CoRE Working Group B. Silverajan Internet-Draft Tampere University of Technology Intended status: Informational T. Savolainen Expires: August 18, 2014 Nokia February 14, 2014 CoAP Communication with Alternative Transports draft-silverajan-core-coap-alternative-transports-04 Abstract CoAP is being standardised as an application level REST-based protocol. A single CoAP message is typically encapsulated and transmitted using UDP or DTLS as transports. These transports are optimal solutions for CoAP use in IP-based constrained environments and nodes. However compelling motivation exists for understanding how CoAP can operate with other transports, such as the need for M2M communication using non-IP networks, improved transport level end-to- end reliability and security, NAT and firewall traversal issues, and mechanisms possibly incurring a lower overhead to CoAP/HTTP translation gateways. This draft examines the requirements for conveying CoAP packets to end points over such alternative transports. It also provides URI solutions for representing CoAP resources over alternative transports. Status of This Memo 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 http://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 August 18, 2014. Copyright Notice Copyright (c) 2014 IETF Trust and the persons identified as the document authors. All rights reserved. Silverajan & Savolainen Expires August 18, 2014 [Page 1] Internet-Draft CoAP Alternative Transports February 2014 This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://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. Table of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 2. Usage Cases . . . . . . . . . . . . . . . . . . . . . . . . . 3 2.1. Use of SMS . . . . . . . . . . . . . . . . . . . . . . . 3 2.2. Use of WebSockets . . . . . . . . . . . . . . . . . . . . 4 2.3. Use of P2P Overlays . . . . . . . . . . . . . . . . . . . 4 2.4. Use of TCP . . . . . . . . . . . . . . . . . . . . . . . 4 2.5. Others . . . . . . . . . . . . . . . . . . . . . . . . . 4 3. Node Types based on Transport Availability . . . . . . . . . 5 4. CoAP Transport URI . . . . . . . . . . . . . . . . . . . . . 6 4.1. Design Considerations . . . . . . . . . . . . . . . . . . 7 4.2. Transport information in URI scheme . . . . . . . . . . . 8 4.3. Transport information as part of the URI authority . . . 9 4.3.1. Usage of DNS records . . . . . . . . . . . . . . . . 9 4.4. Making CoAP Resources Available over Multiple Transports 10 5. Alternative Transport Analysis and Properties . . . . . . . . 12 6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 14 7. Security Considerations . . . . . . . . . . . . . . . . . . . 14 8. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 15 9. Informative References . . . . . . . . . . . . . . . . . . . 15 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 17 1. Introduction The Constrained Application Protocol (CoAP) [I-D.ietf-core-coap] is being standardised by the CoRE WG as a lightweight, HTTP-like protocol providing a request/response model that constrained nodes can use to communicate with other nodes, be those servers, proxies, gateways, less constrained nodes, or other constrained nodes. As the Internet continues taking shape by integrating new kinds of networks, services and devices, the need for a consistent, lightweight method for resource representation, retrieval and manipulation becomes evident. Owing to its simplicity and low overhead, CoAP is a highly suitable protocol for this purpose. However, the CoAP endpoint can reside in a non-IP network, be separated from its peer by NATs and firewalls or simply has no Silverajan & Savolainen Expires August 18, 2014 [Page 2] Internet-Draft CoAP Alternative Transports February 2014 possibility to communicate over UDP. Consequently in addition to UDP, alternative transport channels for conveying CoAP packets could be considered. Extending CoAP-based resource retrieval over alternative transports allows implementations to have a significantly larger relevance in constrained as well as non-constrained networked environments. It leads to better code optimisation in constrained nodes and broader implementation reuse across new transport channels. As opposed to implementing new resource retrieval schemes, an application in an end-node can continue relying on using CoAP's REST-based method calls for this purpose, but lets CoAP's messaging sublayer take into account the change in end point identification and transport protocol. This simplifies development and memory requirements. Resource representations are also visible in an end-to-end manner for any CoAP client. The processing and computational overhead for conveying CoAP packets from one underlying transport to another, would be less than that of an application-level gateway performing individual packet-based, protocol translation between CoAP to another resource retrieval scheme. This document looks at how CoAP can be used by nodes for resource retrieval, in an end-to-end manner regardless of the transport channel available. It looks at current usage of CoAP in this regard today and provides other possible scenarios. A simple transport type classification is provided. Then we look at the various ways in which CoAP resource representations can be formulated in URIs over alternative, which express transport identification in addition to endpoint information and resource paths. Following that, a discussion of the various transport properties which influence how CoAP packets are mapped to transport level payloads, is presented. This draft however, does not discuss on application QoS requirements, user policies or network adaptation, nor does it advocate replacing the current practice of UDP-based CoAP communication. 2. Usage Cases Apart from UDP and DTLS, CoAP usage is being specified for the following environments: 2.1. Use of SMS CoAP Request and Response messages can be sent via SMS between CoAP end-points in a cellular network [I-D.becker-core-coap-sms-gprs]. A CoAP Request message can also be sent via SMS from a CoAP client to a sleeping CoAP Server as a wake-up mechanism and trigger communication via IP. The Open Mobile Alliance (OMA) specifies both UDP and SMS as Silverajan & Savolainen Expires August 18, 2014 [Page 3] Internet-Draft CoAP Alternative Transports February 2014 transports for M2M communication in cellular networks. The OMA Lightweight M2M protocol being drafted uses CoAP, and as transports, specifies both UDP binding as well as Short Message Service (SMS) bindings [OMALWM2M] for the same reason. 2.2. Use of WebSockets The WebSocket protocol is being used as a transport channel between WebSocket enabled CoAP end-points on the Internet [I-D.savolainen-core-coap-websockets]. This is particularly useful as a means for web browsers, particularly in smart devices, to allow embedded client side scripts to upgrade an existing HTTP connection to a WebSocket connection through which CoAP Request and Response messages can be exchanged with a WebSocket-enabled server. This also allows a browser containing an embedded CoAP server to behave as a WebSocket client by opening a connection to a WebSocket enabled CoAP Mirror Server to register and update its resources. 2.3. Use of P2P Overlays [I-D.jimenez-p2psip-coap-reload] specifices how CoAP nodes can use a peer-to-peer overlay network called RELOAD, as a resource caching facility for storing wireless sensor data. When a CoAP node registers its resources with a RELOAD Proxy Node (PN), the node computes a hash value from the CoAP URI and stores it as a structure together with the PN's Node ID as well as the resources. Resource retrieval by CoAP nodes is accomplished by computing the hash key over the Request URI,opening a connection to the overlay and using its message routing system to contact the CoAP server via its PN. 2.4. Use of TCP Using TCP to facilitate the traversal of CoAP Request and Response messages [I-D.bormann-core-coap-tcp]. This allows easier communication between CoAP clients and servers separated by firewalls and NATS. This also allows CoAP messages to be transported over push notification services from a notification server to a client app on a smartphone, that may previously have subscribed to receive change notifications of CoAP resource representations, possibly by using CoAP Observe-functionality [I-D.ietf-core-observe]. 2.5. Others We also envisage CoAP being extended atop other transport channels, such as: 1. The transportation of CoAP messages in Delay-Tolerant Networks [RFC4838], using the Bundle Protocol [RFC5050] for reaching Silverajan & Savolainen Expires August 18, 2014 [Page 4] Internet-Draft CoAP Alternative Transports February 2014 sensors in extremely challenging environments such as acoustic, underwater and deep space networks. 2. Any type of non-IP networks supporting constrained nodes and low- energy sensors, such as Bluetooth and Bluetooth Low Energy (either through L2CAP or with GATT) [BTCorev4.1], ZigBee, Z-Wave, 1-Wire, DASH7 and so on. 3. Instant Messaging and Social Networking channels, such as Jabber and Twitter. 3. Node Types based on Transport Availability The term "alternative transport" in this document thus far has been used to refer to any non-UDP and non-DTLS transport that can convey CoAP messages in its payload. A node however, may in fact possess the capability to utilise CoAP over multiple transport channels at its disposal, simultaneously or otherwise, at any point in time to communicate with a CoAP end-point. Such communication can obviously take place over UDP and DTLS as well. Inevitably, if two CoAP endpoints reside in distinctly separate networks with orthogonal transports, a CoAP proxy node is needed between the 2 networks so that CoAP Requests and Responses can be fulfilled properly. In [I-D.ietf-lwig-terminology], Tables 1, 3 and 4 introduced classification schemes for devices, in terms of their resource constraints, energy limitations and communication power. For this document, in addition to these capabilities, it seems useful to additionally identify devices based on their transport capabilities. +-------+----------------------------+ | Name | Transport Availability | +-------+----------------------------+ | T0 | Single transport | | | | | T1 | Multiple transports, with | | | one or more active at any | | | point in time | | | | | T2 | Multiple active transports| +-------+----------------------------+ Table 1: Classes of Available Transports Nodes falling under Type T0 possess the capability of exactly 1 type of transport channel for CoAP, at all times. These include both Silverajan & Savolainen Expires August 18, 2014 [Page 5] Internet-Draft CoAP Alternative Transports February 2014 active and sleepy nodes, which may choose to perform duty cycling for power saving. Type T1 nodes possess multiple different transports, and can perform CoAP resource retrieval or expose CoAP resources over any or all of these transports. However, not all transports are constantly active and certain transport channels and interfaces could be kept in a mostly-off state for energy-efficiency, such as in section 2.1 Type T2 nodes possess more than 1 transport, and multiple transports are simultaneously active at all times. CoAP proxy nodes which allow CoAP endpoints from disparate transports to communicate with each other, are a good example of this. 4. CoAP Transport URI Based on the usage scenarios as well as the transport classes presented in the preceding sections, this section discusses the formulation of a new URI for representing CoAP resources over alternative transports. CoAP is logically divided into 2 sublayers, whereby a request/ response layer is responsible for the protocol functionality of exchanging request and response messages, while the messaging layer is bound to UDP. These 2 sublayers are tightly coupled, both being responsible for properly encoding the header and body of the CoAP message. The COAP URI is used by both logical sublayers. For a URI that is expressed generically as URI = scheme ":" "//" authority path-abempty ["?"query ] A simple example COAP URI, "coap://server.example.com/sensors/ temperature" can be interpreted as follows: coap :// server.example.com /sensors/temperature \___/ \______ ________/ \______ _________/ | \/ \/ protocol endpoint parameterised identifier identifier resource identifier Figure 1: The CoAP URI format The resource path is explicitly expressed, and the endpoint identifier, which contains the host address at the network-level is Silverajan & Savolainen Expires August 18, 2014 [Page 6] Internet-Draft CoAP Alternative Transports February 2014 also directly bound to the scheme name containing the application- level protocol identifier. The choice of a specific transport for a scheme, however, cannot be embedded with a URI, but is defined by convention or standardisation of the protocol using the scheme. As examples, [RFC5092] defines the 'imap' scheme for the IMAP protocol over TCP, while [RFC2818] requires that the 'https' protocol identifier be used to differentiate using HTTP over TLS instead of TCP. 4.1. Design Considerations Several ways of formulating a URI which express an alternative transport binding to CoAP, can be envisioned. When such a URI is provided from an end-application to its CoAP implementation, the URI component containing transport-specific information can be checked to allow the CoAP to use the appropriate transport for a target endpoint identifier. The following design considerations influence the formulation of a new URI expressing CoAP resources over alternative transports: 1. The generic syntax for a URI is described in [RFC3986]. Conformance to RFC3986 would also obviate the need for custom URI parsers as well as resolution algorithms. In particular, how can a URI format be described in which each URI component clearly meets the syntax and percent-encoding rules described? 2. When relative references are encountered, [RFC3986] also establishes how they can be resolved against a base URI by providing an algorithm. Given this algorithm, how can a URI format be described in which relative reference resolution does not result in a target URI that loses its transport-specific information? 3. URIs are designed to uniquely identify resources. When a single resource is represented with multiple URIs (for example, an owner of two different domains deciding to serve the same resource from both), URI aliasing [WWWArchv1] occurs. Avoiding URI aliasing is considered good practice. However, when a node of Type T1 or T2 exposes a resource over multiple transport end-points, URI aliasing can occur if transport-specific information is embedded in a URI. How can a URI format be described, which avoids or at least minimises occurrences of URI aliasing? 4. The host component of current CoAP URIs can either be a numerical IP address or a fully qualified domain name (FQDN). While the usage of DNS can sometimes be useful for distinguishing transport information (see section 4.3.1), accessing DNS over some alternative transport environments may be challenging. Silverajan & Savolainen Expires August 18, 2014 [Page 7] Internet-Draft CoAP Alternative Transports February 2014 Therefore, how can a URI format be described, which expresses resources without heavy reliance on a naming infrastructure, such as DNS? A CoAP Transport URI can also be supplied as a Proxy-Uri option by a CoAP end-point to a CoAP forward proxy in order to communicate with a CoAP end-point residing in a network using a different transport. Section 6.4 of [I-D.ietf-core-coap] provides an algorithm for parsing a received URI to obtain the request's options. The following sections outline various choices for URIs that meet some of these design goals. 4.2. Transport information in URI scheme The URI scheme can follow a convention of the form "coap+", where the name of the transport is clearly and unambiguously described. Each scheme name formed in this manner can be used to differentiate the use of CoAP over an alternative transport instead of the use of CoAP over UDP or DTLS. The endpoint identifier, path and query components together with each scheme name would be used to uniquely identify each resource. Examples of such URIs are: o coap+tcp://[2001:db8::1]:5683/sensors/temperature for using CoAP over TCP o coap+sms://0015105550101/sensors/temperature for using CoAP over SMS or USSD with the endpoint identifier being a telephone subscriber number o coap+ws://www.example.com/WebSocket?/sensors/temperature for using CoAP over WebSockets with the endpoint at ws://www.example.com/ WebSocket Note: Expressing target address formats other than IPv6 literal addresses with '[' and ']' characters within this URI format, such as Bluetooth, is as yet unresolved. A new URI of this format to distinguish transport types is simple to understand and generally not dissimilar to the CoAP URI format. It is however entirely possible for each new scheme to specify its own rules for how resource and transport endpoint information can be presented. The URI meets many of the design goals described, although URI aliasing cannot be avoided for multiple transport end- points. As the usage of each alternative transport results in an entirely new scheme, IANA intervention is required for the registration of each scheme name. The registration process follows Silverajan & Savolainen Expires August 18, 2014 [Page 8] Internet-Draft CoAP Alternative Transports February 2014 the guidelines stipulated in [RFC4395], particularly where permanent URI scheme registration is concerned. 4.3. Transport information as part of the URI authority A single URI scheme, "coap-at" can be introduced, as part of an absolute URI which expresses the transport information within the authority component. One approach is to structure the component with a transport prefix to the endpoint identifier and a delimiter, such as "-endpoint_identifier". Examples of resulting URIs are: o coap-at://tcp-server.example.com/sensors/temperature o coap-at://sms-0015105550101/sensors/temperature An implementation note here is that some generic URI parsers will fail when encountering a URI such as "coap-at://tcp-[2001:db8::1]/ sensors/temperature". Consequently, an equivalent, but parseable URI from the ip6.arpa domain needs to be formulated instead. For [2001:db8::1] using TCP, this would result in the following URL: coap-at://tcp-1.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.8.b.d.0 .1.0.0.2.ip6.arpa:5683/sensors/temperature Usage of an IPv4-mapped IPv6 address such as [::ffff.192.100.0.1] can similarly be expressed with a URI from the ip6.arpa domain. This URI format allows the usage of a single scheme to represent multiple types of transport end-points. Consequently, it requires consistency in ensuring how various transport-specific endpoints are identified, as a single URI format is used. Attention must be paid towards the syntax rules and encoding for the URI host component. Additionally, against a base URI of the form "coap-at://tcp- server.example.com/sensors/temperature", resolving a relative reference, such as "//example.net/sensors/temperature" would result in the target URI "coap-at://example.net/sensors/temperature", in which transport information is lost. As with the URI presented previously in Section 4.2, URI aliasing issues persist with this format too. 4.3.1. Usage of DNS records DNS names can be used instead of IPv6 address literals to mitigate lengthy URLs referring to the ip6.arpa domain, if usage of DNS is possible. Silverajan & Savolainen Expires August 18, 2014 [Page 9] Internet-Draft CoAP Alternative Transports February 2014 DNS SRV records can also be employed to formulate a URL such as: coap-at://srv-_coap._tcp.example.com/sensors/temperature in which the "srv" prefix is used to indicate that a DNS SRV lookup should be used for _coap._tcp.example.com, where usage of CoAP over TCP is specified for example.com, and is eventually resolved to a numerical IPv4 or IPv6 address. 4.4. Making CoAP Resources Available over Multiple Transports The CoAP URI used thus far is as follows: URI = scheme ":" hier-part [ "?" query ] hier-part = "//" authority path-abempty A new URI format could be introduced, that does not possess an "authority" component, and instead defining "hier-part" to instead use another component, "path-rootless", as specified by RFC3986 [RFC3986]. The partial ABNF format of this URI would then be: URI = scheme ":" hier-part [ "?" query ] hier-part = path-rootless path-rootless = segment-nz *( "/" segment ) The full syntax of "path-rootless" is described in [RFC3986]. A generic URI defined this way would conform to the syntax of [RFC3986], while the path component can be treated as an opaque string to indicate transport types, endpoints as well as paths to CoAP resources. A single scheme can similarly be used. A constrained node that is capable of communicating over several types of transports (such as UDP, TCP and SMS) would be able to convey a single CoAP resources over multiple transports. This is also beneficial for nodes performing caching and proxying from one type of transport to another. Requesting and retrieving the same CoAP resource representation over multiple transports could be rendered possible by prefixing the transport type and endpoint identifier information to the CoAP URI. This would result in the following example representation: Silverajan & Savolainen Expires August 18, 2014 [Page 10] Internet-Draft CoAP Alternative Transports February 2014 coap-at:tcp://example.com?coap://example.com/sensors/temperature \_______ ______/ \________________ __________________/ \/ \/ Transport-specific CoAP Resource Prefix Figure 2: Prefixing a CoAP URI with TCP transport Such a representation would result in the URI being decomposed into its constituent components, with the CoAP resource residing within the query component as follows: Scheme: coap-at Path: tcp://example.com Query: coap://example.com/sensors/temperature The same CoAP resource, if requested over a WebSocket transport, would result the following URI: coap-at:ws://example.com/endpoint?coap://example.com/sensors/temperature \___________ __________/ \_______________ ___________________/ \/ \/ Transport-specific CoAP Resource Prefix Figure 3: Prefixing a CoAP URI with WebSocket transport While the transport prefix changes, the CoAP resource representation remains the same in the query component: Scheme: coap-at Path: ws://example.com/endpoint Query: coap://example.com/sensors/temperature The URI format described here overcomes the URI aliasing when multiple transports are used, by ensuring each CoAP resource representation remains the same, but is prefixed with different transports. However, against a base URI of this format, resolving Silverajan & Savolainen Expires August 18, 2014 [Page 11] Internet-Draft CoAP Alternative Transports February 2014 relative references of the form "//example.net/sensors/temperature" and "/sensor2/temperature" would again result in target URIs which lose transport-specific information. Implementation note: While square brackets are disallowed within the path component, the '[' and ']' characters needed to enclose a literal IPv6 address can be percent-encoded into their respective equivalents. The ':' character does not need to be percent-encoded. This results in a significantly simpler URI string compared to section 2.2, particularly for compressed IPv6 addresses. Additionally, the URI format can be used to specify other similar address families and formats, such as Bluetooth addresses [BTCorev4.1]. 5. Alternative Transport Analysis and Properties In this section we consider the various characteristics of alternative transports for successfully supporting various kinds of functionality for CoAP. CoAP factors lossiness, unreliability, small packet sizes and connection statelessness into its protocol logic. We discuss general transport differences and their impact on carrying CoAP packets here. Note that Properties 1, 2, and 3 are related. Property 1: Uniqueness of an end-point identifier. Transport protocols providing non-unique end-point IDs for nodes may only convey a subset of the CoAP functionality. Such nodes may only serve as CoAP servers that announce data at specific intervals to a pre-specified end point, or to a shared medium. Property 2: Unidirectional or bidirectional CoAP communication support. This refers to the ability of the CoAP end-point to use a single transport channel for both request and response messages. Depending on the scenario, having a unidirectional transport layer would mean the CoAP end-point might utilise it only for outgoing data or incoming data. Should both functionalities be needed, 2 unidirectional transport channels would be necessary. Property 3: 1:N communication support. This refers to the ability of the transport protocol to support broadcast and multicast communication. CoAP's request/response behaviour depends on unicast messaging. Group communication in CoAP is bound to using multicasting. Therefore a protocol such as TCP would be ill-suited for group communications using multicast. Anycast support, where a message is sent to a well defined Silverajan & Savolainen Expires August 18, 2014 [Page 12] Internet-Draft CoAP Alternative Transports February 2014 destination address to which several nodes belong, on the other hand, is supported by TCP. Property 4: Transport-level reliability. This refers to the ability of the transport protocol to provide a guarantee of reliability against packet loss, ensuring ordered packet delivery and having error control. When CoAP Request and Response messages are delivered over such transports, the CoAP implementations elide certain fields in the packet header. As an example, if the usage of a connection-oriented transport renders it unnecessary to specify the various CoAP message types, the Type field can be elided. For some connection-oriented transports, such as WebSockets, the version of CoAP being used can be negotiated during the opening transfer. Consequently, the Version field in CoAP packets can also be elided. Property 5: Message encoding. While parts of the CoAP payload are human readable or are transmitted in XML, JSON or SenML format, CoAP is essentially a low overhead binary protocol. Efficient transmission of such packets would therefore be met with a transport offering binary encoding support, although techniques exist in allowing binary payloads to be transferred over text-based transport protocols such as base-64 encoding. A fuller discussion about performing CoAP message encoding for SMS can be found in Appendix A.5 of [I-D.bormann-coap-misc] Property 6: Network byte order. CoAP, as well as transports based on the IP stack use a Big Endian byte order for transmitting packets over the air or wire, while transports based on Bluetooth and Zigbee prefer Little Endian byte ordering for packet fields and transmission. Any CoAP implementation that potentially uses multiple transports has to ensure correct byte ordering for the transport used. Property 7: MTU correlation with CoAP PDU size. Section 4.6 of [I-D.ietf-core-coap] discusses the avoidance of IP fragmentation by ensuring CoAP message fit into a single UDP datagram. End-points on constrained networks using 6LoWPAN may use blockwise transfers to accommodate even smaller packet sizes to avoid fragmentation. The MTU sizes for Bluetooth Low Energy as well as Classic Bluetooth are provided in Section 2.4 of [I-D.ietf-6lo-btle]. Transport MTU correlation with CoAP messages helps ensure minimal to no fragmentation at the transport layer. On the other hand, allowing a CoAP message to be delivered using a delay-tolerant transport Silverajan & Savolainen Expires August 18, 2014 [Page 13] Internet-Draft CoAP Alternative Transports February 2014 service such as the Bundle Protocol [RFC5050] would imply that the CoAP message may be fragmented (or reconstituted) along various nodes in the DTN as various sized bundles and bundle fragments. Property 8: Framing When using CoAP over a streaming transport protocol such as TCP, as opposed to datagram based protocols, care must be observed in preserving message boundaries. Commonly applied techniques at the transport level include the use of delimiting characters for this purpose as well as message framing and length prefixing. Property 9: Transport latency. A confirmable CoAP request would be retransmitted by a CoAP end-point if a response is not obtained within a certain time. A CoAP end- point registering to a Resource Directory uses a POST message that could include a lifetime value. A sleepy end-point similarly uses a lifetime value to indicate the freshness of the data to a CoAP Mirror Server. Care needs to be exercised to ensure the latency of the transport being used to carry CoAP packets is small enough not to interfere with these values for the proper operation of these functionalities. Property 10: Connection Management. A CoAP endpoint using a connection-oriented transport should be responsible for proper connection establishment prior to sending a CoAP Request message. Both communicating endpoints may monitor the connection health during the Data Transfer phase. Finally, once data transfer is complete, at least one end point should perform connection teardown gracefully. 6. IANA Considerations This memo includes no request to IANA. 7. Security Considerations While we envisage no new security risks simply from the introduction of support for alternative transports, end-applications and CoAP implementations should take note if certain transports require privacy trade-offs that may arise if identifiers such as MAC addresses or phone numbers are made public in addition to FQDNs. Silverajan & Savolainen Expires August 18, 2014 [Page 14] Internet-Draft CoAP Alternative Transports February 2014 8. Acknowledgements Feedback, ideas and ongoing discussions with Klaus Hartke, Martin Thomson, Mark Nottingham, Graham Klyne, Carsten Bormann, Markus Becker and Golnaz Karbaschi provided useful insights and ideas for this work. 9. Informative References [BTCorev4.1] BLUETOOTH Special Interest Group, "BLUETOOTH Specification Version 4.1", December 2013. [I-D.becker-core-coap-sms-gprs] Becker, M., Li, K., Poetsch, T., and K. Kuladinithi, "Transport of CoAP over SMS", draft-becker-core-coap-sms- gprs-04 (work in progress), August 2013. [I-D.bormann-coap-misc] Bormann, C. and K. Hartke, "Miscellaneous additions to CoAP", draft-bormann-coap-misc-26 (work in progress), December 2013. [I-D.bormann-core-coap-tcp] Bormann, C., "A TCP transport for CoAP", draft-bormann- core-coap-tcp-00 (work in progress), July 2013. [I-D.ietf-6lo-btle] Nieminen, J., Savolainen, T., Isomaki, M., Patil, B., Shelby, Z., and C. Gomez, "Transmission of IPv6 Packets over BLUETOOTH Low Energy", draft-ietf-6lo-btle-00 (work in progress), November 2013. [I-D.ietf-core-coap] Shelby, Z., Hartke, K., and C. Bormann, "Constrained Application Protocol (CoAP)", draft-ietf-core-coap-18 (work in progress), June 2013. [I-D.ietf-core-groupcomm] Rahman, A. and E. Dijk, "Group Communication for CoAP", draft-ietf-core-groupcomm-18 (work in progress), December 2013. [I-D.ietf-core-observe] Hartke, K., "Observing Resources in CoAP", draft-ietf- core-observe-11 (work in progress), October 2013. Silverajan & Savolainen Expires August 18, 2014 [Page 15] Internet-Draft CoAP Alternative Transports February 2014 [I-D.ietf-lwig-terminology] Bormann, C., Ersue, M., and A. Keranen, "Terminology for Constrained Node Networks", draft-ietf-lwig-terminology-06 (work in progress), December 2013. [I-D.jimenez-p2psip-coap-reload] Jimenez, J., Lopez-Vega, J., Maenpaa, J., and G. Camarillo, "A Constrained Application Protocol (CoAP) Usage for REsource LOcation And Discovery (RELOAD)", draft-jimenez-p2psip-coap-reload-03 (work in progress), February 2013. [I-D.savolainen-core-coap-websockets] Savolainen, T., Hartke, K., and B. Silverajan, "CoAP over WebSockets", draft-savolainen-core-coap-websockets-01 (work in progress), October 2013. [OMALWM2M] Open Mobile Alliance (OMA), "Lightweight Machine to Machine Technical Specification", 2013. [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. [RFC2609] Guttman, E., Perkins, C., and J. Kempf, "Service Templates and Service: Schemes", RFC 2609, June 1999. [RFC2818] Rescorla, E., "HTTP Over TLS", RFC 2818, May 2000. [RFC3986] Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform Resource Identifier (URI): Generic Syntax", STD 66, RFC 3986, January 2005. [RFC4395] Hansen, T., Hardie, T., and L. Masinter, "Guidelines and Registration Procedures for New URI Schemes", BCP 35, RFC 4395, February 2006. [RFC4838] Cerf, V., Burleigh, S., Hooke, A., Torgerson, L., Durst, R., Scott, K., Fall, K., and H. Weiss, "Delay-Tolerant Networking Architecture", RFC 4838, April 2007. [RFC5050] Scott, K. and S. Burleigh, "Bundle Protocol Specification", RFC 5050, November 2007. [RFC5092] Melnikov, A. and C. Newman, "IMAP URL Scheme", RFC 5092, November 2007. Silverajan & Savolainen Expires August 18, 2014 [Page 16] Internet-Draft CoAP Alternative Transports February 2014 [RFC6455] Fette, I. and A. Melnikov, "The WebSocket Protocol", RFC 6455, December 2011. [RFC6568] Kim, E., Kaspar, D., and JP. Vasseur, "Design and Application Spaces for IPv6 over Low-Power Wireless Personal Area Networks (6LoWPANs)", RFC 6568, April 2012. [RFC6733] Fajardo, V., Arkko, J., Loughney, J., and G. Zorn, "Diameter Base Protocol", RFC 6733, October 2012. [WWWArchv1] http://www.w3.org/TR/webarch/#uri-aliases, "Architecture of the World Wide Web, Volume One", December 2004. Authors' Addresses Bilhanan Silverajan Tampere University of Technology Korkeakoulunkatu 10 FI-33720 Tampere Finland Email: bilhanan.silverajan@tut.fi Teemu Savolainen Nokia Hermiankatu 12 D FI-33720 Tampere Finland Email: teemu.savolainen@nokia.com Silverajan & Savolainen Expires August 18, 2014 [Page 17]