Internet-Draft | In-situ OAM Deployment | October 2022 |
Brockners, et al. | Expires 15 April 2023 | [Page] |
In-situ Operations, Administration, and Maintenance (IOAM) collects operational and telemetry information in the packet while the packet traverses a path between two points in the network. This document provides a framework for IOAM deployment and discusses best current practices.¶
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
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."¶
This Internet-Draft will expire on 15 April 2023.¶
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.¶
"In-situ" Operations, Administration, and Maintenance (IOAM) collects OAM information within the packet while the packet traverses a particular network domain. The term "in-situ" refers to the fact that the OAM data is added to the data packets rather than is being sent within packets specifically dedicated to OAM. IOAM is to complement mechanisms such as Ping, Traceroute, or other active probing mechanisms. In terms of "active" or "passive" OAM, "in-situ" OAM can be considered a hybrid OAM type. "In-situ" mechanisms do not require extra packets to be sent. IOAM adds information to the already available data packets and therefore cannot be considered passive. In terms of the classification given in [RFC7799] IOAM could be portrayed as Hybrid Type I. IOAM mechanisms can be leveraged where mechanisms using e.g. ICMP do not apply or do not offer the desired results, such as proving that a certain traffic flow takes a pre-defined path, SLA verification for the live data traffic, detailed statistics on traffic distribution paths in networks that distribute traffic across multiple paths, or scenarios in which probe traffic is potentially handled differently from regular data traffic by the network devices.¶
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.¶
Abbreviations used in this document:¶
IOAM is focused on "limited domains" as defined in [RFC8799]. IOAM is not targeted for a deployment on the global Internet. The part of the network which employs IOAM is referred to as the "IOAM-Domain". For example, an IOAM-domain can include an enterprise campus using physical connections between devices or an overlay network using virtual connections / tunnels for connectivity between said devices. An IOAM-domain is defined by its perimeter or edge. The operator of an IOAM-domain is expected to put provisions in place to ensure that packets which contain IOAM data fields do not leak beyond the edge of an IOAM domain, e.g. using for example packet filtering methods. The operator should consider the potential operational impact of IOAM to mechanisms such as ECMP load-balancing schemes (e.g., load-balancing schemes based on packet length could be impacted by the increased packet size due to IOAM), path MTU (i.e., ensure that the MTU of all links within a domain is sufficiently large to support the increased packet size due to IOAM) and ICMP message handling.¶
An IOAM-Domain consists of "IOAM encapsulating nodes", "IOAM decapsulating nodes" and "IOAM transit nodes". The role of a node (i.e., encapsulating, transit, decapsulating) is defined within an IOAM-Namespace (see below). A node can have different roles in different IOAM-Namespaces.¶
An "IOAM encapsulating node" incorporates one or more IOAM-Option-Types into packets that IOAM is enabled for. If IOAM is enabled for a selected subset of the traffic, the IOAM encapsulating node is responsible for applying the IOAM functionality to the selected subset.¶
An "IOAM transit node" updates one or more of the IOAM-Data-Fields. If both the Pre-allocated and the Incremental Trace Option-Types are present in the packet, each IOAM transit node will update at most one of these Option-Types. A transit node does not add new IOAM-Option-Types to a packet, and does not change the IOAM-Data-Fields of an IOAM Edge-to-Edge Option-Type.¶
An "IOAM decapsulating node" removes IOAM-Option-Type(s) from packets.¶
The role of an IOAM-encapsulating, IOAM-transit or IOAM-decapsulating node is always performed within a specific IOAM-Namespace. This means that an IOAM node which is e.g., an IOAM-decapsulating node for IOAM-Namespace "A" but not for IOAM-Namespace "B" will only remove the IOAM-Option-Types for IOAM-Namespace "A" from the packet. An IOAM decapsulating node situated at the edge of an IOAM domain removes all IOAM-Option-Types and associated encapsulation headers for all IOAM-Namespaces from the packet.¶
IOAM-Namespaces allow for a namespace-specific definition and interpretation of IOAM-Data-Fields. An interface-id could for example point to a physical interface (e.g., to understand which physical interface of an aggregated link is used when receiving or transmitting a packet) whereas in another case it could refer to a logical interface (e.g., in case of tunnels). Please refer to Section 7.1 for a discussion of IOAM-Namespaces.¶
Export of Export of Export of Export of IOAM data IOAM data IOAM data IOAM data (optional) (optional) (optional) (optional) ^ ^ ^ ^ | | | | | | | | User +---+----+ +---+----+ +---+----+ +---+----+ packets |Encapsu-| | Transit| | Transit| |Decapsu-| -------->|lating |====>| Node |====>| Node |====>|lating |--> |Node | | A | | B | |Node | +--------+ +--------+ +--------+ +--------+
IOAM nodes which add or remove the IOAM-Data-Fields can also update the IOAM-Data-Fields at the same time. Or in other words, IOAM encapsulating or decapsulating nodes can also serve as IOAM transit nodes at the same time. Note that not every node in an IOAM domain needs to be an IOAM transit node. For example, a deployment might require that packets traverse a set of firewalls which support IOAM. In that case, only the set of firewall nodes would be IOAM transit nodes rather than all nodes.¶
IOAM supports different modes of operation, which are differentiated by the type of IOAM data fields being carried in the packet, the data being collected, the type of nodes which collect or update data as well as whether and how nodes export IOAM information.¶
Per-hop tracing: OAM information about each IOAM node a packet traverses is collected and stored within the user data packet as the packet progresses through the IOAM domain. Potential uses of IOAM per-hop tracing include:¶
"IOAM tracing data" is expected to be collected at every IOAM transit node that a packet traverses to ensure visibility into the entire path a packet takes within an IOAM-Domain. I.e., in a typical deployment all nodes in an IOAM-Domain would participate in IOAM and thus be IOAM transit nodes, IOAM encapsulating or IOAM decapsulating nodes. If not all nodes within a domain are IOAM capable, IOAM tracing information (i.e., node data, see below) will only be collected on those nodes which are IOAM capable. Nodes which are not IOAM capable will forward the packet without any changes to the IOAM-Data-Fields. The maximum number of hops and the minimum path MTU of the IOAM domain is assumed to be known.¶
IOAM offers two different trace Option-Types, the "incremental" Option-Type as well as the "pre-allocated" Option-Type. For a discussion which of the two option types is the most suitable for an implementation and/or deployment, see Section 7.3.¶
Every node data entry holds information for a particular IOAM transit node that is traversed by a packet. The IOAM decapsulating node removes the IOAM-Option-Type(s) and processes and/or exports the associated data. All IOAM-Data-Fields are defined in the context of an IOAM-Namespace.¶
IOAM tracing can collect the following types of information:¶
The Option-Types of incremental tracing and pre-allocated tracing are defined in [RFC9197].¶
IOAM Proof of Transit Option-Type is to support path or service function chain [RFC7665] verification use cases. Proof-of-transit uses methods like nested hashing or nested encryption of the IOAM data or mechanisms such as Shamir's Secret Sharing Schema (SSSS).¶
The IOAM Proof of Transit Option-Type consist of a fixed size "IOAM proof of transit option header" and "IOAM proof of transit option data fields". For details see [RFC9197].¶
The IOAM Edge-to-Edge Option-Type is to carry data that is added by the IOAM encapsulating node and interpreted by IOAM decapsulating node. The IOAM transit nodes may process the data but must not modify it.¶
The IOAM Edge-to-Edge Option-Type consist of a fixed size "IOAM Edge-to-Edge Option-Type header" and "IOAM Edge-to-Edge Option-Type data fields". For details see [RFC9197].¶
Direct Export is an IOAM mode of operation within which IOAM data to be directly exported to a collector rather than being collected within the data packets. The IOAM Direct Export Option-Type consist of a fixed size "IOAM direct export option header". Direct Export for IOAM is defined in [I-D.ietf-ippm-ioam-direct-export].¶
IOAM data fields and associated data types for in-situ OAM are defined in [RFC9197]. The in-situ OAM data field can be transported by a variety of transport protocols, including NSH, Segment Routing, Geneve, BIER, IPv6, etc.¶
IOAM encapsulation for IPv6 is defined in [I-D.ietf-ippm-ioam-ipv6-options], which also discussed IOAM deployment considerations for IPv6 networks¶
IOAM encapsulation for NSH is defined in [I-D.ietf-sfc-ioam-nsh].¶
IOAM encapsulation for BIER is defined in [I-D.xzlnp-bier-ioam].¶
IOAM encapsulation for GRE is outlined as part of the "EtherType Protocol Identification of In-situ OAM Data" in [I-D.weis-ippm-ioam-eth], though no example protocol header stacks are provided in the document. When used with GRE, the IOAM-Option-Types (the below diagram uses "IOAM" as shorthand for IOAM-Option-Types) are sequenced in behind the GRE header that follows the "outer" IP header. Figure 2 shows two example protocol header stacks that use GRE along with IOAM.¶
Example 1 Example 2 | ... | | ... | +----------------+ +----------------+ | TCP/UDP header | | IP, ... | +----------------+ +----------------+ | IP header | | Eth. header | +----------------+ +----------------+ | IOAM | | IOAM | +----------------+ +----------------+ | GRE header | | GRE header | +----------------+ +----------------+ | IP header | | IP header | +----------------+ +----------------+ | Layer 2 | | Layer 2 | +----------------+ +----------------+ | Layer 1 | | Layer 1 | +----------------+ +----------------+
IOAM encapsulation for Geneve is defined in [I-D.brockners-ippm-ioam-geneve].¶
IOAM encapsulation for Segment Routing is defined in [I-D.gandhi-spring-ioam-sr-mpls].¶
IOAM encapsulation for Segment Routing over IPv6 is defined in [I-D.ali-spring-ioam-srv6].¶
IOAM encapsulation for VXLAN-GPE is defined in [I-D.brockners-ippm-ioam-vxlan-gpe].¶
IOAM nodes collect information for packets traversing a domain that supports IOAM. IOAM decapsulating nodes as well as IOAM transit nodes can choose to retrieve IOAM information from the packet, process the information further and export the information using e.g., IPFIX.¶
Raw data export of IOAM data using IPFIX is discussed in [I-D.spiegel-ippm-ioam-rawexport]. "Raw export of IOAM data" refers to a mode of operation where a node exports the IOAM data as it is received in the packet. The exporting node neither interprets, aggregates nor reformats the IOAM data before it is exported. Raw export of IOAM data is to support an operational model where the processing and interpretation of IOAM data is decoupled from the operation of encapsulating/updating/decapsulating IOAM data, which is also referred to as IOAM data-plane operation. The figure below shows the separation of concerns for IOAM export: Exporting IOAM data is performed by the "IOAM node" which performs IOAM data-plane operation, whereas the interpretation of IOAM data is performed by one or several IOAM data processing systems. The separation of concerns is to off-load interpretation, aggregation and formatting of IOAM data from the node which performs data-plane operations. In other words, a node which is focused on data-plane operations, i.e. forwarding of packets and handling IOAM data will not be tasked to also interpret the IOAM data, but can leave this task to another system or a set of systems. For scalability reasons, a single IOAM node could choose to export IOAM data to several IOAM data processing systems. Similarly, there several monitoring systems or analytics systems can be used to further process the data received from the IOAM preprocessing systems. Figure 3 shows an overview of IOAM export, including IOAM data processing systems and monitoring/analytics systems.¶
+--------------+ +-------------+ | | Monitoring/ | | | Analytics | | | system(s) |-+ +-------------+ ^ | Processed/interpreted/ | aggregated IOAM data | +--------------+ +-------------+ | | IOAM data | | | processing | | | system(s) |-+ +-------------+ ^ | Raw export of | IOAM data | +--------------+-------+------+--------------+ | | | | | | | | User +---+----+ +---+----+ +---+----+ +---+----+ packets |Encapsu-| | Transit| | Transit| |Decapsu-| -------->|lating |====>| Node |====>| Node |====>|lating |--> |Node | | A | | B | |Node | +--------+ +--------+ +--------+ +--------+
This section discusses several aspects of an IOAM deployment, including IOAM Namespaces, IOAM Layering, traffic-sets that IOAM is applied to and IOAM loopback mode.¶
IOAM-Namespaces add further context to IOAM-Option-Types and associated IOAM-Data-Fields. IOAM-Namespaces support several different uses:¶
IOAM-Namespaces can be used to identify different sets of devices (e.g., different types of devices) in a deployment: If an operator desires to insert different IOAM-Data-Fields based on the device, the devices could be grouped into multiple IOAM-Namespaces. This could be due to the fact that the IOAM feature set differs between different sets of devices, or it could be for reasons of optimized space usage in the packet header. It could also stem from hardware or operational limitations on the size of the trace data that can be added and processed, preventing collection of a full trace for a flow.¶
If several encapsulation protocols (e.g., in case of tunneling) are stacked on top of each other, IOAM-Data-Fields could be present in different protocol fields at different layers. Layering allows operators to instrument the protocol layer they want to measure. The behavior follows the ships-in-the-night model, i.e. IOAM-Data-Fields in one layer are independent from IOAM-Data-Fields in another layer. Or in other words: Even though the term "layering" often implies some form of hierarchy and relationship, in IOAM, layers are independent from each other and don't assume any relationship among them. The different layers could, but do not have to share the same IOAM encapsulation mechanisms. Similarly, the sematics of the IOAM-Data-Fields, can, do not have to be associated to across different layers. For example, a node which inserts node-id information into two different layers could use "node-id=10" for one layer and "node-id=1000" for the second layer.¶
Figure 4 shows an example of IOAM layering. The figure shows a Geneve tunnel carried over IPv6 which starts at node A and ends at node D. IOAM information is encapsulated in IPv6 as well as in Geneve. At the IPv6 layer, node A is IOAM encapsulating node (into IPv6), node D is the IOAM decapsulating node and node B and node C are IOAM transit nodes. At the Geneve layer, node A is IOAM encapsulating node (into Geneve) and node D is IOAM decapsulating node (from Geneve). The use of IOAM at both layers as shown in the example here could be used to reveal which nodes of an underlay (here the IPv6 network) are traversed by tunneled packet in an overlay (here the Geneve network) - which assumes that the IOAM information encapsulated by nodes A and D into Geneve and IPv6 is associated to each other.¶
+---+----+ +---+----+ User |Geneve | |Geneve | packets |Encapsu-| |Decapsu-| -------->|lating |==================================>|lating |--> | Node | | Node | | A | | D | +--------+ +--------+ ^ ^ | | v v +--------+ +--------+ +--------+ +--------+ |IPv6 | | IPv6 | | IPv6 | |IPv6 | |Encapsu-| | Transit| | Transit| |Decapsu-| |lating |====>| Node |====>| Node |====>|lating | | Node | | | | | | Node | | A | | B | | C | | D | +--------+ +--------+ +--------+ +--------+
IOAM offers two different IOAM Option-Types for tracing: "Incremental" Trace-Option-Type and "Pre-allocated" Trace-Option-Type. "Incremental" refers to a mode of operation where the packet is expanded at every IOAM node that addes IOAM-Data-Fields. "Pre-Allocated" describes a mode of operation where the IOAM encapsulating node allocates room for all IOAM-Data-Fields in the entire IOAM domain. More specifically:¶
A deployment can choose to configure and support one or both of the IOAM Trace-Option-Types. The operator decides by means of configuration which Trace-Option-Type(s) will be used for a particular domain. Deployments can mix devices which include either the Incremental Trace-Option-Type or the Pre-allocated Trace-Option-Type, e.g., in case different types of packet forwarders and associated different types of IOAM implementations exist in a deployment. As a result, both Option-Types can be present in a packet. IOAM decapsulating nodes remove both types of Trace-Option-Types from the packet.¶
The two different Option-Types cater to different packet forwarding infrastructures and are to allow an optimized implementation of IOAM tracing:¶
IOAM can be deployed on all or only on subsets of the live user traffic, e.g., per interface, based on an access control list or flow specification defining a specific set of traffic, etc.¶
IOAM Loopback is used to trigger each transit device along the path of a packet to send a copy of the data packet back to the source. 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 the return path. Loopback is enabled by the encapsulating node setting the loopback flag. Looped-back packets use the source address of the original packet as destination address and the address of the node which performs the loopback operation as source address. Nodes which loop back a packet clear the loopback flag before sending the copy back towards the source. Loopack applies to IOAM deployments where the encapsulating node is either a host or the start of a tunnel: For details on IOAM loopback, please refer to [I-D.ietf-ippm-ioam-flags].¶
The IOAM active mode flag indicates that a packet is an active OAM packet as opposed to regular user data traffic. Active mode is expected to be used for active measurement using IOAM. For details on IOAM active mode, please refer to [I-D.ietf-ippm-ioam-flags].¶
Example use-cases for IOAM Active Mode include:¶
A network can consist of a mix of IOAM aware and IOAM unaware nodes. The encapsulation of IOAM-Data-Fields into different protocols (see also Section 5) are defined such that data packets that include IOAM-Data-Fields do not get dropped by IOAM unaware nodes. For example, packets which contain the IOAM-Trace-Option-Types in IPv6 Hop by Hop extension headers are defined with bits to indicate "00 - skip over this option and continue processing the header". This will ensure that when a node that is IOAM unaware receives a packet with IOAM-Data-Fields included, does not drop the packet.¶
Deployments which leverage the IOAM-Trace-Option-Type(s) could benefit from the ability to detect the presence of IOAM unaware nodes, i.e. nodes which forward the packet but do not update/add IOAM-Data-Fields in IOAM-Trace-Option-Type(s). The node data that is defined as part of the IOAM-Trace-Option-Type(s) includes a Hop_Lim field associated to the node identifier to detect missed nodes, i.e. "holes" in the trace. Monitoring/Analytics system(s) could utilize this information to account for the presence of IOAM unaware nodes in the network.¶
The YANG model for configuring IOAM in network nodes which support IOAM is defined in [I-D.zhou-ippm-ioam-yang].¶
A deployment can leverage IOAM profiles is to limit the scope of IOAM features, allowing simpler implementation, verification, and interoperability testing in the context of specific use cases that do not require the full functionality of IOAM. An IOAM profile defines a use case or a set of use cases for IOAM, and an associated set of rules that restrict the scope and features of the IOAM specification, thereby limiting it to a subset of the full functionality. IOAM profiles are defined in [I-D.mizrahi-ippm-ioam-profile].¶
For deployments where the IOAM capabilities of a node are unknown, [I-D.ietf-ippm-ioam-conf-state] [I-D.ietf-ippm-ioam-flags] could be used to discover the enabled IOAM capabilities of nodes.¶
This document does not request any IANA actions.¶
As discussed in [RFC7276], a successful attack on an OAM protocol in general, and specifically on IOAM, can prevent the detection of failures or anomalies, or create a false illusion of nonexistent ones.¶
The Proof of Transit Option-Type (Section Section 4.2) is used for verifying the path of data packets. The security considerations of POT are further discussed in [I-D.ietf-sfc-proof-of-transit].¶
Security considerations related to the use of IOAM flags, in particular the loopback flag are found in [I-D.ietf-ippm-ioam-flags].¶
IOAM data can be subject to eavesdropping. Although the confidentiality of the user data is not at risk in this context, the IOAM data elements can be used for network reconnaissance, allowing attackers to collect information about network paths, performance, queue states, buffer occupancy and other information. Recon is an improbable security threat in an IOAM deployment that is within a confined physical domain. However, in deployments that are not confined to a single LAN, but span multiple inter-connected sites (for example, using an overlay network), the inter-site links can be secured (e.g., by IPsec) in order to avoid external eavesdropping. Another possible mitigation approach is to use the "direct exporting" mode [I-D.ietf-ippm-ioam-direct-export]. In this case the IOAM related trace information would not be available in the customer data packets, but would trigger exporting of (secured) packet-related IOAM information at every node. IOAM data export and securing IOAM data export is outside the scope of this document.¶
IOAM can be used as a means for implementing Denial of Service (DoS) attacks, or for amplifying them. For example, a malicious attacker can add an IOAM header to packets or modify an IOAM header in en route packets in order to consume the resources of network devices that take part in IOAM or collectors that analyze the IOAM data. Another example is a packet length attack, in which an attacker pushes headers associated with IOAM Option-Types into data packets, causing these packets to be increased beyond the MTU size, resulting in fragmentation or in packet drops. Such DoS attacks can be mitigated by deploying IOAM in confined administrative domains, and by defining performance limits on IOAM encapsulation and IOAM exporting. By limiting the rate and/or percentage of packets that are subject to IOAM encpasulation and the rate of exported traffic, it is possible to cofine the impact of such attacks.¶
Since IOAM options may include timestamps, if network devices use synchronization protocols then any attack on the time protocol [RFC7384] can compromise the integrity of the timestamp-related data fields. Synchronization attacks can be mitigated by combining a secured time distribution scheme, e.g., [RFC8915], and by using redundant clock sources [RFC5905] and/or redundant network paths for the time distribution protocol [RFC8039].¶
At the management plane, attacks may be implemented by misconfiguring or by maliciously configuring IOAM-enabled nodes in a way that enables other attacks. Thus, IOAM configuration should be secured in a way that authenticates authorized users and verifies the integrity of configuration procedures.¶
Notably, IOAM is expected to be deployed in specific network domains, thus confining the potential attack vectors to within the network domain. Indeed, in order to limit the scope of threats to within the current network domain the network operator is expected to enforce policies that prevent IOAM traffic from leaking outside of the IOAM domain, and prevent IOAM data from outside the domain to be processed and used within the domain. Note that the Immediate Export mode (reference to be added in a future revision) can mitigate the potential threat of IOAM data leaking through data packets.¶
The authors would like to thank Tal Mizrahi, Eric Vyncke, Nalini Elkins, Srihari Raghavan, Ranganathan T S, Barak Gafni, Karthik Babu Harichandra Babu, Akshaya Nadahalli, LJ Wobker, Erik Nordmark, Vengada Prasad Govindan, Andrew Yourtchenko, Aviv Kfir, Tianran Zhou, Zhenbin (Robin), Joe Clarke, Al Morton, Tom Herbet, Haoyu song, and Mickey Spiegel for the comments and advice on IOAM.¶