| Link State Routing | L. Ginsberg, Ed. |
| Internet-Draft | Cisco Systems, Inc. |
| Obsoletes: 7810 (if approved) | S. Previdi, Ed. |
| Intended status: Standards Track | |
| Expires: October 1, 2018 | S. Giacolone |
| Microsoft | |
| D. Ward | |
| Cisco Systems, Inc. | |
| J. Drake | |
| Juniper Networks | |
| Q. Wu | |
| Huawei | |
| March 30, 2018 |
IS-IS Traffic Engineering (TE) Metric Extensions
draft-ginsberg-isis-rfc7810bis-00
In certain networks, such as, but not limited to, financial information networks (e.g., stock market data providers), network- performance criteria (e.g., latency) are becoming as critical to data-path selection as other metrics.
This document describes extensions to IS-IS Traffic Engineering Extensions (RFC 5305) such that network-performance information can be distributed and collected in a scalable fashion. The information distributed using IS-IS TE Metric Extensions can then be used to make path-selection decisions based on network performance.
Note that this document only covers the mechanisms with which network-performance information is distributed. The mechanisms for measuring network performance or acting on that information, once distributed, are outside the scope of this document.
This document obsoletes RFC 7810.
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.
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 October 1, 2018.
Copyright (c) 2018 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 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.
In certain networks, such as, but not limited to, financial information networks (e.g., stock market data providers), network- performance information (e.g., latency) is becoming as critical to data-path selection as other metrics.
In these networks, extremely large amounts of money rest on the ability to access market data in "real time" and to predictably make trades faster than the competition. Because of this, using metrics such as hop count or cost as routing metrics is becoming only tangentially important. Rather, it would be beneficial to be able to make path-selection decisions based on performance data (such as latency) in a cost-effective and scalable way.
This document describes extensions (hereafter called "IS-IS TE Metric Extensions") to the IS-IS Extended Reachability TLV defined in [RFC5305], that can be used to distribute network-performance information (such as link delay, delay variation, packet loss, residual bandwidth, and available bandwidth).
The data distributed by the IS-IS TE Metric Extensions proposed in this document is meant to be used as part of the operation of the routing protocol (e.g., by replacing cost with latency or considering bandwidth as well as cost), to enhance Constrained-SPF (CSPF), or for other uses such as supplementing the data used by an ALTO server [RFC7285]. With respect to CSPF, the data distributed by IS-IS TE Metric Extensions can be used to set up, fail over, and fail back data paths using protocols such as RSVP-TE [RFC3209].
Note that the mechanisms described in this document only disseminate performance information. The methods for initially gathering that performance information, such as described in [RFC6375], or acting on it once it is distributed are outside the scope of this document. Example mechanisms to measure latency, delay variation, and loss in an MPLS network are given in [RFC6374]. While this document does not specify how the performance information should be obtained, the measurement of delay SHOULD NOT vary significantly based upon the offered traffic load. Thus, queuing delays SHOULD NOT be included in the delay measurement. For links such as Forwarding Adjacencies, care must be taken that measurement of the associated delay avoids significant queuing delay; that could be accomplished in a variety of ways, including either by measuring with a traffic class that experiences minimal queuing or by summing the measured link delays of the components of the link's path.
This document registers new IS-IS TE sub-TLVs that can be announced in the "Sub-TLVs for TLVs 22, 23, 141, 222, and 223" registry in order to distribute network-performance information. The extensions in this document build on the ones provided in IS-IS TE [RFC5305] and GMPLS [RFC4203].
IS-IS Extended Reachability TLV 22 (defined in [RFC5305]), Inter-AS Reachability Information TLV 141 (defined in [RFC5316]), and MT-ISIS TLV 222 (defined in [RFC5120]) have nested sub-TLVs that permit the TLVs to be readily extended. This document registers several sub- TLVs:
Type Description ---------------------------------------------------- 33 Unidirectional Link Delay 34 Min/Max Unidirectional Link Delay 35 Unidirectional Delay Variation 36 Unidirectional Link Loss 37 Unidirectional Residual Bandwidth 38 Unidirectional Available Bandwidth 39 Unidirectional Utilized Bandwidth
When network performance violates configurable link-local thresholds, a sub-TLV with the A bit set is advertised. These sub-TLVs could be used by the receiving node to determine whether to fail traffic to a backup path or whether to calculate an entirely new path. From an MPLS perspective, the intent of the A bit is to permit label switched path ingress nodes to determine whether the link referenced in the sub-TLV affects any of the label switched paths for which it is ingress. If they are affected, then they can determine whether those label switched paths still meet end-to-end performance objectives. If not, then the node could conceivably move affected traffic to a pre-established protection label switched path or establish a new label switched path and place the traffic in it.
If link performance then improves beyond a configurable minimum value (reuse threshold), that sub-TLV can be re-advertised with the A bit cleared. In this case, a receiving node can conceivably do whatever re-optimization (or failback) it wishes to do (including nothing).
Note that when a sub-TLV does not include the A bit, that sub-TLV cannot be used for failover purposes. The A bit was intentionally omitted from some sub-TLVs to help mitigate oscillations. See Section 5 for more information.
Consistent with existing IS-IS TE specification [RFC5305], the bandwidth advertisements defined in this document MUST be encoded as IEEE floating-point values. The delay and delay-variation advertisements defined in this document MUST be encoded as integer values. Delay values MUST be quantified in units of microseconds, packet loss MUST be quantified as a percentage of packets sent, and bandwidth MUST be sent as bytes per second. All values (except residual bandwidth) MUST be calculated as rolling averages where the averaging period MUST be a configurable period of time. See Section 5 for more information.
The use of IS-IS TE Metric Extensions sub-TLVs is not confined to the TE context. In other words, IS-IS TE Metric Extensions sub-TLVs defined in this document can also be used for computing paths in the absence of a TE subsystem.
However, as for the TE case, Interface Address and Neighbor Address sub-TLVs (IPv4 or IPv6) MUST be present. The encoding is defined in [RFC5305] for IPv4 and in [RFC6119] for IPv6.
This sub-TLV advertises the average link delay between two directly connected IS-IS neighbors. The delay advertised by this sub-TLV MUST be the delay from the local neighbor to the remote one (i.e., the forward-path latency). The format of this sub-TLV is shown in the following diagram:
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type | Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |A| RESERVED | Delay | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 1
Type: 33
Length: 4
A bit: The A bit represents the Anomalous (A) bit. The A bit is set when the measured value of this parameter exceeds its configured maximum threshold. The A bit is cleared when the measured value falls below its configured reuse threshold. If the A bit is clear, the sub-TLV represents steady-state link performance.
RESERVED: This field is reserved for future use. It MUST be set to 0 when sent and MUST be ignored when received.
Delay: This 24-bit field carries the average link delay over a configurable interval in microseconds, encoded as an integer value. When set to the maximum value 16,777,215 (16.777215 sec), then the delay is at least that value and may be larger.
This sub-TLV advertises the minimum and maximum delay values between two directly connected IS-IS neighbors. The delay advertised by this sub-TLV MUST be the delay from the local neighbor to the remote one (i.e., the forward-path latency). The format of this sub-TLV is shown in the following diagram:
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type | Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |A| RESERVED | Min Delay | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | RESERVED | Max Delay | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 2
Type: 34
Length: 8
A bit: This field represents the Anomalous (A) bit. The A bit is set when one or more measured values exceed a configured maximum threshold. The A bit is cleared when the measured value falls below its configured reuse threshold. If the A bit is clear, the sub-TLV represents steady-state link performance.
RESERVED: This field is reserved for future use. It MUST be set to 0 when sent and MUST be ignored when received.
Min Delay: This 24-bit field carries the minimum measured link delay value (in microseconds) over a configurable interval, encoded as an integer value.
Max Delay: This 24-bit field carries the maximum measured link delay value (in microseconds) over a configurable interval, encoded as an integer value.
Implementations MAY also permit the configuration of an offset value (in microseconds) to be added to the measured delay value, to facilitate the communication of operator-specific delay constraints.
It is possible for the Min and Max delay to be the same value.
When the delay value (Min or Max) is set to the maximum value 16,777,215 (16.777215 sec), then the delay is at least that value and may be larger.
This sub-TLV advertises the average link delay variation between two directly connected IS-IS neighbors. The delay variation advertised by this sub-TLV MUST be the delay from the local neighbor to the remote one (i.e., the forward-path latency). The format of this sub- TLV is shown in the following diagram:
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type | Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | RESERVED | Delay Variation | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 3
where
Type: 35
Length: 4
RESERVED: This field is reserved for future use. It MUST be set to 0 when sent and MUST be ignored when received.
Delay Variation: This 24-bit field carries the average link delay variation over a configurable interval in microseconds, encoded as an integer value. When set to 0, it has not been measured. When set to the maximum value 16,777,215 (16.777215 sec), then the delay is at least that value and may be larger.
This sub-TLV advertises the loss (as a packet percentage) between two directly connected IS-IS neighbors. The link loss advertised by this sub-TLV MUST be the packet loss from the local neighbor to the remote one (i.e., the forward-path loss). The format of this sub-TLV is shown in the following diagram:
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type | Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |A| RESERVED | Link Loss | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 4
Type: 36
Length: 4
A bit: The A bit represents the Anomalous (A) bit. The A bit is set when the measured value of this parameter exceeds its configured maximum threshold. The A bit is cleared when the measured value falls below its configured reuse threshold. If the A bit is clear, the sub-TLV represents steady-state link performance.
RESERVED: This field is reserved for future use. It MUST be set to 0 when sent and MUST be ignored when received.
Link Loss: This 24-bit field carries link packet loss as a percentage of the total traffic sent over a configurable interval. The basic unit is 0.000003%, where (2^24 - 2) is 50.331642%. This value is the highest packet-loss percentage that can be expressed (the assumption being that precision is more important on high-speed links than the ability to advertise loss rates greater than this, and that high- speed links with over 50% loss are unusable). Therefore, measured values that are larger than the field maximum SHOULD be encoded as the maximum value.
This sub-TLV advertises the residual bandwidth between two directly connected IS-IS neighbors. The residual bandwidth advertised by this sub-TLV MUST be the residual bandwidth from the system originating the Link State Advertisement (LSA) to its neighbor.
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type | Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Residual Bandwidth | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 5
Type: 37
Length: 4
Residual Bandwidth: This field carries the residual bandwidth on a link, forwarding adjacency [RFC4206], or bundled link in IEEE floating-point format with units of bytes per second. For a link or forwarding adjacency, residual bandwidth is defined to be the Maximum Bandwidth [RFC5305] minus the bandwidth currently allocated to RSVP- TE label switched paths. For a bundled link, residual bandwidth is defined to be the sum of the component link residual bandwidths.
The calculation of residual bandwidth is different than that of unreserved bandwidth [RFC5305]. Residual bandwidth subtracts tunnel reservations from maximum bandwidth (i.e., the link capacity) [RFC5305] and provides an aggregated remainder across priorities. Unreserved bandwidth, on the other hand, is subtracted from the maximum reservable bandwidth (the bandwidth that can theoretically be reserved) and provides per-priority remainders. Residual bandwidth and unreserved bandwidth [RFC5305] can be used concurrently and each has a separate use case (e.g., the former can be used for applications like Weighted ECMP while the latter can be used for call admission control).
This sub-TLV advertises the available bandwidth between two directly connected IS-IS neighbors. The available bandwidth advertised by this sub-TLV MUST be the available bandwidth from the system originating this sub-TLV. The format of this sub-TLV is shown in the following diagram:
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type | Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Available Bandwidth | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 6
Type: 38
Length: 4
Available Bandwidth: This field carries the available bandwidth on a link, forwarding adjacency, or bundled link in IEEE floating-point format with units of bytes per second. For a link or forwarding adjacency, available bandwidth is defined to be residual bandwidth (see Section 4.5) minus the measured bandwidth used for the actual forwarding of non-RSVP-TE label switched path packets. For a bundled link, available bandwidth is defined to be the sum of the component link available bandwidths minus the measured bandwidth used for the actual forwarding of non-RSVP-TE label switched path packets. For a bundled link, available bandwidth is defined to be the sum of the component link available bandwidths.
This sub-TLV advertises the bandwidth utilization between two directly connected IS-IS neighbors. The bandwidth utilization advertised by this sub-TLV MUST be the bandwidth from the system originating this sub-TLV. The format of this sub-TLV is shown in the following diagram:
0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type | Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Utilized Bandwidth | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 7
Type: 39
Length: 4
Utilized Bandwidth: This field carries the bandwidth utilization on a link, forwarding adjacency, or bundled link in IEEE floating-point format with units of bytes per second. For a link or forwarding adjacency, bandwidth utilization represents the actual utilization of the link (i.e., as measured by the advertising node). For a bundled link, bandwidth utilization is defined to be the sum of the component link bandwidth utilizations.
The values advertised in all sub-TLVs (except min/max delay and residual bandwidth) MUST represent an average over a period or be obtained by a filter that is reasonably representative of an average. For example, a rolling average is one such filter.
Min and max delay MUST each be derived in one of the following ways: by taking the lowest and/or highest measured value over a measurement interval or by making use of a filter or other technique to obtain a reasonable representation of a min and max value representative of the interval, with compensation for outliers.
The measurement interval, any filter coefficients, and any advertisement intervals MUST be configurable per sub-TLV.
In addition to the measurement intervals governing re-advertisement, implementations SHOULD provide configurable accelerated advertisement thresholds per sub-TLV, such that:
To prevent oscillations, only the high threshold or the low threshold (but not both) may be used to trigger any given sub-TLV that supports both.
Additionally, once outside the bounds of the threshold, any re-advertisement of a measurement within the bounds would remain governed solely by the measurement interval for that sub-TLV.
When link-performance values change by small amounts that fall under thresholds that would cause the announcement of a sub-TLV, implementations SHOULD suppress sub-TLV re-advertisement and/or lengthen the period within which they are refreshed.
Only the accelerated advertisement threshold mechanism described in Section 5 may shorten the re-advertisement interval. All suppression and re-advertisement interval backoff timer features SHOULD be configurable.
Sections 5 and 6 provide configurable mechanisms to bound the number of re-advertisements. Instability might occur in very large networks if measurement intervals are set low enough to overwhelm the processing of flooded information at some of the routers in the topology. Therefore, care should be taken in setting these values.
Additionally, the default measurement interval for all sub-TLVs SHOULD be 30 seconds.
Announcements MUST also be able to be throttled using configurable inter-update throttle timers. The minimum announcement periodicity is 1 announcement per second. The default value SHOULD be set to 120 seconds.
Implementations SHOULD NOT permit the inter-update timer to be lower than the measurement interval.
Furthermore, it is RECOMMENDED that any underlying performance- measurement mechanisms not include any significant buffer delay, any significant buffer-induced delay variation, or any significant loss due to buffer overflow or due to active queue management.
Implementations MUST make it possible to individually enable or disable each sub-TLV based on configuration.
Implementations SHOULD permit static configuration and/or manual override of dynamic measurements for each sub-TLV in order to simplify migration and to mitigate scenarios where dynamic measurements are not possible.
As per [RFC5305], unrecognized sub-TLVs should be silently ignored.
The sub-TLVs introduced in this document allow an operator to advertise state information of links (bandwidth, delay) that could be sensitive and that an operator may not want to disclose.
Section 7 describes a mechanism to ensure network stability when the new sub-TLVs defined in this document are advertised. Implementation SHOULD follow the described guidelines to mitigate the instability risk.
[RFC5304] describes an authentication method for IS-IS Link State PDUs that allows cryptographic authentication of IS-IS Link State PDUs.
It is anticipated that in most deployments, the IS-IS protocol is used within an infrastructure entirely under control of the same operator. However, it is worth considering that the effect of sending IS-IS Traffic Engineering sub-TLVs over insecure links could result in a man-in-the-middle attacker delaying real-time data to a given site or destination, which could negatively affect the value of the data for that site or destination. The use of Link State PDU cryptographic authentication allows mitigation the risk of man-in- the-middle attack.
IANA maintains the registry for the sub-TLVs. IANA has registered the following sub-TLVs in the "Sub-TLVs for TLVs 22, 23, 141, 222, and 223" registry:
Type Description ---------------------------------------------------- 33 Unidirectional Link Delay 34 Min/Max Unidirectional Link Delay 35 Unidirectional Delay Variation 36 Unidirectional Link Loss 37 Unidirectional Residual Bandwidth 38 Unidirectional Available Bandwidth 39 Unidirectional Utilized Bandwidth
For the original version of this document the authors recognized Ayman Soliman, Nabil Bitar, David McDysan, Edward Crabbe, Don Fedyk, Hannes Gredler, Uma Chunduri, Alvaro Retana, Brian Weis, and Barry Leiba for their contribution and review of this document.
The authors also recognized Curtis Villamizar for significant comments and direct content collaboration.
For the second version of this document the authors thank Jeff Haas for identifying and reporting the incorrect encoding of the bandwidth related sub-TLVs.
The following people contributed substantially to the content of this document and should be considered co-authors:
Alia Atlas Juniper Networks United States Email: akatlas@juniper.net Clarence Filsfils Cisco Systems Inc. Belgium Email: cfilsfil@cisco.com
Errata ID: 5293 (https://www.rfc-editor.org/errata_search.php?rfc=7810) correctly identified that in [RFC7810] the length associated with the following sub-TLVs did not match the figures associated with each:
37 Unidirectional Residual Bandwidth
38 Unidirectional Available Bandwidth
39 Unidirectional Utilized Bandwidth
Because these different implementation choices are not interoperable, it was decided that a bis version should be generated which resolved the ambiguity.
The choice made here is to omit the unused RESERVED field from these sub-TLVs and use the length of 4. This matches the corresponding advertisements specified in the equivalent OSPF specification [RFC7471] and the corresponding BGP-LS specification [I-D.ietf-idr-te-pm-bgp].
Some minor editorial corrections have also been made.