Internet-Draft | CoAP Protocol Indication | October 2021 |
Amsüss | Expires 28 April 2022 | [Page] |
The Constrained Application Protocol (CoAP, [RFC7252]) is available over different transports (UDP, DTLS, TCP, TLS, WebSockets), but lacks a way to unify these addresses. This document provides terminology based on Web Linking [RFC8288] to express alternative transports available to a device, and to optimize exchanges using these.¶
This note is to be removed before publishing as an RFC.¶
Discussion of this document takes place on the Constrained RESTful Environments Working Group mailing list (core@ietf.org), which is archived at https://mailarchive.ietf.org/arch/browse/core/.¶
Source for this draft and an issue tracker can be found at https://gitlab.com/chrysn/transport-indication.¶
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 28 April 2022.¶
Copyright (c) 2021 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.¶
The Constrained Application Protocol (CoAP) provides transports mechanisms (UDP and DTLS since [RFC7252], TCP, TLS and WebSockets since [RFC8323]), with some additional being used in LwM2M [lwm2m] and even more being explored ([I-D.bormann-t2trg-slipmux], [I-D.amsuess-core-coap-over-gatt]). These are mutually incompatible on the wire, but CoAP implementations commonly support several of them, and proxies can translate between them.¶
CoAP currently lacks a way to indicate which transports are available for a given resource, and to indicate that a device is prepared to serve as a proxy; this document solves both by introducing the "has-proxy" terminology to Web Linking [RFC8288] that expresses the former through the latter. The additional "has-unique-proxy" term is introduced to negate any per-request overhead that would otherwise be introduced in the course of this.¶
CoAP also lacks a unified scheme to label a resource in a transport-indepenent way. This document does not attempt to introduce any new scheme here, or raise a scheme to be the canonical one. Instead, each host or application can pick a canonical address for its resources, and advertise other transports in addition.¶
A CoAP server that accepts forward proxy requests (i.e., requests carrying the Proxy-Scheme option) exclusively for URIs that it is also the authoritative server for is defined as a "same-host proxy".¶
The distinction between a same-host and any other proxy is only relevant on a practical, server-implementation and illustrative level; this specification does not use the distinction in normative requirements, and clients need not make the distinction at all.¶
The verb "to host" is used here in the sense of the link relation of the same name defined in [RFC6690].¶
For resources discovered via CoAP's discovery interface,
a hosting statement is typically provided by the defaults implied by [RFC6690]
where a link like </sensor/temp>
is implied to have the relation "hosts" and the anchor /
,
such that a statement "coap://hostname hosts coap://hostname/sensor/temp" can be inferred.¶
For many application it can make sense to assume that any resource has a "host" relation leading from the root path of the server without having performed that discovery explicitly.¶
[ TBD: The last paragraph could be a contentuous point; things should still work without it, and maybe that's even better because it precludes a dynamic resource created with one transport from use with another transport unless explicitly cleared. ]¶
When talking of proxy requests, this document only talks of the Proxy-Scheme option. Given that all URIs this is usable with can be expressed in decomposed CoAP URIs, the need for using the Proxy-URI option should never arise. The Proxy-URI option is still equivalent to the decomposed options, and can be used if the server supports it.¶
This document introduces provisions for the seamless use of different transport mechanisms for CoAP. Combined, these provide:¶
For all these functions, security policies must be described that allow the client to use them as securely as the original transport.¶
This document will not concern itself with changes in transport availability over time, neither in causing them ("Please take up your TCP interface, I'm going to send a firmware update") nor in advertising them (other than by the server putting suitable Max-Age values on any of its statements).¶
While CoAP can indicate the authority component of the requested URI in all requests (by means of Uri-Host), indicating the scheme of a requested URI (by means of Proxy-Scheme) makes the request implicitly a proxy request. However, this needs to be of only little practical concern: Any device can serve as a proxy for itself (a "same-host proxy") by accepting requests that carry the Proxy-Scheme option. If it is to be a well-behaved as a proxy, the device should then check whether it recognizes the name indicated in Uri-Host as one of its own (as it should if no Proxy-Scheme option accompanied it), reject the request with 5.05 when it is not recognized, and otherwise process it as it would process a request coming in on that protocol (which, for many hosts, is the same as if the option were absent completely).¶
A server can indicate support for same-host proxying (or any kind of proxying, really) by serving a Web Link with the "has-proxy" relation.¶
The semantics of a link from C to T with relations has-proxy ("C has-proxy T", <T>;rel=has-proxy;anchor="C"
)
are that for any resource R hosted on C ("C hosts R"), T is can be used as a proxy to obtain R.¶
Note that HTTP and CoAP proxies are not located at a particular resource,
but at a host in general.
Thus, a proxy URI T
in these protocols has the path "/", and no query component or fragment identifier.
This is true even for CoAP over WebSockets (which uses the concrete resource /.well-known/coap
, but that is not expressed in "coap+ws" URI).¶
Future protocols for which CoAP proxying is defined may use more components. As this is designed primarily for CoAP, proxies that perform URI mapping (as described in Section 5 of [RFC8075], especially using URI templates) are not supported in this document.¶
A constrained device at the address 2001:db8::1 that supports CoAP over TCP in addition to CoAP can self-describe like this:¶
Note that generating this discovery file needs to be dynamic based on its available addresses; only if queried using a link-local source address, it may also respond with a link-local address in the authority component of the proxy URI.¶
Unless the device makes resources discoverable at coap+tcp://[2001:db8::1]/.well-known/core
or another discovery mechanism,
clients may not assume that coap+tcp://[2001:db8::1]/sensors/temp
is a valid resource (let alone is equivalent to the other resource on the same path).
The server advertising itself like this may reject any request on CoAP-over-TCP unless they contain a Proxy-Scheme option.¶
Clients that want to access the device using CoAP-over-TCP would send a request by connecting to 2001:db8::1 TCP port 5683 and sending a GET with the options Proxy-Scheme: coap, no Uri-Host or -Port options (utilizing their default values), and the Uri-Paths "sensors" and "temp".¶
If the originally requested URI R or the application requirements demand a security mechanism is used, the client MUST only use the proxy T if the proxy can provide suitable credentials. (The hosting URI C is immaterial to these considerations).¶
For example, if the application uses the host name and a public key infrastructure and R is coap://example.com/
the proxy accessed as coap+tcp://[2001:db8::1]
still needs to provide a certificate chain for the name example.com to one of the system's trust anchors.
If, on the other hand, the application is doing a firmware update and requires any certificate from its configured firmware update issuer,
the proxy needs to provide such a firmware update certificate.¶
It is up to the client whether to use an advertised proxy transport, or (if multiple are provided) which to pick.¶
Links to proxies may be annotated with additional metadata that may help guide such a choice; defining such metadata is out of scope for this document.¶
Clients MAY switch between advertised transports as long as the document describing them is fresh; they may even do so per request. (For example, they may perform individual requests using CoAP-over-UDP, but choose CoAP-over-TCP for requests with large expected responses). When the describing document approaches expiry, the client can use the representation's ETag renew their justification for using the alternative transport.¶
While a server is at liberty to provide the same resource independently on different transports (i.e. to create aliases), it may make sense for it to pick a single scheme and authority under which it announces its resources. Using only one address helps proxies keep their caches efficient, and makes it easier for clients to avoid exploring the same server twice from different angles.¶
When there is a predominant scheme and authority through which an existing service is discovered, it makes sense to use these for the canonical addresses.¶
Otherwise,
it is suggested to use the coap
or coaps
scheme
(given that these are the most basic and widespread ones),
and the most stable usable name the host has.¶
In the Resource Directory specification [I-D.ietf-core-resource-directory], protocol negotiation was anticipated to use multiple base values. This approach was abandoned since then, as it would incur heavy URI aliasing.¶
Instead, devices can submit their has-proxy links to the Resource Directory like all their other metadata.¶
A client performing resource lookup can ask the RD to provide available (same-host-)proxies in a follow-up request
by asking for ?anchor=<the-discovered-host>&rel=has-proxy
.
The RD may also volunteer that information during resource lookups even though the has-proxy link itself does not match the search criteria.¶
[ It may be useful to define RD parameters for use with lookup here, which'd guide which available proxies to include.
For example, asking ?if=tag:example.com,sensor&proxy-links=tcp
could give as a result:
<coap://[2001:db8::1]/s>;rt=tag:example.com,sensor,<coap+tcp://[2001:db8::1]/>;rel=has-proxy;anchor="coap://[2001:db8::1]/"
]¶
A CoAP server may publish and accept multiple URIs for the same resource, for example when it accepts requests on different IP addresses that do not carry a Uri-Host option, or when it accepts requests both with and without the Uri-Host option carrying a registered name. Likewise, the server may serve the same resources on different transports. This makes for efficient requests (with no Proxy-Scheme or Uri-Host option), but in general is discouraged [aliases].¶
To make efficient requests possible without creating URI aliases that propagate, the "has-unique-proxy" specialization of the has-proxy relation is defined.¶
If a proxy is unique, it means that it unconditionally forwards to the server indicated in the link context, even if the Proxy-Scheme and Uri-Host options are elided.¶
[ The following two paragraphs are both true but follow different approaches to explaining the observable and implementable behavior; it may later be decided to focus on one or the other in this document. ]¶
While this creates URI aliasing in the requests as they are sent over the network, applications that discover a proxy this way should not "think" in terms of these URIs, but retain the originally discovered URIs (which, because Cool URIs Don't Change[cooluris], should be long-term usable). They use the proxy for as long as they have fresh knowledge of the has-(unique-)proxy statement.¶
In a way, advertising has-unique-proxy
can be viewed as a description of the link target in terms of SCHC [I-D.ietf-lpwan-coap-static-context-hc]:
In requests to that target, the link source's scheme and host are implicitly present.¶
While applications retain knowledge of the originally requested URI (even if it is not expressed in full on the wire), the original URI is not accessible to caches both within the host and on the network (the latter see Section 5). Thus, cached responses to the canonical and any aliased URI are mutually interchangable as long as both the response and the proxy statement are fresh.¶
A client MAY use a unique-proxy like a proxy and still send the Proxy-Scheme and Uri-Host option; such a client needs to recognize both relation types, as relations of the has-unique-proxy type are a specialization of has-proxy and typically don't carry the latter (redundant) annotation. [ To be evaluated -- one one hand, supporting it this way means that the server needs to identify all of its addresses and reject others. Then again, is a server that (like many now do) fully ignore any set Uri-Host correct at all? ]¶
Example:¶
It is noteworthy that when the URI reference /sensors/temperature
is resolved,
the base URI is coap://device0815.example.com
and not its coaps+ws counterpart --
as the request is still for that URI, which both the client and the server are aware of.
However, this detail is of little practical importance:
A simplistic client that uses coaps+ws://device0815.proxy.rd.example.com
as a base URI will still arrive at an identical follow-up request with no ill effect,
as long as it only uses the wrongly assembled URI for dereferencing resources,
the security context is the same,
the state is kept no longer than the has-unique-proxy statement is fresh,
and it does not (for example) pass the URI on to other devices.¶
A server that is aware of a suitable cross proxy may use the has-proxy relation to advertise that proxy. If the protocol used towards the proxy provides name indication (as CoAP over TLS or WebSockets does), or by using a large number of addresses or ports, it can even advertise a (more efficient) has-unique-proxy relation. This is particularly interesting when the advertisements are made available across transports, for example in a Resource Directory.¶
How the server can discover and trust such a proxy is out of scope for this document, but generally involves the same kind of links. In particular, a server may obtain a link to a third party proxy from an administrator as part of its configuration.¶
The proxy may advertise itself without the origin server's involvement; in that case, the client needs to take additional care (see Section 7.2).¶
A third party proxy may advertise its availability to act as a proxy for arbitrary CoAP requests. This use is not directly related to the protocol indication in other parts of this document, but sufficiently similar to warrant being described in the same document.¶
The resource type "TBDcore.proxy" can be used to describe such a proxy. The link target attribute "proxy-schemes" can be used to indicate the scheme(s) supported by the proxy, separated by the space character.¶
The considerations of Section 7.2 apply here.¶
A generic advertised proxy is always a forward proxy, and can not be advertised as a "unique" proxy as it would lack information about where to forward.¶
The use of a generic proxy can be limited to a set of devices that have permission to use it. Clients can be allowed by their network address if they can be verified, or by using explicit client authentication using the methods of [I-D.tiloca-core-oscore-capable-proxies].¶
When a resource is accessed through an "actual" proxy (i.e., a host between the client and the server, which itself may have a same-host proxy in addition to that), the proxy's choice of the upstream server is originally (i.e., without the mechanisms of this document) either configured (as in a "chain" of proxies) or determined by the request URI (where a proxy picks CoAP over TCP for a request aimed at a coap+tcp URI).¶
A proxy that has learned, by active solicitation of the information or by consulting links in its cache, that the requested URI is available through a (possibly same-host) proxy, or that has learned of advertised URI aliasings, may use that information in choosing the upstream transport, and to use responses obtained through one transport to satisfy requests on another.¶
For example, if a host at coap://h1.example.com has advertised </res>,<coap+tcp://h1.example.com>;rel=has-proxy;anchor="/"
,
then a proxy that has an active CoAP-over-TCP connection to h1.example.com
can forward an incoming request for coap://h1.example.com/res through that CoAP-over-TCP connection
with a suitable Proxy-Scheme on that connection.¶
If the host had marked the proxy point as <coap+tcp://h1.example.com>;rel=has-unique-proxy
,
then the proxy could elide the Proxy-Scheme and Uri-Host options,
and would (from the original CoAP caching rules) also be allowed
to use any fresh cache representation of coap+tcp://h1.example.com/res
to satisfy requests for coap://h1.example.com/res.¶
Clients need to strictly enforce the rules of Section 2.2. Failure to do so, in particular using a thusly announced proxy based on a certificate that attests the proxy's name, would allow attackers to circumvent the client's security expectation.¶
When security is terminated at proxies (as is in DTLS and TLS), a third party proxy can usually not satisfy this requirement; these transports are limited to same-host proxies.¶
Accepting arbitrary proxies, even with security context propagation performed properly, would attackers to redirect traffic through systems under their control. Not only does that impact availability, it also allows an attacker to observe traffic patterns.¶
This affects both OSCORE and (D)TLS, as neither protect the participants' network addresses.¶
Other than the security context propagation rules, there are no hard and general rules about when an advertised proxy is a suitable candidate. Aspects for consideration are:¶
When no direct connection is possible (e.g. because the resource to be accessed is served as coap+tcp and TCP is not implemented in the client, or because the resource's host is available on IPv6 while the client has no default IPv6 route), using a proxy is necessary if complete service disruption is to be avoided.¶
While an adversary can cause such a situation (e.g. by manipulating routing or DNS entries), such an adversary is usually already in a position to observe traffic patterns.¶
A proxy advertised by the device hosting the resource to be accessed is less risky to use than one advertised by a third party.¶
Note that in some applications, servers produce representations based on unverified user input. In such cases, and more so when multiple applications share a security context, the advertisements' provenance may need to be considered.¶
A widely published statement about a host's availability as a proxy can cause many clients to attempt to use it.¶
This is mitigated in well-behaved clients by observing the rate limits of [RFC7252], and by ceasing attempts to reach a proxy for the Max-Age of received errors.¶
Operators can further limit ill-effects by ensuring that their client systems do not needlessly use proxies advertised in an unsecured way, and by providing own proxies when their clients need them.¶
IANA is asked to add two entries into the Link Relation Type Registry last updated in [RFC8288]:¶
Relation Name | Description | Reference |
---|---|---|
has-proxy | The link target can be used as a proxy to reach the link context. | RFCthis |
has-unique-proxy | Like has-proxy, and using this proxy implies scheme and host of the target. | RFCthis |
IANA is asked to add an entry into the "Resource Type (rt=) Link Target Attribute Values" registry under the Constrained RESTful Environments (CoRE) Parameters:¶
[ The RFC Editor is asked to replace any occurrence of TBDcore.proxy with the actually registered attribute value. ]¶
Attribute Value: core.proxy¶
Description: Forward proxying services¶
Reference: [ this document ]¶
Notes: The schemes for which the proxy is usable may be indicated using the proxy-schemes target attribute as per Section 4.1 of [ this document ].¶
Since -01:¶
Since -00:¶
OSCORE interaction: [RFC8613] Section 4.1.3.2 requirements place OSCORE use in a weird category between has-proxy and has-unique-proxy (because if routing still works, the result will be correct). Not sure how to write this down properly, or whether it's actionable at all.¶
Possibly there is an inbetween category of "The host needs the Uri-Host etc. when accessed through CoAP, but because the host does not use the same OSCORE KID across different virtual hosts, it's has-unique-proxy as soon as you talk OSCORE".¶
Self-uniqueness:¶
A host that wants to indicate that it doesn't care about Uri-Host
can probably publish something like </>;rel=has-unique-proxy
to do so.¶
This'd help applications justify when they can elide the Uri-Host, even when no different protocols are involved.¶
Advertising under a stable name:¶
If a host wants to advertise its host name rather than its IP address during multicast, how does it best do that?¶
Options, when answering from 2001:db8::1 to a request to ff02::fd are:¶
<coap://myhostname/foo>,..., <coap://[2001:db8::1]>;rel=has-unique-proxy;anchor="coap://myhostname"¶
which is verbose but formally clear, and¶
</foo>,..., <coap://[2001:db8::1]>;rel=has-unique-proxy;anchor="coap://myhostname"¶
which is compatible with unaware clients, but its correctnes with respect to canonical URIs needs to be argued by the client, in this sequence¶
(Not that it'd need to happen in software in that sequence,
but that's the sequence needed to understand how the /foo
here is really coap://myhostname/foo
).¶
If CoRAL is used during discovery, a base directive or reverse relation to has-unique-proxy would make this easier.¶
This document heavily builds on concepts explored by Bill Silverajan and Mert Ocak in [I-D.silverajan-core-coap-protocol-negotiation], and together with Ines Robles and Klaus Hartke inside T2TRG.¶