| Internet-Draft | CDNI Capacity Capability Advertisment Ex | March 2022 |
| Ryan, et al. | Expires 4 September 2022 | [Page] |
Open Caching architecture is a use case of Content Delivery Networks Interconnection (CDNI) in which the commercial Content Delivery Network (CDN) is the upstream CDN (uCDN) and the ISP caching layer serves as the downstream CDN (dCDN). This document supplements to the CDNI Capability Objects defined in RFC 8008 the defined capability objects structure and interface for advertisments and managment of a downstream CDN capacity.¶
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 4 September 2022.¶
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.¶
The Streaming Video Alliance [SVA] is a global association that works to solve streaming video challenges in an effort to improve end-user experience and adoption. The Open Caching Working Group [OCWG] of the Streaming Video Alliance [SVA] is focused on the delegation of video delivery requests from commerical CDNs to a caching layer at the ISP's network. Open Caching architecture is a specific use case of CDNI where the commercial CDN is the upstream CDN (uCDN) and the ISP caching layer is the downstream CDN (dCDN). While delegating traffic from one CDN to the other, it is important to make sure that an appropriate amount of traffic is delegated. In order to achive that, the SVA Open Caching Capacity Insight Specification [OC-CII] defines a feedback mechanism to inform the delegator how much traffic is appropriate to delegate. The traffic level information provided by that interface will be consumed by entities, such as the Open Caching Request router [OC-RR], to help inform that entity's traffic delegation decisions. This document defines and registers CDNI Payload Types (as defined at section 7.1 of [RFC8006]). These Payload types are used for Capability Objects added to those defined at section 4 of [RFC8008], which are required for the Open Caching Capacity Insights Interface [OC-CII].¶
For consistency with other CDNI documents this document follows the CDNI convention of uCDN (upstream CDN) and dCDN (downstream CDN) to represent the commercial CDN and ISP caching layer respectively.¶
This document registers two CDNI Payload Types (section 7.1 of [RFC8006]) for the defined capability objects:¶
The following terms are used throughout this document:¶
Additionally, this document reuses the terminology defined in [RFC6707], [RFC7336], [RFC8006], [RFC8007], [RFC8008], and [RFC8804]. Specifically, we use the following CDNI acronyms:¶
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.¶
In order to enable information exchange between a uCDN and a dCDN about acceptable levels of traffic to delegate, the following process has been defined:¶
In normal operation a uCDN will communicate with a dCDN, via an interface, to collect and understand any limits that a dCDN has set forth for traffic delegation from a uCDN. These limits will come in the form of metrics such as bits per second, requests per second, etc.. These limits can be thought of as Not to Exceed (NTE) limits.¶
The dCDN should provide access to a telemetry source, of near real time metrics, that the uCDN can use to track current usage. The uCDN should compare it's current usage to the limits the dCDN has put forth and adjust traffic delegation decisions accordingly to keep current usage under the specified limits.¶
In summary, the dCDN will provide the uCDN of limits of how much traffic it should delegate towards the dCDN and then also provide a telemetry source that is coupled to the same scope as the limit, so that the uCDN can use to track its current usage against the advertised limit. Having a limit and a corresponding telemetry source for that limit allows for a non ambiguous definition of what a particular limit means for both the uCDN and dCDN.¶
Limits that are communicated from the dCDN to the uCDN should be considered valid based on the TTL of the response. The TTL of the response will be provided by the transport mechanism for the response i.e. an HTTP Cache-Control header. The intention is that the limits would have a long lived TTL and would represent a reasonable peak utilization limit that the uCDN should target.¶
In the event that a dCDN needs to inform a uCDN of an update to a previously communicated limit, the dCDN SHOULD be able to leverage a uCDN callback endpoint to inform the uCDN of adjusted limits. The most common use case for this would be related to dCDN infrastructure issues which reduced the amount of capacity previously advertised as being available.¶
Section 5 of [RFC8008] describes the FCI Capability Advertisement Object, which contains a CDNI Capability Object as well as the capability object type (a CDNI Paylod Type). The section also defines the Capability Objects per such type. Below we define two additional Capability Objects.¶
Note: In the following sections, the term "mandatory-to-specify" is used to convey which properties MUST be included when serializing a given capability object. When mandatory-to-specify is defined as "Yes" for an individual property, it means that if the object containing that property is included in an FCI message, then the mandatory-to-specify property MUST also be included.¶
The Telemetry Capability Object is used to define a list of telemetry sources made available by the dCDN to the uCDN. In this document, Telemetry data is being defined as near real time aggregated metics of dCDN utilization, such as bits per second egress, and should be specific to the uCDN and dCDN traffic delegation relationship. Telemetry data is uniqiely defined by a source id, a metrics name, along with the footprints that are associated with an FCI.Capability advertisement. When defining a Capacity Limit, the meaning of a limit might be considered ambiguous if the uCDN and dCDN are defining current usage via different data sources. Having the dCDN provide a data source defining usage that both itself and the uCDN reference, allows a non ambiguous metric to use when determing current usage and how that compares to a limit Telemetry data is not only an important component for making informed traffic delegation decisions but also for providing visiblity to traffic that has been delegated back through to upstream providers. In situations where there are mutiple CDNi delegations, a uCDN will need to incorporate the usage information from any dCDN's it delegated to when itself is asked to provide usage information otherwise, the traffic may seem unaccounted for. An example of this situation is when a Content Provider delegates traffic directly to a CDN, and that CDN decides to further delegate that traffic to a dCDN, if the Content Provider polls the uCDN for traffic usage, if the uCDN does not integrate the Telemetry data of the dCDN it delegated to, any of the traffic the uCDN delegated to it's dCDN would become invisible to the Content Provider.¶
Property: sources¶
The Telemetry Source Object is built of an associated type, a list of exposed metrics, and type-specific configuration data.¶
Property: id¶
Property: type¶
Property: metrics¶
Property: configuration¶
Below are the listed valid telemetry source types. At the time of this draft, the type registry is limit to a single type of Generic. The intention of this type registry is to allow for future extension to reference a yet to be drafted specification for a CDNI Telemetry interface, which would standardize the definition, format,etc of Telemetry data between participants of a CDNI workflow.¶
| Source Type | Description |
|---|---|
| generic | An object which allows for advertisement of generic datasources |
The following shows an example of Telemetry Capability including 2 metrics for a source, that is scoped to a footprint.¶
"capabilities": [
{
"capability-type": "FCI.Telemetry",
"capability-value": {
"sources": [
{
"id": "capacity_metrics_region1",
"type": "generic",
"metrics": [
{
"name": "egress_5m",
"time-granularity": 300,
"data-percentile": 50,
"latency": 1500
},
{
"name": "requests_5m",
...
}
]
}
]
},
"footprints": [
<footprint objects>
]
}
]
¶
The Capacity Limits Capability Object enables the dCDN to specify traffic delegation limits to a uCDN within an FCI.Capabilities advertisement. The limits specified by the dCDN will inform the uCDN on how much traffic can be delegated to the dCDN. The limits specified by the dCDN should be considered Not To Exceed (NTE) limits. The limits should be based on near real time telemetry data that the dCDN provides to the uCDN, or in other words, for each limit that is advertised, there should also exist a telemetry source which provides data of current utilization against the particular advertised limit.¶
Property: limits¶
A CapacityLimit object is used to represent traffic limits for delegation from the uCDN towards the dCDN. By default the limit object will be scoped to the footprint associated with the FCI capability advertisement encompassing this object. The limit object can contain an optional scoping parameter which will allow specification of a limit for a subset of the encompassing footprint. Limits will be considered using a logical AND, such that a uCDN will need to ensure that all the limits are considered and honored rather than choosing the most specific only. There SHOULD be at least one limit object without an optional scope per capability-value which will define the overall limit for the footprint.¶
Property: scope¶
Property: limit-type¶
Property: id¶
Property: maximum-hard¶
Property: maximum-soft¶
Property: current¶
Property: telemetry-source¶
Below are listed the valid capacity limit types. Additional limits would need to be specified and extended into this list. The values specified here represent the types that were identified as being the most relevant metrics for the purposes of traffic delegation between CDNs.¶
| Limit Type | Units |
|---|---|
| egress | Bits per second |
| requests | Requests per second |
| storage-size | Total bytes |
| storage-objects | Count |
| sessions | Count |
| cache-size | Total bytes |
The Capacity Limit Telemetry Source Object refers to a specific metric within a Telementry Source.¶
A CapacityLimitScope object is used to define a more granular scope for a limit within the encompassing footprint. The object will define what type of scope is being declared (published host, service ids, etc..) and will then contain an array of items of the type defined (publishedhostA.cdn.com, publishedhostB.cdn.com,..). The scope types are meant to be extendible and reference already established identifiers commonly used in delivery via a CDN. The scope types MUST be mutually understood between the uCDN and dCDN.¶
Property: type¶
Property: values¶
Below are listed the valid capacity limit types. Additional limits would need to be specified and extended into this list. The values specified here represent the types that were identified as being the most relevant metrics for the purposes of traffic delegation between CDNs. The most recognized value will be the published-host, while the other values listed correspond to identifiers commonly used with commercial CDN providers that either reference a specific configuration or a logical grouping of configurations.¶
| Scope Type | Description |
|---|---|
| published-host | CDN-Domain |
| service-id | an identifier associated with a group of configurations |
| property-id | an identifier associated with a specific configuration |
The following shows an example of an FCI.CapacityLimits object.¶
"capabilities":[
{
"capability-type":"FCI.CapacityLimits",
"capability-value":{
"limits":[
{
"id":"capacity_limit_region1",
"limit-type":"egress",
"maximum-hard":50000000000,
"maximum-soft":25000000000,
"telemetry-source":{
"id":"capacity_metrics_region1",
"metric":"egress_5m"
}
},
{
"id":"capacity_limit_region1",
"scope":{
"type":"published-host",
"values":[
"serviceA.cdn.example.com"
]
},
"limit-type":"egress",
"maximum-hard":20000000000,
"maximum-soft":10000000000,
"telemetry-source":{
"id":"capacity_metrics_region1",
"metric":"egress_service2_5m"
}
},
{
"scope":{
"type":"service-id",
"Values":[
"abcd4567ef9a"
]
},
"limit-type":"egress",
"maximum-hard":30000000000,
"maximum-soft":15000000000,
"current":20000000000,
"telemetry-source":{
"id":"capacity_metrics_region3",
"metric":"egress_service3_5m"
}
}
]
},
"footprints":[
"<footprint objects>"
]
}
]
¶
similar to the type definitions described in section 7.1 of [RFC8006] as well as the types described in section 6.1 of [RFC8008].¶
This document requests the registration of the two additional payload types:¶
| Payload Type | Specification |
|---|---|
| FCI.Telemetry | RFCthis |
| FCI.CapacityLimits | RFCthis |
[RFC Editor: Please replace RFCthis with the published RFC number for this document.]¶
This specification is in accordance with the CDNI Request Routing: Footprint and Capabilities Semantics. As such, it is subject to the security and privacy considerations as defined in Section 8 of [RFC8006] and in Section 7 of [RFC8008] respectively.¶
The authors would like to express their gratitude to TBD for TBD (their guidance / contribution / reviews ...)¶