Quality of Service Parameters for Usage with the AAA Framework

This document defines a number of Quality of Service (QoS) parameters that can be reused for conveying QoS information within RADIUS and Diameter. The subsequent section give an overview of the parameters defined by this document.

The Traffic Model (TMOD) parameter is a container consisting of four sub-parameters: rate (r) bucket size (b) peak rate (p) minimum policed unit (m) The TMOD parameter is a mathematically complete way to describe the traffic source. If, for example, TMOD is set to specify bandwidth only, then set r = peak rate = p, b = large, m = large. As another example if TMOD is set for TCP traffic, then set r = average rate, b = large, p = large.

<Path Latency>, <Path Jitter>, <Path PLR>, and <Path PER> are QoS parameters describing the desired path latency, path jitter and path bit error rate respectively. The <Path Latency> parameter refers to the accumulated latency of the packet forwarding process associated with each QoS aware node along the path, where the latency is defined to be the mean packet delay added by each such node. This delay results from speed-of-light propagation delay, from packet processing limitations, or both. The mean delay reflects the variable queuing delay that may be present. The purpose of this parameter is to provide a minimum path latency for use with services which provide estimates or bounds on additional path delay . The <Path Jitter> parameter refers to the accumulated jitter of the packet forwarding process associated with each QoS aware node along the path, where the jitter is defined to be the nominal jitter added by each such node. IP packet jitter, or delay variation, is defined in Section 3.4 of RFC 3393 , (Type-P-One-way-ipdv), and where the selection function includes the packet with minimum delay such that the distribution is equivalent to 2-point delay variation in . The suggested evaluation interval is 1 minute. This jitter results from packet processing limitations, and includes any variable queuing delay which may be present. The purpose of this parameter is to provide a nominal path jitter for use with services that provide estimates or bounds on additional path delay . The procedures for collecting path jitter information are outside the scope of this document. The <Path PLR> parameter refers to the accumulated packet loss rate (PLR) of the packet forwarding process associated with each QoS aware node along the path where the PLR is defined to be the PLR added by each such node. The <Path PER> parameter refers to the accumulated packet error rate (PER) of the packet forwarding process associated with each QoS aware node, where the PER is defined to be the PER added by each such node. The <Slack Term> parameter refers to the difference between desired delay and delay obtained by using bandwidth reservation, and which is used to reduce the resource reservation for a flow . The <Preemption Priority> parameter refers to the priority of the new flow compared with the <Defending Priority> of previously admitted flows. Once a flow is admitted, the preemption priority becomes irrelevant. The <Defending Priority> parameter is used to compare with the preemption priority of new flows. For any specific flow, its preemption priority MUST always be less than or equal to the defending priority. <Admission Priority> and <RPH Priority> provide an essential way to differentiate flows for emergency services, ETS, E911, etc., and assign them a higher admission priority than normal priority flows and best-effort priority flows.

The <Excess Treatment< parameter describes how a QoS aware node will process excess traffic, that is, out-of-profile traffic. Excess traffic MAY be dropped, shaped and/or remarked.

Resource reservations might refer to a packet processing with a particular DiffServ per-hop behavior (PHB) or to a particular QoS class, e.g., Y.1541 QoS class or DiffServ-aware MPLS traffic engineering (DSTE) class type , .

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 RFC2119 .

The TMOD-1 AVP is obtained from and . The structure of the AVP is as follows:

{ TMOD-Rate-1 } { TMOD-Size-1 } { Peak-Data-Rate-1 } { Minimum-Policed-Unit-1 } ]]>

The TMOD-Rate-1 AVP (AVP Code TBD) is of type Float32 and contains the rate (r).

The TMOD-Size-1 AVP (AVP Code TBD) is of type Float32 and contains the bucket size (b).

The Peak-Data-Rate-1 AVP (AVP Code TBD) is of type Float32 and contains the peak rate (p).

The Minimum-Policed-Unit-1 AVP (AVP Code TBD) is of type Unsigned32 and contains the minimum policed unit (m).

The values r, b, and p are represented as IEEE floating point values and the sign bit MUST be zero (all values MUST be non-negative). Exponents less than 127 (i.e., 0) are prohibited. Exponents greater than 162 (i.e., positive 35) are discouraged, except for specifying a peak rate of infinity. Infinity is represented with an exponent of all ones (255) and a sign bit and mantissa of all zeroes.

A description of the semantic of the parameter values can be found in . The TMOD-2 AVP is useful in a DiffServ environment. The coding for the TMOD-2 AVP is as follows:

{ TMOD-Rate-2 } { TMOD-Size-2 } { Peak-Data-Rate-2 } { Minimum-Policed-Unit-2 } ]]>

The TMOD-Rate-2 AVP (AVP Code TBD) is of type Float32 and contains the rate (r).

The TMOD-Size-2 AVP (AVP Code TBD) is of type Float32 and contains the bucket size (b).

The Peak-Data-Rate-2 AVP (AVP Code TBD) is of type Float32 and contains the peak rate (p).

The Minimum-Policed-Unit-2 AVP (AVP Code TBD) is of type Unsigned32 and contains the minimum policed unit (m).

The semantic of the parameter values can be found in and . The Path-Latency AVP (AVP Code TBD) is of type Integer32. The composition rule for the path latency is summation with a clamp of (2**32 - 1) on the maximum value. The latencies are average values reported in units of one microsecond. A system with resolution less than one microsecond MUST set unused digits to zero. The total latency added across all QoS aware nodes along the path can range as high as (2**32)-2.

The coding for the Path-Jitter AVP is as follows:

{ Path-Jitter-STAT1 } { Path-Jitter-STAT2 } { Path-Jitter-STAT3 } { Path-Jitter-STAT4 } ]]>

The Path-Jitter-STAT1 AVP (AVP Code TBD) is of type Integer32 and contains the variance.

The Path-Jitter-STAT2 AVP (AVP Code TBD) is of type Integer32 and contains the 99.9%-ile.

The Path-Jitter-STAT3 AVP (AVP Code TBD) is of type Integer32 and contains the minimum latency.

The Path-Jitter-STAT4 AVP (AVP Code TBD) is of type Integer32 and is reserved for future use.

The Path-Jitter AVP is the combination of four statistics describing the jitter distribution with a clamp of (2**32 - 1) on the maximum of each value. The jitter STATs are reported in units of one microsecond.

The Path-PLR AVP (AVP Code TBD) is of type Float32 and contains the path packet loss ratio. The PLRs are reported in units of 10^-11. A system with resolution less than one microsecond MUST set unused digits to zero. The total PLR added across all QoS aware nodes can range as high as 10^-1.

The Path-PER AVP (AVP Code TBD) is of type Float32 and contains the path packet error ratio. The PERs are reported in units of 10^-11. A system with resolution less than one microsecond MUST set unused digits to zero. The total PER added across all QoS aware nodes can range as high as 10^-1.

The Slack-Term AVP (AVP Code TBD) is of type Integer32 and its semantic can be found in and . The Slack-Term AVP contains values measured in microseconds and its value can range from 0 to (2**32)-1 microseconds.

The Priority AVP is a grouped AVP consisting of two AVPs, the Preemption-Priority and the Defending-Priority AVP. A description of the semantic can be found in .

{ Preemption-Priority } { Defending-Priority } ]]>

The Preemption-Priority AVP (AVP Code TBD) is of type Unsigned32 and it indicates the priority of the new flow compared with the defending priority of previously admitted flows. Higher values represent higher priority.

The Defending-Priority AVP (AVP Code TBD) is of type Unsigned32. Once a flow is admitted, the preemption priority becomes irrelevant. Instead, its defending priority is used to compare with the preemption priority of new flows.

The Admission-Priority AVP (AVP Code TBD) is of type Unsigned32. The admission control priority of the flow, in terms of access to network bandwidth in order to provide higher probability of call completion to selected flows. Higher values represent higher priority. A given admission priority is encoded in this information element using the same value as when encoded in the Admission-Priority AVP defined in Section 6.2.9 of , or in the Admission Priority parameter defined in Section 3.1 of . In other words, a given value inside the Admission-Priority AVP, inside the admission priority parameter or inside the admission priority parameter, refers to the same admission priority.

The Application-Level Resource Priority (ALRP) AVP is a grouped AVP consisting of two AVPs, the ALRP-Namespace and the ALRP-Priority AVP. A description of the semantic of the parameter values can be found in and in . The coding for parameter is as follows:

{ ALRP-Namespace } { ALRP-Priority } ]]>

The ALRP-Namespace AVP (AVP Code TBD) is of type Unsigned32.

The Path-Jitter-STAT4 AVP (AVP Code TBD) is of type Unsigned32.

defines a resource priority header and established the initial registry; that registry was later extended by .

The Excess-Treatment AVP is a grouped AVP consisting of two AVPs, the Treatment and the Remark-Value AVP. The coding for the Excess-Treatment AVP is as follows:

{ Excess-Treatment-Value } [ Remark-Value ] ]]>

The Excess-Treatment-Value AVP (AVP Code TBD) is of type Enumerated and indicates how a QoS aware node should process out-of-profile traffic. The following values are currently defined:

The default treatment in case that none is specified is that there are no guarantees to excess traffic, i.e., a QoS aware node can do what it finds suitable. When the treatment is set to 'drop', all marked traffic MUST be dropped by a QoS aware node. When the treatment is set to 'shape', it is expected that QoS parameters conveyed as part of QoS-Desired are used to reshape the traffic (for example a TMOD parameter indicated as QoS desired). When the shaping causes unbounded queue growth at the shaper traffic can be dropped. When the treatment is set to 'remark', the excess treatment parameter MUST carry the remark value. For example, packets may be remarked to drop remarked to pertain to a particular QoS class. In the latter case, remarking relates to a DiffServ-type model, where packets arrive marked as belonging to a certain QoS class, and when they are identified as excess, they should then be remarked to a different QoS Class. The Remark-Value AVP carries the information used for re-marking. If 'no metering or policing is permitted' is indicated, the QoS aware node should accept the treatment set by the sender with special care so that excess traffic should not cause a problem. To request the Null Meter is especially strong, and should be used with caution.

The Remark-Value AVP (AVP Code TBD) is of type Unsigned32 and contains the DSCP value the excess traffic should be remarked to.

The PHB-Class AVP (AVP Code TBD) is of type OctetString and is two octets long. A description of the semantic of the parameter values can be found in . The coding for the values is as follows:

As prescribed in , the encoding for a single PHB is the recommended DSCP value for that PHB, left-justified in the 16 bit field, with bits 6 through 15 set to zero. The encoding for a set of PHBs is the numerically smallest of the set of encodings for the various PHBs in the set, with bit 14 set to 1. (Thus for the AF1x PHBs, the encoding is that of the AF11 PHB, with bit 14 set to 1.)

PHBs not defined by standards action, i.e., experimental or local use PHBs as allowed by . In this case an arbitrary 12 bit PHB identification code, assigned by the IANA, is placed left-justified in the 16 bit field. Bit 15 is set to 1, and bit 14 is zero for a single PHB or 1 for a set of PHBs. Bits 12 and 13 are zero.

Bits 12 and 13 are reserved either for expansion of the PHB identification code, or for other use, at some point in the future. In both cases, when a single PHBID is used to identify a set of PHBs (i.e., bit 14 is set to 1), that set of PHBs MUST constitute a PHB Scheduling Class (i.e., use of PHBs from the set MUST NOT cause intra-microflow traffic reordering when different PHBs from the set are applied to traffic in the same microflow). The set of AF1x PHBs is an example of a PHB Scheduling Class. Sets of PHBs that do not constitute a PHB Scheduling Class can be identified by using more than one PHBID. The registries needed to use already exist. Hence, no new registry needs to be created for this purpose.

The DSTE-Class-Type AVP (AVP Code TBD) is of type Unsigned32. A description of the semantic of the parameter values can be found in . Currently, the values of alues currently allowed are 0, 1, 2, 3, 4, 5, 6, 7.

The Y.1541-QoS-Class AVP (AVP Code TBD) is of type Unsigned32. A description of the semantic of the parameter values can be found in . Currently, the allowed values of the Y.1541 QoS class are 0, 1, 2, 3, 4, 5, 6, 7. Mean delay <= 100 ms, delay variation <= 50 ms, loss ratio <= 10^-3. Real-time, highly interactive applications, sensitive to jitter. Application examples include VoIP, Video Teleconference. Mean delay <= 400 ms, delay variation <= 50 ms, loss ratio <= 10^-3. Real-time, interactive applications, sensitive to jitter. Application examples include VoIP, Video Teleconference. Mean delay <= 100 ms, delay variation unspecified, loss ratio <= 10^-3. Highly interactive transaction data. Application examples include signaling. Mean delay <= 400 ms, delay variation unspecified, loss ratio <= 10^-3. Interactive transaction data. Application examples include signaling. Mean delay <= 1 sec, delay variation unspecified, loss ratio <= 10^-3. Low Loss Only applications. Application examples include short transactions, bulk data, video streaming. Mean delay unspecified, delay variation unspecified, loss ratio unspecified. Unspecified applications. Application examples include traditional applications of default IP networks. Mean delay <= 100 ms, delay variation <= 50 ms, loss ratio <= 10^-5. Applications that are highly sensitive to loss, such as television transport, high-capacity TCP transfers, and TDM circuit emulation. Mean delay <= 400 ms, delay variation <= 50 ms, loss ratio <= 10^-5. Applications that are highly sensitive to loss, such as television transport, high-capacity TCP transfers, and TDM circuit emulation.

This document is designed with extensibility in mind given that different organizations and groups are used to define their own Quality of Service parameters. This document provides an initial QoS profile with common set of QoS parameters. Ideally, these parameters should be used whenever possible but there are cases where additional parameters might be needed, or where the parameters specified in this document are used with a different semantic. In this case it is advisable to define a new QoS profile that may consist of new parameters in addition to parameters defined in this document or an entirely different set of parameters. To enable the definition of new QoS profiles a 8 octet registry is defined field that is represented by a 4-octet vendor and 4-octet specifier field. A QoS profile groups together a bunch of QoS parameters for usage in a specific environment. The vendor field indicates the type as either standards-specified or vendor-specific. If the four octets of the vendor field are 0x00000000, then the value is standards-specified and the registry is maintained by IANA, and any other value represents a vendor-specific Object Identifier (OID). IANA created registry is split into two value ranges; one range uses the "Standards Action" and the second range uses "Specification Required" allocation policy. The latter range is meant to be used by organizations outside the IETF.

This section defines the registries and initial codepoint assignments, in accordance with BCP 26 RFC 2434 . It also defines the procedural requirements to be followed by IANA in allocating new codepoints. IANA is requested to create the following registries listed in the subsections below.

The QoS Profile refers to a 64 bit long field that is represented by a 4-octet vendor and 4-octet specifier field. The vendor field indicates the type as either standards-specified or vendor-specific. If the four octets of the vendor field are 0x00000000, then the value is standards-specified and the registry is maintained by IANA, and any other value represents a vendor-specific Object Identifier (OID). The specifier field indicates the actual QoS profile. The vendor field 0x00000000 is reserved to indicate that the values in the specifier field are maintained by IANA. This document requests IANA to create such a registry and to allocate the value zero (0) for the QoS profile defined in this document. For any other vendor field, the specifier field is maintained by the vendor. For the IANA maintained QoS profiles the following allocation policy is defined: 1 to 511: Standards Action 512 to 4095: Specification Required Standards action is required to depreciate, delete, or modify existing QoS profile values in the range of 0-511 and a specification is required to depreciate, delete, or modify existing QoS profile values in the range of 512-4095.

This specification assigns the values TBD1 to TBD2 from the AVP Code namespace defined in . See for the assignment of the namespace in this specification.

The following values are allocated by this specification: Excess Treatment Value 0: drop Excess Treatment Value 1: shape Excess Treatment Value 2: remark Excess Treatment Value 3: no metering or policing is permitted Excess Treatment Values 4-63: Standards Action Excess Treatment Value 64-2^32-1: Reserved

The following values are allocated by this specification: DSTE Class Type Value 0: DSTE Class Type 0 DSTE Class Type Value 1: DSTE Class Type 1 DSTE Class Type Value 2: DSTE Class Type 2 DSTE Class Type Value 3: DSTE Class Type 3 DSTE Class Type Value 4: DSTE Class Type 4 DSTE Class Type Value 5: DSTE Class Type 5 DSTE Class Type Value 6: DSTE Class Type 6 DSTE Class Type Value 7: DSTE Class Type 7 DSTE Class Type Values 8-63: Standards Action DSTE Class Type Values 64-2^32-1: Reserved

The following values are allocated by this specification: Y.1541 QoS Class Value 0: Y.1541 QoS Class 0 Y.1541 QoS Class Value 1: Y.1541 QoS Class 1 Y.1541 QoS Class Value 2: Y.1541 QoS Class 2 Y.1541 QoS Class Value 3: Y.1541 QoS Class 3 Y.1541 QoS Class Value 4: Y.1541 QoS Class 4 Y.1541 QoS Class Value 5: Y.1541 QoS Class 5 Y.1541 QoS Class Value 6: Y.1541 QoS Class 6 Y.1541 QoS Class Value 7: Y.1541 QoS Class 7 Y.1541 QoS Class Values 8-63: Standards Action Y.1541 QoS Class Values 64-2^32-1: Reserved

The values in the ALRP-Namespace and ALRP-Priority AV{ inside the ALRP AVP take their values from the registry created by and extended with No additional actions are required by IANA by this specification.

This document does not raise any security concerns as it only defines QoS parameters.

The authors would like to thank the NSIS QSPEC authors (Cornelia Kappler, Jerry Ash, Attila Bader, Dave Oran), the NSIS working group chairs (John Loughney and Martin Stiemerling) and the former Transport Area Directors (Allison Mankin, Jon Peterson) for their help. We would like to thank Francois Le Faucheur, John Loughney, Martin Stiemerling, Dave Oran, An Nguyen, Ken Carlberg, James Polk, Lars Eggert, and Magnus Westerlund for their help with resolving problems regarding the Admission Priority and the ALRP parameter.