rfc9322







Internet Engineering Task Force (IETF)                        T. Mizrahi
Request for Comments: 9322                                        Huawei
Category: Standards Track                                   F. Brockners
ISSN: 2070-1721                                                    Cisco
                                                             S. Bhandari
                                                             Thoughtspot
                                                                B. Gafni
                                                                  Nvidia
                                                              M. Spiegel
                                                       Barefoot Networks
                                                           November 2022


In Situ Operations, Administration, and Maintenance (IOAM) Loopback and
                              Active Flags

Abstract

   In situ Operations, Administration, and Maintenance (IOAM) collects
   operational and telemetry information in packets while they traverse
   a path between two points in the network.  This document defines two
   new flags in the IOAM Trace Option headers, specifically the Loopback
   and Active flags.

Status of This Memo

   This is an Internet Standards Track document.

   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).  Further information on
   Internet Standards is available in Section 2 of RFC 7841.

   Information about the current status of this document, any errata,
   and how to provide feedback on it may be obtained at
   https://www.rfc-editor.org/info/rfc9322.

Copyright Notice

   Copyright (c) 2022 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
   (https://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 Revised BSD License text as described in Section 4.e of the
   Trust Legal Provisions and are provided without warranty as described
   in the Revised BSD License.

Table of Contents

   1.  Introduction
   2.  Conventions
     2.1.  Requirements Language
     2.2.  Terminology
   3.  New IOAM Trace Option Flags
   4.  Loopback in IOAM
     4.1.  Loopback: Encapsulating Node Functionality
       4.1.1.  Loopback Packet Selection
     4.2.  Receiving and Processing Loopback
     4.3.  Loopback on the Return Path
     4.4.  Terminating a Looped-Back Packet
   5.  Active Measurement with IOAM
   6.  IANA Considerations
   7.  Performance Considerations
   8.  Security Considerations
   9.  References
     9.1.  Normative References
     9.2.  Informative References
   Acknowledgments
   Contributors
   Authors' Addresses

1.  Introduction

   IOAM [RFC9197] is used for monitoring traffic in the network by
   incorporating IOAM data fields into in-flight data packets.

   IOAM data may be represented in one of four possible IOAM options:
   Pre-allocated Trace, Incremental Trace, Proof of Transit (POT), and
   Edge-to-Edge.  This document defines two new flags in the Pre-
   allocated and Incremental Trace options: the Loopback and Active
   flags.

   The Loopback flag is used to request that each transit device along
   the path loops back a truncated copy of the data packet to the
   sender.  The Active flag indicates that a packet is used for active
   measurement.  The term "active measurement" in the context of this
   document is as defined in [RFC7799].

2.  Conventions

2.1.  Requirements Language

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and
   "OPTIONAL" in this document are to be interpreted as described in
   BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all
   capitals, as shown here.

2.2.  Terminology

   Abbreviations used in this document:

   IOAM:   In situ Operations, Administration, and Maintenance

   OAM:    Operations, Administration, and Maintenance [RFC6291]

3.  New IOAM Trace Option Flags

   This document defines two new flags in the Pre-allocated and
   Incremental Trace options:

   Bit 1 "Loopback" (L-bit):  When set, the Loopback flag triggers the
      sending of a copy of a packet back towards the source, as further
      described in Section 4.

   Bit 2 "Active" (A-bit):  When set, the Active flag indicates that a
      packet is an active measurement packet rather than a data packet,
      where "active" is used in the sense defined in [RFC7799].  The
      packet may be an IOAM probe packet or a replicated data packet
      (the second and third use cases of Section 5).

4.  Loopback in IOAM

   The Loopback flag is used to request that each transit device along
   the path loops back a truncated copy of the data packet to the
   sender.  Loopback allows an IOAM encapsulating node to trace the path
   to a given destination and to receive per-hop data about both the
   forward and return paths.  Loopback is intended to provide an
   accelerated alternative to Traceroute that allows the encapsulating
   node to receive responses from multiple transit nodes along the path
   in less than one round-trip time (RTT) and by sending a single
   packet.

   As illustrated in Figure 1, an IOAM encapsulating node can push an
   IOAM encapsulation that includes the Loopback flag onto some or all
   of the packets it forwards using one of the IOAM encapsulation types,
   e.g., [IOAM-NSH] or [IOAM-IPV6-OPTIONS].  The IOAM transit node and
   the decapsulating node both create copies of the packet and loop them
   back to the encapsulating node.  The decapsulating node also
   terminates the IOAM encapsulation and then forwards the packet
   towards the destination.  The two IOAM looped-back copies are
   terminated by the encapsulating node.

 +--------+     +--------+     +--------+     +--------+     +--------+
 |        |     |  IOAM  |.....|  IOAM  |.....|  IOAM  |     |        |
 +--------+     +--------+     +--------+     +--------+     +--------+
 | L2/L3  |<===>| L2/L3  |<===>| L2/L3  |<===>| L2/L3  |<===>| L2/L3  |
 +--------+     +--------+     +--------+     +--------+     +--------+
   Source      Encapsulating    Transit      Decapsulating   Destination
                    Node           Node           Node

                <------------  IOAM-Domain  ----------->

                     IOAM encap. with Loopback flag
 Data packet  ------->============================>----------->
                                   |             |
                  IOAM looped back |             |
                     <=============+             |
                                 IOAM looped back|
                     <===========================+

                       Figure 1: Loopback in IOAM

   Loopback can be used only if a return path from transit nodes and
   destination nodes towards the source (encapsulating node) exists.
   Specifically, loopback is only applicable in encapsulations in which
   the identity of the encapsulating node is available in the
   encapsulation header.  If an encapsulating node receives a looped-
   back packet that was not originated from the current encapsulating
   node, the packet is dropped.

4.1.  Loopback: Encapsulating Node Functionality

   The encapsulating node either generates synthetic packets with an
   IOAM trace option that has the Loopback flag set or sets the Loopback
   flag in a subset of the in-transit data packets.  Loopback is used
   either proactively or on-demand, i.e., when a failure is detected.
   The encapsulating node also needs to ensure that sufficient space is
   available in the IOAM header for loopback operation, which includes
   transit nodes adding trace data on the original path and again on the
   return path.

   An IOAM trace option that has the Loopback flag set MUST have the
   value '1' in the most significant bit of IOAM-Trace-Type and '0' in
   the rest of the bits of IOAM-Trace-Type.  Thus, every transit node
   that processes this trace option only adds a single data field, which
   is the Hop_Lim and node_id data field.  A transit node that receives
   a packet with an IOAM trace option that has the Loopback flag set and
   the IOAM-Trace-Type is not equal to '1' in the most significant bit
   and '0' in the rest of the bits MUST NOT loop back a copy of the
   packet.  The reason for allowing only a single data field per hop is
   to minimize the impact of amplification attacks.

   IOAM encapsulating nodes MUST NOT push an IOAM encapsulation with the
   Loopback flag onto data packets that already include an IOAM
   encapsulation.  This requirement is intended to prevent IOAM Loopback
   nesting where looped-back packets may be subject to loopback in a
   nested IOAM-Domain.

4.1.1.  Loopback Packet Selection

   If an IOAM encapsulating node incorporates the Loopback flag into all
   the traffic it forwards, it may lead to an excessive amount of looped
   back packets, which may overload the network and the encapsulating
   node.  Therefore, an IOAM encapsulating node that supports the
   Loopback flag MUST support the ability to incorporate the Loopback
   flag selectively into a subset of the packets that are forwarded by
   it.

   Various methods of packet selection and sampling have been previously
   defined, such as [RFC7014] and [RFC5475].  Similar techniques can be
   applied by an IOAM encapsulating node to apply loopback to a subset
   of the forwarded traffic.

   The subset of traffic that is forwarded or transmitted with a
   Loopback flag SHOULD NOT exceed 1/N of the interface capacity on any
   of the IOAM encapsulating node's interfaces.  This requirement
   applies to the total traffic that incorporates a Loopback flag,
   including traffic that is forwarded by the IOAM encapsulating node
   and probe packets that are generated by the IOAM encapsulating node.
   In this context, N is a parameter that can be configurable by network
   operators.  If there is an upper bound, M, on the number of IOAM
   transit nodes in any path in the network, then configuring N such
   that N >> M (i.e., N is much greater than M) is RECOMMENDED.  The
   rationale is that a packet that includes the Loopback flag triggers a
   looped-back packet from each IOAM transit node along the path for a
   total of M looped-back packets.  Thus, if N >> M, then the number of
   looped-back packets is significantly lower than the number of data
   packets forwarded by the IOAM encapsulating node.  It is RECOMMENDED
   that the default value of N satisfies N>100 to be used in the absence
   of explicit operator configuration or if there is no prior knowledge
   about the network topology or size.

   An IOAM-Domain in which the Loopback flag is used MUST be configured
   such that there is expected to be a return path from each of the IOAM
   transit and IOAM decapsulating nodes; if this expectation does not
   apply, or if the encapsulating node's identity is not available in
   the encapsulation header, then configuration MUST NOT enable the
   Loopback flag to be set.

4.2.  Receiving and Processing Loopback

   A Loopback flag that is set indicates to the transit nodes processing
   this option that they are to create a copy of the received packet and
   send the copy back to the source of the packet.  In this context, the
   source is the IOAM encapsulating node and it is assumed that the
   source address is available in the encapsulation header.  Thus, the
   source address of the original packet is used as the destination
   address in the copied packet.  If IOAM is used over an encapsulation
   that does not include the address of the encapsulating node, then the
   transit/decapsulating node does not loop back a copy of the original
   packet.  The address of the node performing the copy operation is
   used as the source address; the specific method of source address
   assignment is encapsulation specific, e.g., if an IPv6 encapsulation
   is used, then the source address can be assigned as specified in
   [RFC6724].  The copy is also truncated, i.e., any payload that
   resides after the IOAM option(s) is removed before transmitting the
   looped-back packet back towards the encapsulating node.  Creating the
   copy that is looped back, and specifically the truncation, may
   require some encapsulation-specific updates in the encapsulation
   header.  The original packet continues towards its destination.  The
   L-bit MUST be cleared in the copy of the packet that a node sends
   back towards the source.

   An IOAM node that supports the reception and processing of the
   Loopback flag MUST support the ability to limit the rate of the
   looped-back packets.  The rate of looped-back packets SHOULD be
   limited so that the number of looped-back packets is significantly
   lower than the number of packets that are forwarded by the device.
   The looped-back data rate SHOULD NOT exceed 1/N of the interface
   capacity on any of the IOAM node's interfaces.  Using N>100 is
   RECOMMENDED.  Depending on the IOAM node's architecture
   considerations, the loopback response rate may be limited to a lower
   number in order to avoid overloading the IOAM node.

4.3.  Loopback on the Return Path

   On its way back towards the source, the copied packet is processed
   like any other packet with IOAM information, including adding
   requested data at each transit node (assuming there is sufficient
   space).

4.4.  Terminating a Looped-Back Packet

   Once the return packet reaches the IOAM-Domain boundary, IOAM
   decapsulation occurs as with any other packet containing IOAM
   information.  Note that the looped-back packet does not have the
   L-bit set.  The IOAM encapsulating node that initiated the original
   loopback packet recognizes a received packet as an IOAM looped-back
   packet by checking the Node ID in the Hop_Lim/node_id field that
   corresponds to the first hop.  If the Node ID and IOAM-Namespace
   match the current IOAM node, it indicates that this is a looped-back
   packet that was initiated by the current IOAM node and processed
   accordingly.  If there is no match in the Node ID, the packet is
   processed like a conventional IOAM-encapsulated packet.

   Note that an IOAM encapsulating node may be either an endpoint (such
   as an IPv6 host) or a switch/router that pushes a tunnel
   encapsulation onto data packets.  In both cases, the functionality
   that was described above avoids IOAM data leaks from the IOAM-Domain.
   Specifically, if an IOAM looped-back packet reaches an IOAM boundary
   node that is not the IOAM node that initiated the loopback, the node
   does not process the packet as a loopback; the IOAM encapsulation is
   removed, preventing IOAM information from leaking out from the IOAM-
   Domain.  Since the packet does not have any payload, it is
   terminated.

5.  Active Measurement with IOAM

   Active measurement methods [RFC7799] make use of synthetically
   generated packets in order to facilitate measurement.  This section
   presents use cases of active measurement using the IOAM Active flag.

   The Active flag indicates that a packet is used for active
   measurement.  An IOAM decapsulating node that receives a packet with
   the Active flag set in one of its Trace options must terminate the
   packet.  The Active flag is intended to simplify the implementation
   of decapsulating nodes by indicating that the packet should not be
   forwarded further.  It is not intended as a replacement for existing
   active OAM protocols, which may run in higher layers and make use of
   the Active flag.

   An example of an IOAM deployment scenario is illustrated in Figure 2.
   The figure depicts two endpoints: a source and a destination.  The
   data traffic from the source to the destination is forwarded through
   a set of network devices, including an IOAM encapsulating node (which
   incorporates one or more IOAM options), a decapsulating node (which
   removes the IOAM options), and optionally one or more transit nodes.
   The IOAM options are encapsulated in one of the IOAM encapsulation
   types, e.g., [IOAM-NSH] or [IOAM-IPV6-OPTIONS].


 +--------+     +--------+     +--------+     +--------+     +--------+
 |        |     |  IOAM  |.....|  IOAM  |.....|  IOAM  |     |        |
 +--------+     +--------+     +--------+     +--------+     +--------+
 | L2/L3  |<===>| L2/L3  |<===>| L2/L3  |<===>| L2/L3  |<===>| L2/L3  |
 +--------+     +--------+     +--------+     +--------+     +--------+
   Source      Encapsulating    Transit      Decapsulating   Destination
                   Node           Node           Node

                <------------  IOAM-Domain  ----------->

                      Figure 2: Network Using IOAM

   This document focuses on three possible use cases of active
   measurement using IOAM.  These use cases are described using the
   example of Figure 2.

   Endpoint active measurement:
      synthetic probe packets are sent between the source and
      destination, traversing the IOAM-Domain.  Since the probe packets
      are sent between the endpoints, these packets are treated as data
      packets by the IOAM-Domain and do not require special treatment at
      the IOAM layer.  Specifically, the Active flag is not used in this
      case and the IOAM layer does not need to be aware that an active
      measurement mechanism is used at a higher layer.

   IOAM active measurement using probe packets within the IOAM-
   Domain:
      probe packets are generated and transmitted by the IOAM
      encapsulating node and are expected to be terminated by the
      decapsulating node.  IOAM data related to probe packets may be
      exported by one or more nodes along its path by an exporting
      protocol that is outside the scope of this document (e.g.,
      [IOAM-RAWEXPORT]).  Probe packets include a Trace Option that has
      its Active flag set, indicating that the decapsulating node must
      terminate them.  The specification of these probe packets and the
      processing of these packets by the encapsulating and decapsulating
      nodes is outside the scope of this document.

   IOAM active measurement using replicated data packets:
      probe packets are created by the encapsulating node by selecting
      some or all of the en route data packets and replicating them.  A
      selected data packet and its (possibly truncated) copy is
      forwarded with one or more IOAM options while the original packet
      is forwarded normally without IOAM options.  To the extent
      possible, the original data packet and its replica are forwarded
      through the same path.  The replica includes a Trace Option that
      has its Active flag set, indicating that the decapsulating node
      should terminate it.  The current document defines the role of the
      Active flag in allowing the decapsulating node to terminate the
      packet, but the replication functionality and the functionality of
      the decapsulating node in this context is outside the scope of
      this document.

   If the volume of traffic that incorporates the Active flag is large,
   it may overload the network and the IOAM node(s) that process the
   active measurement packet.  Thus, the rate of the traffic that
   includes the Active flag SHOULD NOT exceed 1/N of the interface
   capacity on any of the IOAM node's interfaces.  Using N>100 is
   RECOMMENDED.  Depending on the IOAM node's architecture
   considerations, the rate of Active-enabled IOAM packets may be
   limited to a lower number in order to avoid overloading the IOAM
   node.

6.  IANA Considerations

   IANA has allocated the following bits in the "IOAM Trace-Flags"
   registry as follows:

   Bit 1  "Loopback" (L-bit)

   Bit 2  "Active" (A-bit)

   This document is specified as the "Reference" in the registry for
   both bits.

   Note that bit 0 is the most significant bit in the "IOAM Trace-Flags"
   registry.  This bit was allocated by [RFC9197] as the 'Overflow' bit.

7.  Performance Considerations

   Each of the flags that are defined in this document may have
   performance implications.  When using the loopback mechanism, a copy
   of the data packet is sent back to the sender (thus, generating more
   traffic than originally sent by the endpoints).  Using active
   measurement with the Active flag requires the use of synthetic
   (overhead) traffic.

   Each of the mechanisms that use the flags above has a cost in terms
   of the network bandwidth and may potentially load the node that
   analyzes the data.  Therefore, it MUST be possible to use each of the
   mechanisms on a subset of the data traffic; an encapsulating node
   needs to be able to set the Loopback and Active flags selectively in
   a way that considers the effect on the network performance, as
   further discussed in Sections 4.1.1 and 5.

   Transit and decapsulating nodes that support loopback need to be able
   to limit the looped-back packets (as discussed in Section 4.2) so as
   to ensure that the mechanisms are used at a rate that does not
   significantly affect the network bandwidth and does not overload the
   source node in the case of loopback.

8.  Security Considerations

   The security considerations of IOAM in general are discussed in
   [RFC9197].  Specifically, an attacker may try to use the
   functionality that is defined in this document to attack the network.

   IOAM is assumed to be deployed in a restricted administrative domain,
   thus limiting the scope of the threats above and their effect.  This
   is a fundamental assumption with respect to the security aspects of
   IOAM as further discussed in [RFC9197].  However, even given this
   limited scope, security threats should still be considered and
   mitigated.  Specifically, an attacker may attempt to overload network
   devices by injecting synthetic packets that include an IOAM Trace
   Option with one or more of the flags defined in this document.
   Similarly, an on-path attacker may maliciously set one or more of the
   flags of transit packets.

   Loopback flag:
      an attacker that sets this flag, either in synthetic packets or
      transit packets, can potentially cause an amplification since each
      device along the path creates a copy of the data packet and sends
      it back to the source.  The attacker can potentially leverage the
      Loopback flag for a DDoS attack as multiple devices send looped-
      back copies of a packet to a single victim.

   Active flag:
      the impact of synthetic packets with the Active flag is no worse
      than synthetic data packets in which the Active flag is not set.
      By setting the Active flag in en route packets, an attacker can
      prevent these packets from reaching their destination since the
      packet is terminated by the decapsulating device.  However, note
      that an on-path attacker may achieve the same goal by changing the
      destination address of a packet.  Another potential threat is
      amplification; if an attacker causes transit switches to replicate
      more packets than they are intended to replicate (either by
      setting the Active flag or by sending synthetic packets), then
      traffic is amplified, causing bandwidth degradation.  As mentioned
      in Section 5, the specification of the replication mechanism is
      not within the scope of this document.  A specification that
      defines the replication functionality should also address the
      security aspects of this mechanism.

   Some of the security threats that were discussed in this document may
   be worse in a wide area network in which there are nested IOAM-
   Domains.  For example, if there are two nested IOAM-Domains that use
   loopback, then a looped-back copy in the outer IOAM-Domain may be
   forwarded through another (inner) IOAM-Domain and may be subject to
   loopback in that (inner) IOAM-Domain, causing the amplification to be
   worse than in the conventional case.

   In order to mitigate the performance-related attacks described in
   Section 7, it should be possible for IOAM-enabled devices to
   selectively apply the mechanisms that use the flags defined in this
   document to a subset of the traffic and to limit the performance of
   synthetically generated packets to a configurable rate.
   Specifically, IOAM nodes should be able to:

   *  Limit the rate of IOAM packets with the Loopback flag (IOAM
      encapsulating nodes) as discussed in Section 4.1.1.

   *  Limit the rate of looped back packets (IOAM transit and
      decapsulating nodes) as discussed in Section 4.2.

   *  Limit the rate of IOAM packets with the Active flag (IOAM
      encapsulating nodes) as discussed in Section 5.

   As defined in Section 4, transit nodes that process a packet with the
   Loopback flag only add a single data field and truncate any payload
   that follows the IOAM option(s), thus significantly limiting the
   possible impact of an amplification attack.

9.  References

9.1.  Normative References

   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119,
              DOI 10.17487/RFC2119, March 1997,
              <https://www.rfc-editor.org/info/rfc2119>.

   [RFC8174]  Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC
              2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174,
              May 2017, <https://www.rfc-editor.org/info/rfc8174>.

   [RFC9197]  Brockners, F., Ed., Bhandari, S., Ed., and T. Mizrahi,
              Ed., "Data Fields for In Situ Operations, Administration,
              and Maintenance (IOAM)", RFC 9197, DOI 10.17487/RFC9197,
              May 2022, <https://www.rfc-editor.org/info/rfc9197>.

9.2.  Informative References

   [IOAM-IPV6-OPTIONS]
              Bhandari, S., Ed. and F. Brockners, Ed., "In-situ OAM IPv6
              Options", Work in Progress, Internet-Draft, draft-ietf-
              ippm-ioam-ipv6-options-09, 11 October 2022,
              <https://datatracker.ietf.org/doc/html/draft-ietf-ippm-
              ioam-ipv6-options-09>.

   [IOAM-NSH] Brockners, F., Ed. and S. Bhandari, Ed., "Network Service
              Header (NSH) Encapsulation for In-situ OAM (IOAM) Data",
              Work in Progress, Internet-Draft, draft-ietf-sfc-ioam-nsh-
              11, 30 September 2022,
              <https://datatracker.ietf.org/doc/html/draft-ietf-sfc-
              ioam-nsh-11>.

   [IOAM-RAWEXPORT]
              Spiegel, M., Brockners, F., Bhandari, S., and R.
              Sivakolundu, "In-situ OAM raw data export with IPFIX",
              Work in Progress, Internet-Draft, draft-spiegel-ippm-ioam-
              rawexport-06, 21 February 2022,
              <https://datatracker.ietf.org/doc/html/draft-spiegel-ippm-
              ioam-rawexport-06>.

   [RFC5475]  Zseby, T., Molina, M., Duffield, N., Niccolini, S., and F.
              Raspall, "Sampling and Filtering Techniques for IP Packet
              Selection", RFC 5475, DOI 10.17487/RFC5475, March 2009,
              <https://www.rfc-editor.org/info/rfc5475>.

   [RFC6291]  Andersson, L., van Helvoort, H., Bonica, R., Romascanu,
              D., and S. Mansfield, "Guidelines for the Use of the "OAM"
              Acronym in the IETF", BCP 161, RFC 6291,
              DOI 10.17487/RFC6291, June 2011,
              <https://www.rfc-editor.org/info/rfc6291>.

   [RFC6724]  Thaler, D., Ed., Draves, R., Matsumoto, A., and T. Chown,
              "Default Address Selection for Internet Protocol Version 6
              (IPv6)", RFC 6724, DOI 10.17487/RFC6724, September 2012,
              <https://www.rfc-editor.org/info/rfc6724>.

   [RFC7014]  D'Antonio, S., Zseby, T., Henke, C., and L. Peluso, "Flow
              Selection Techniques", RFC 7014, DOI 10.17487/RFC7014,
              September 2013, <https://www.rfc-editor.org/info/rfc7014>.

   [RFC7799]  Morton, A., "Active and Passive Metrics and Methods (with
              Hybrid Types In-Between)", RFC 7799, DOI 10.17487/RFC7799,
              May 2016, <https://www.rfc-editor.org/info/rfc7799>.

Acknowledgments

   The authors thank Martin Duke, Tommy Pauly, Donald Eastlake, Paul
   Kyzivat, Bernard Aboba, Greg Mirsky, and other members of the IPPM
   working group for many helpful comments.

Contributors

   The Editors would like to recognize the contributions of the
   following individuals to this document.

   Ramesh Sivakolundu
   Cisco Systems, Inc.
   170 West Tasman Dr.
   San Jose, CA 95134
   United States of America
   Email: sramesh@cisco.com


   Carlos Pignataro
   Cisco Systems, Inc.
   7200-11 Kit Creek Road
   Research Triangle Park, NC 27709
   United States of America
   Email: cpignata@cisco.com


   Aviv Kfir
   Nvidia
   Email: avivk@nvidia.com


   Jennifer Lemon
   Broadcom
   270 Innovation Drive
   San Jose, CA 95134
   United States of America
   Email: jennifer.lemon@broadcom.com


Authors' Addresses

   Tal Mizrahi
   Huawei
   Israel
   Email: tal.mizrahi.phd@gmail.com


   Frank Brockners
   Cisco Systems, Inc.
   3rd Floor
   Hansaallee 249
   40549 Duesseldorf
   Germany
   Email: fbrockne@cisco.com


   Shwetha Bhandari
   Thoughtspot
   3rd Floor
   Indiqube Orion
   Garden Layout
   HSR Layout
   24th Main Rd
   Bangalore 560 102
   Karnataka
   India
   Email: shwetha.bhandari@thoughtspot.com


   Barak Gafni
   Nvidia
   Suite 100
   350 Oakmead Parkway
   Sunnyvale, CA 94085
   United States of America
   Email: gbarak@nvidia.com


   Mickey Spiegel
   Barefoot Networks, an Intel company
   4750 Patrick Henry Drive
   Santa Clara, CA 95054
   United States of America
   Email: mickey.spiegel@intel.com


ERRATA