime-TRx(Rt-A) = WdrawTime2(Rt-A)
(5) DUT processes the withdraw and sends it to the Helper Node.
(6) Record the time when no traffic is observed as determined by the emulator. This is:
This test measures the convergence time taken by an implementation to service a BGP Path Attribute Change.
Reference Test Setup:
This test uses the setup as shown in Figure 1.
Procedure:
A. This test only applies to Well-Known Mandatory Attributes like origin, AS path, and next hop.
B. In each iteration of the test, only one of these mandatory attributes need to be varied whereas the others remain the same.
C. All devices MUST be synchronized using NTP or some local reference clock.
D. All variables should be set to basic-test parameters.
E. Advertise the same route, routeA, over both adjacencies with preferences so that the Best Egress Interface for the preferred next hop is (Emp1) interface.
F. To ensure adjacency establishment, wait for three keepalives to be received from the DUT or a configurable delay before proceeding with the rest of the test.
G. Start the traffic from the emulator towards the DUT targeted at the specific route (e.g., routeA). Initially, traffic would be observed on the Best Egress Interface.
H. Now advertise the same route, routeA, on the Next-Best Egress Interface but by varying one of the well-known mandatory attributes to have a preferred value over that interface. We call this Tbetter. The other values need to be the same as what was advertised on the Best-Egress adjacency.
This test measures the route convergence time taken by an implementation during a Graceful Restart Event as detailed in the terminology document [RFC4098].
Reference Test Setup:
This test uses the setup as shown in Figure 4.
Procedure:
A. It measures the time taken by an implementation to service a BGP Graceful Restart Event and advertise a route.
B. The Helper Nodes are the same model as the DUT and run the same BGP implementation as the DUT.
C. The BGP implementation on the DUT and Helper Node needs to support the BGP Graceful Restart Mechanism [RFC4724].
D. All devices MUST be synchronized using NTP or some local reference clock.
E. All variables are set to basic-test values.
F. The DUT and Helper Node 1 (HLP1) are configured in the same AS, whereas the emulator and Helper Node 2 (HLP2) are configured under different ASs.
G. Establish BGP adjacency between the DUT and Helper Nodes.
Papneja, et al. Informational [Page 27]
RFC 7747 BGP Convergence Methodology April 2016
H. Establish BGP adjacency between the Helper Node 2 and the emulator.
I. To ensure adjacency establishment, wait for three keepalives to be received from the DUT or a configurable delay before proceeding with the rest of the test.
J. Configure a policy under the BGP on Helper Node 1 to deny routes received from the DUT.
K. Advertise routeA from the emulator to Helper Node 2.
L. Helper Node 2 advertises the route to the DUT and the DUT will try to advertise the route to Helper Node 1, which will be denied.
N. Start the traffic from the emulator towards the Helper Node 1 targeted at the specific route (e.g., routeA). Initially, no traffic would be observed on the egress interface as routeA is not present.
O. Perform a Graceful Restart Trigger Event on the DUT and note the time. This is the GREventTime.
S. It is assumed in this test case that after a switchover is triggered on the DUT, it will not have any cycles to process the BGP Refresh messages. The reason for this assumption is that there is a narrow window of time where after switchover, when we remove the policy from Helper Node 1, implementations might generate Route Refresh automatically and this request might be serviced before the DUT actually switches over and re-establishes BGP adjacencies with the peers.
Papneja, et al. Informational [Page 28]
RFC 7747 BGP Convergence Methodology April 2016
6. Reporting Format
For each test case, it is recommended that the reporting tables below are completed, and all time values SHOULD be reported with resolution as specified in [RFC4098].
Parameter Units or Description =========================== ========================== Test case Test case number
Test topology 1, 2, 3, or 4
Parallel links Number of parallel links
Interface type Gigabit Ethernet (GigE), Packet over SONET (POS), ATM, other
Convergence Event Hard reset, soft reset, link failure, or other defined
eBGP sessions Number of eBGP sessions
iBGP sessions Number of iBGP sessions
eBGP neighbor Number of eBGP neighbors
iBGP neighbor Number of iBGP neighbors
Routes per peer Number of routes
Total unique routes Number of routes
Total non-unique routes Number of routes
IGP configured IS-IS, OSPF, static, or other
Route mixture Description of route mixture
Route packing Number of routes included in an update
Interface failure Seconds indication delay Hold time Seconds MinRouteAdvertisementInterval Seconds (MRAI) MinASOriginationInterval Seconds (MAOI) Keepalive time Seconds ConnectRetry Seconds
TCP parameters for DUT and Tester Maximum Segment Size (MSS) Bytes Slow start threshold Bytes Maximum window size Bytes
Test Details:
a. If the Offered Load matches a subset of routes, describe how this subset is selected.
b. Describe how the convergence event is applied; does it cause instantaneous traffic loss or not?
c. If there is any policy configured, describe the configured policy.
Papneja, et al. Informational [Page 30]
RFC 7747 BGP Convergence Methodology April 2016
Complete the table below for the initial convergence event and the reversion convergence event.
Parameter Unit =========================== ========================== Convergence Event Initial or reversion
Traffic Forwarding Metrics Total number of packets Number of packets offered to the DUT Total number of packets Number of packets forwarded by the DUT Connectivity packet loss Number of packets Convergence packet loss Number of packets Out-of-order packets Number of packets Duplicate packets Number of packets
Convergence Benchmarks
Rate-Derived Method [RFC6412]: First route convergence Seconds time Full convergence time Seconds
Loss-Derived Method [RFC6412]: Loss-Derived convergence Seconds time
Route-Specific (R-S) Loss-Derived Method: Minimum R-S convergence Seconds time Maximum R-S convergence Seconds time Median R-S convergence Seconds time Average R-S convergence Seconds time
Loss of Connectivity (LoC) Benchmarks
Loss-Derived Method: Loss-Derived loss of Seconds connectivity period
Papneja, et al. Informational [Page 31]
RFC 7747 BGP Convergence Methodology April 2016
Route-Specific Loss-Derived Method: Minimum LoC period [n] Array of seconds Minimum Route LoC period Seconds Maximum Route LoC period Seconds Median Route LoC period Seconds Average Route LoC period Seconds
7. Security Considerations
Benchmarking activities as described in this memo are limited to technology characterization using controlled stimuli in a laboratory environment, with dedicated address space and the constraints specified in the sections above.
The benchmarking network topology is an independent test setup and MUST NOT be connected to devices that may forward the test traffic into a production network or misroute traffic to the test management network.
Further, benchmarking is performed on a "black-box" basis, relying solely on measurements observable and external to the DUT/SUT.
Special capabilities SHOULD NOT exist in the DUT/SUT specifically for benchmarking purposes. Any implications for network security arising from the DUT/SUT SHOULD be identical in the lab and in production networks.
[IEEE.802.11] IEEE, "IEEE Standard for Information technology -- Telecommunications and information exchange between systems Local and metropolitan area networks -- Specific requirements Part 11: Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications", IEEE 802.11-2012, DOI 10.1109/ieeestd.2012.6178212, April 2012, <http://ieeexplore.ieee.org/servlet/ opac?punumber=6178209>.
[RFC4098] Berkowitz, H., Davies, E., Ed., Hares, S., Krishnaswamy, P., and M. Lepp, "Terminology for Benchmarking BGP Device Convergence in the Control Plane", RFC 4098, DOI 10.17487/RFC4098, June 2005, <http://www.rfc-editor.org/info/rfc4098>.
We would like to thank Anil Tandon, Arvind Pandey, Mohan Nanduri, Jay Karthik, and Eric Brendel for their input and discussions on various sections in the document. We also like to acknowledge Will Liu, Hubert Gee, Semion Lisyansky, and Faisal Shah for their review and feedback on the document.