RFC 7452






Internet Architecture Board (IAB)                          H. Tschofenig
Request for Comments: 7452                                      ARM Ltd.
Category: Informational                                         J. Arkko
ISSN: 2070-1721                                                D. Thaler
                                                            D. McPherson
                                                              March 2015


        Architectural Considerations in Smart Object Networking

Abstract



   The term "Internet of Things" (IoT) denotes a trend where a large
   number of embedded devices employ communication services offered by
   Internet protocols.  Many of these devices, often called "smart
   objects", are not directly operated by humans but exist as components
   in buildings or vehicles, or are spread out in the environment.
   Following the theme "Everything that can be connected will be
   connected", engineers and researchers designing smart object networks
   need to decide how to achieve this in practice.

   This document offers guidance to engineers designing Internet-
   connected smart objects.

Status of This Memo



   This document is not an Internet Standards Track specification; it is
   published for informational purposes.

   This document is a product of the Internet Architecture Board (IAB)
   and represents information that the IAB has deemed valuable to
   provide for permanent record.  It represents the consensus of the
   Internet Architecture Board (IAB).  Documents approved for
   publication by the IAB are not a candidate for any level of Internet
   Standard; see Section 2 of RFC 5741.

   Information about the current status of this document, any errata,
   and how to provide feedback on it may be obtained at
   http://www.rfc-editor.org/info/rfc7452.












Tschofenig, et al.            Informational                     [Page 1]

RFC 7452        Smart Object Architectural Considerations     March 2015


Copyright Notice



   Copyright (c) 2015 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
   (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.

Table of Contents



   1.  Introduction  . . . . . . . . . . . . . . . . . . . . . . . .   2
   2.  Smart Object Communication Patterns . . . . . . . . . . . . .   4
     2.1.  Device-to-Device Communication Pattern  . . . . . . . . .   4
     2.2.  Device-to-Cloud Communication Pattern . . . . . . . . . .   6
     2.3.  Device-to-Gateway Communication Pattern . . . . . . . . .   7
     2.4.  Back-End Data Sharing Pattern . . . . . . . . . . . . . .   9
   3.  Reuse Internet Protocols  . . . . . . . . . . . . . . . . . .  10
   4.  The Deployed Internet Matters . . . . . . . . . . . . . . . .  13
   5.  Design for Change . . . . . . . . . . . . . . . . . . . . . .  14
   6.  Security Considerations . . . . . . . . . . . . . . . . . . .  16
   7.  Privacy Considerations  . . . . . . . . . . . . . . . . . . .  18
   8.  Informative References  . . . . . . . . . . . . . . . . . . .  19
   Appendix A.  IAB Members at the Time of Approval  . . . . . . . .  23
   Acknowledgements  . . . . . . . . . . . . . . . . . . . . . . . .  23
   Authors' Addresses  . . . . . . . . . . . . . . . . . . . . . . .  24

1.  Introduction



   RFC 6574 [RFC6574] refers to smart objects as devices with
   constraints on energy, bandwidth, memory, size, cost, etc.  This is a
   fuzzy definition, as there is clearly a continuum in device
   capabilities and there is no hard line to draw between devices that
   can run Internet protocols and those that can't.

   Interconnecting smart objects with the Internet enables exciting new
   use cases and products.  An increasing number of products put the
   Internet Protocol Suite on smaller and smaller devices and offer the
   ability to process, visualize, and gain insight from the collected
   sensor data.  The network effect can be increased if the data
   collected from many different devices can be combined.







Tschofenig, et al.            Informational                     [Page 2]

RFC 7452        Smart Object Architectural Considerations     March 2015


   Developing embedded systems is a complex task, and designers must
   make a number of design decisions such as:

   o  How long is the device designed to operate?

   o  How does it interact with the physical world?  Is it a sensor or
      actuator or both?

   o  How many "owners" does it have?  One?  Many?  Is the owner likely
      to change over the lifetime of the device?

   o  Is it continuously or intermittently powered?  Does it sleep?

   o  Is it connected to a network, and if so, how?

   o  Will it be physically accessible for direct maintenance after
      deployment?  How does that affect the security model?

   While developing embedded systems is itself a complex task, designing
   Internet-connected smart objects is even harder since it requires
   expertise with Internet protocols in addition to software programming
   and hardware skills.  To simplify the development task, and thereby
   to lower the cost of developing new products and prototypes, we
   believe that reuse of prior work is essential.  Therefore, we provide
   high-level guidance on the use of Internet technology for the
   development of smart objects, and connected systems in general.

   Utilize Existing Design Patterns

      Design patterns are generally reusable solutions to a commonly
      occurring design problem (see [Gamma] for more discussion).
      Existing smart object deployments show communication patterns that
      can be reused by engineers with the benefit of lowering the design
      effort.  As discussed in the sections below, individual patterns
      also have an implication on the required interoperability between
      the different entities.  Depending on the desired functionality,
      already-existing patterns can be reused and adjusted.  Section 2
      talks about various communication patterns.

   Reuse Internet Protocols

      Most smart object deployments can make use of the already-
      standardized Internet Protocol Suite.  Internet protocols can be
      applied to almost any environment due to their generic design and
      typically offer plenty of potential for reconfiguration, which
      allows them to be tailored for the specific needs.  Section 3
      discusses this topic.




Tschofenig, et al.            Informational                     [Page 3]

RFC 7452        Smart Object Architectural Considerations     March 2015


   The Deployed Internet Matters

      When connecting smart objects to the Internet, take existing
      deployment into consideration to avoid unpleasant surprises.
      Assuming an ideal, clean-slate deployment is, in many cases, far
      too optimistic since the already-deployed infrastructure is
      convenient to use.  In Section 4, we highlight the importance of
      this topic.

   Design for Change

      The Internet infrastructure, applications, and preferred building
      blocks evolve over time.  Especially long-lived smart object
      deployments need to take this change into account, and Section 5
      is dedicated to that topic.

2.  Smart Object Communication Patterns



   This section illustrates a number of communication patterns utilized
   in the smart object environment.  It is possible that more than one
   pattern can be applied at the same time in a product.  Developers
   reusing those patterns will benefit from the experience of others as
   well as from documentation, source code, and available products.

2.1.  Device-to-Device Communication Pattern



   Figure 1 illustrates a communication pattern where two devices
   developed by different manufacturers are desired to interoperate and
   communicate directly.  To pick an example from [RFC6574], consider a
   light switch that talks to a light bulb with the requirement that
   each may be manufactured by a different company, represented as
   Manufacturer A and B.  Other cases can be found with fitness
   equipment, such as heart rate monitors and cadence sensors.

                        _,,,,    ,,,,
                       /     -'``    \
                      |  Wireless    |
                      \  Network     |
                      /               \
    ,''''''''|       /                 .       ,''''''''|
    | Light  | ------|------------------\------| Light  |
    | Bulb   |        .                 |      | Switch |
    |........'         `'-              /      |........'
                          \      _-...-`
    Manufacturer           `. ,.'              Manufacturer
        A                    `                      B

             Figure 1: Device-to-Device Communication Pattern



Tschofenig, et al.            Informational                     [Page 4]

RFC 7452        Smart Object Architectural Considerations     March 2015


   In order to fulfill the promise that devices from different
   manufacturers are able to communicate out of the box, these vendors
   need to agree on the protocol stack.  They need to make decisions
   about the following protocol-design aspects:

   o  Which physical layer(s) should be supported?  Does it use low-
      power radio technologies (e.g., Bluetooth Smart, IEEE 802.15.4)?

   o  Can devices be IPv6-only, or must they also support IPv4 for
      backward-compatibility reasons?  What IPv4-IPv6 transition
      technologies are needed?

   o  Which IP address configuration mechanism(s) is integrated into the
      device?

   o  Which communication architectures shall be supported?  Which
      devices are constrained, and what are those constraints?  Is there
      a classical client-server model or rather a peer-to-peer model?

   o  Is there a need for a service-discovery mechanism to allow users
      to discover light bulbs they have in their home or office?

   o  Which transport-layer protocol (e.g., UDP) is used for conveying
      the sensor readings/commands?

   o  Which application-layer protocol is used (for example, the
      Constrained Application Protocol (CoAP) [RFC7252])?

   o  What information model is used for expressing the different light
      levels?

   o  What data model is used to encode information?  (See [RFC3444] for
      a discussion about the difference between data models and
      information models.)

   o  Finally, security and privacy require careful thought.  This
      includes questions like: What are the security threats?  What
      security services need to be provided to deal with the identified
      threats?  Where do the security credentials come from?  At what
      layer(s) in the protocol stack should the security mechanism(s)
      reside?  What privacy implications are caused by various design
      decisions?

   This list is not meant to be exhaustive but aims to illustrate that
   for every usage scenario, many design decisions will have to be made
   in order to accommodate the constrained nature of a specific device
   in a certain usage scenario.  Standardizing such a complete solution




Tschofenig, et al.            Informational                     [Page 5]

RFC 7452        Smart Object Architectural Considerations     March 2015


   to accomplish a full level of interoperability between two devices
   manufactured by different vendors takes time, but there are obvious
   rewards for end customers and vendors.

2.2.  Device-to-Cloud Communication Pattern



   Figure 2 shows a communication pattern for uploading sensor data to
   an application service provider.  Often the application service
   provider (example.com in our illustration) also sells smart objects.
   In that case, the entire communication happens internal to the
   provider and no need for interoperability arises.  Still, it is
   useful for example.com to reuse existing specifications to lower the
   design, implementation, testing, and development effort.

   While this pattern allows using IP-based communication end to end, it
   may still lead to silos.  To prevent silos, example.com may allow
   third-party device vendors to connect to their server infrastructure
   as well.  For those cases, the protocol interface used to communicate
   with the server infrastructure needs to be made available, and
   various standards are available, such as CoAP, Datagram Transport
   Layer Security (DTLS) [RFC6347], UDP, IP, etc., as shown in Figure 2.
   A frequent concern from end users is that a change in the business
   model (or bankruptcy) of the IoT device/service provide might make
   the hardware become unusable.  Companies might consider the
   possibility of releasing their source code for the IoT device or
   allowing other IoT operating systems (plus application software) to
   be installed on the IoT device.

   Similarly, in many situations it is desirable to change which cloud
   service a device connects to, such as when an application service
   provider changes its hosting provider.  Again, standard Internet
   protocols are needed.

   Since the access networks to which various smart objects are
   connected are typically not under the control of the application
   service provider, commonly used radio technologies (such as WLAN,
   wired Ethernet, and cellular radio) together with the network access
   authentication technology have to be reused.  The same applies to
   standards used for IP address configuration.












Tschofenig, et al.            Informational                     [Page 6]

RFC 7452        Smart Object Architectural Considerations     March 2015


            .................
            |  Application  |
            |  Service      |
            |  Provider     |
            |  example.com  |
            |_______________|
                _,   .
     HTTP     ,'      `.        CoAP
     TLS    _,'          `.     DTLS
     TCP  ,'               `._  UDP
     IP -'                    - IP
    ,'''''''''''''|       ,'''''''''''''''''|
    | Device with |       | Device with     |
    | Temperature |       | Carbon Monoxide |
    | Sensor      |       | Sensor          |
    |.............'       |.................'

   TLS = Transport Layer Security

              Figure 2: Device-to-Cloud Communication Pattern

2.3.  Device-to-Gateway Communication Pattern



   The device-to-cloud communication pattern, described in Section 2.2,
   is convenient for vendors of smart objects and works well if they
   choose a radio technology that is widely deployed in the targeted
   market, such as Wi-Fi based on IEEE 802.11 for smart home use cases.
   Sometimes, less-widely-available radio technologies are needed (such
   as IEEE 802.15.4) or special application-layer functionality (e.g.,
   local authentication and authorization) has to be provided or
   interoperability is needed with legacy, non-IP-based devices.  In
   those cases, some form of gateway has to be introduced into the
   communication architecture that bridges between the different
   technologies and performs other networking and security
   functionality.  Figure 3 shows this pattern graphically.  Often,
   these gateways are provided by the same vendor that offers the IoT
   product, for example, because of the use of proprietary protocols, to
   lower the dependency on other vendors or to avoid potential
   interoperability problems.  It is expected that in the future, more
   generic gateways will be deployed to lower cost and infrastructure
   complexity for end consumers, enterprises, and industrial
   environments.  Such generic gateways are more likely to exist if IoT
   device designs make use of generic Internet protocols and not require
   application-layer gateways that translate one application-layer
   protocol to another one.  The use of application-layer gateways will,
   in general, lead to a more fragile deployment, as has been observed
   in the past with [RFC3724] and [RFC3238].




Tschofenig, et al.            Informational                     [Page 7]

RFC 7452        Smart Object Architectural Considerations     March 2015


   This communication pattern can frequently be found with smart object
   deployments that require remote configuration capabilities and real-
   time interactions.  The gateway is thereby assumed to be always
   connected to the Internet.


                .................
                |  Application  |
                |  Service      |
                |  Provider     |
                |  example.com  |
                |_______________|
                       |
                       |
                       | IPv4/IPv6
                .................
                |    Local      |
                |   Gateway     |
                |               |
                |_______________|
                   _,         .
     HTTP       ,'              `.         CoAP
     TLS      _,' Bluetooth Smart  `.      DTLS
     TCP    ,'     IEEE 802.11       `._   UDP
     IPv6 -'       IEEE 802.15.4         - IPv6
    ,'''''''''''''|          ,'''''''''''''''''|
    | Device with |          | Device with     |
    | Temperature |          | Carbon Monoxide |
    | Sensor      |          | Sensor          |
    |.............'          |.................'

             Figure 3: Device-to-Gateway Communication Pattern

   If the gateway is mobile, such as when the gateway is a smartphone,
   connectivity between the devices and the Internet may be
   intermittent.  This limits the applicability of such a communication
   pattern but is nevertheless very common with wearables and other IoT
   devices that do not need always-on Internet or real-time Internet
   connectivity.  From an interoperability point of view, it is worth
   noting that smartphones, with their sophisticated software update
   mechanism via app stores, allow new functionality to be updated
   regularly at the smartphone and sometimes even at the IoT device.
   With special apps that are tailored to each specific IoT device,
   interoperability is mainly a concern with regard to the lower layers
   of the protocol stack, such as the radio interface, and less so at
   the application layer (if users are willing to download a new app for
   each IoT device).




Tschofenig, et al.            Informational                     [Page 8]

RFC 7452        Smart Object Architectural Considerations     March 2015


   It is also worth pointing out that a gateway allows supporting both
   IPv6 and IPv4 (for compatibility with legacy application service
   providers) externally, while allowing devices to be IPv6-only to
   reduce footprint requirements.  If devices do not have the resources
   to support both IPv4 and IPv6 themselves, being IPv6-only (rather
   than IPv4-only) with a gateway enables the most flexibility, avoiding
   the need to update devices to support IPv6 later, whereas IPv4
   address exhaustion makes it ill-suited to scale to smart object
   networks.  See [RFC6540] for further discussion.

2.4.  Back-End Data Sharing Pattern



   The device-to-cloud pattern often leads to silos; IoT devices upload
   data only to a single application service provider.  However, users
   often demand the ability to export and to analyze data in combination
   with data from other sources.  Hence, the desire for granting access
   to the uploaded sensor data to third parties arises.  This design is
   shown in Figure 4.  This pattern is known from the Web in case of
   mashups and is, therefore, reapplied to the smart object context.  To
   offer familiarity for developers, typically a RESTful API design in
   combination with a federated authentication and authorization
   technology (like OAuth 2.0 [RFC6749]) is reused.  While this offers
   reuse at the level of building blocks, the entire protocol stack
   (including the information/data model and RESTful Web APIs) is often
   not standardized.


























Tschofenig, et al.            Informational                     [Page 9]

RFC 7452        Smart Object Architectural Considerations     March 2015


                                              .................
                                              |  Application  |
                                             .|  Service      |
                                          ,-` |  Provider     |
                                        .`    | b-example.com |
                                     ,-`      |_______________|
                                   .`
             .................  ,-`
             |  Application  |-` HTTPS
             |  Service      |   OAuth 2.0
             |  Provider     |   JSON
             |  example.com  |-,
             |_______________|  '.
                  _,              `',
                ,'                   '.
             _,' CoAP or               `',    .................
           ,'   HTTP                      '.  |  Application  |
         -'                                 `'|  Service      |
      ,''''''''|                              |  Provider     |
      | Light  |                              | c-example.com |
      | Sensor |                              |_______________|
      |........'

                  Figure 4: Back-End Data Sharing Pattern

3.  Reuse Internet Protocols



   When discussing the need for reuse of available standards versus
   extending or redesigning protocols, it is useful to look back at the
   criteria for success of the Internet.

   RFC 1958 [RFC1958] provides lessons from the early days of the
   Internet and says:

      The Internet and its architecture have grown in evolutionary
      fashion from modest beginnings, rather than from a Grand Plan.

   And adds:

      A good analogy for the development of the Internet is that of
      constantly renewing the individual streets and buildings of a
      city, rather than razing the city and rebuilding it.

   Yet, because building very small, battery-powered devices is
   challenging, it may be difficult to resist the temptation to build
   solutions tailored to specific applications, or even to redesign
   networks from scratch to suit a particular application.




Tschofenig, et al.            Informational                    [Page 10]

RFC 7452        Smart Object Architectural Considerations     March 2015


   While developing consensus-based standards in an open and transparent
   process takes longer than developing proprietary solutions, the
   resulting solutions often remain relevant over a longer period of
   time.

   RFC 1263 [RFC1263] considers protocol-design strategy and the
   decision to design new protocols or to use existing protocols in a
   non-backward compatible way:

      We hope to be able to design and distribute protocols in less time
      than it takes a standards committee to agree on an acceptable
      meeting time.  This is inevitable because the basic problem with
      networking is the standardization process.  Over the last several
      years, there has been a push in the research community for
      lightweight protocols, when in fact what is needed are lightweight
      standards.  Also note that we have not proposed to implement some
      entirely new set of 'superior' communications protocols, we have
      simply proposed a system for making necessary changes to the
      existing protocol suites fast enough to keep up with the
      underlying change in the network.  In fact, the first standards
      organization that realizes that the primary impediment to
      standardization is poor logistical support will probably win.

   While [RFC1263] was written in 1991 when the standardization process
   was more lightweight than today, these thoughts remain relevant in
   smart object development.

   Interestingly, a large number of already-standardized protocols are
   relevant for smart object deployments.  RFC 6272 [RFC6272], for
   example, made the attempt to identify relevant IETF specifications
   for use in smart grids.

   Still, many commercial products contain proprietary or industry-
   specific protocol mechanisms, and researchers have made several
   attempts to design new architectures for the entire Internet system.
   There are several architectural concerns that deserve to be
   highlighted:

   Vertical Profiles

      The discussions at the IAB workshop (see Section 3.1.2 of
      [RFC6574]) revealed the preference of many participants to develop
      domain-specific profiles that select a minimum subset of protocols
      needed for a specific operating environment.  Various
      standardization organizations and industry fora are currently
      engaged in activities of defining their preferred profile(s).





Tschofenig, et al.            Informational                    [Page 11]

RFC 7452        Smart Object Architectural Considerations     March 2015


      Ultimately, however, the number of domains where smart objects can
      be used is essentially unbounded.  There is also an ever-evolving
      set of protocols and protocol extensions.

      However, merely changing the networking protocol to IP does not
      necessarily bring the kinds of benefits that industries are
      looking for in their evolving smart object deployments.  In
      particular, a profile is rigid and leaves little room for
      interoperability among slightly differing or competing technology
      variations.  As an example, Layer 1 through 7 type profiles do not
      account for the possibility that some devices may use different
      physical media than others, and that in such situations, a simple
      router could still provide an ability to communicate between the
      parties.

   Industry-Specific Solutions

      The Internet Protocol Suite is more extensive than merely the use
      of IP.  Often, significant benefits can be gained from using
      additional, widely available, generic technologies, such as the
      Web. Benefits from using these kinds of tools include access to a
      large available workforce, software, and education already geared
      towards employing the technology.

   Tight Coupling

      Many applications are built around a specific set of servers,
      devices, and users.  However, often the same data and devices
      could be useful for many purposes, some of which may not be easily
      identifiable at the time the devices are deployed.

   In addition to the architectural concerns, developing new protocols
   and mechanisms is generally more risky and expensive than reusing
   existing standards, due to the additional costs involved in design,
   implementation, testing, and deployment.  Secondary costs, such as
   the training of technical staff and, in the worst case, the training
   of end users, can be substantial.

   As a result, while there are some cases where specific solutions are
   needed, the benefits of general-purpose technology are often
   compelling, be it choosing IP over some more specific communication
   mechanism, a widely deployed link layer (such as wireless LAN) over a
   more specific one, web technology over application-specific
   protocols, and so on.

   However, when employing these technologies, it is important to
   embrace them in their entirety, allowing for the architectural
   flexibility that is built into them.  As an example, it rarely makes



Tschofenig, et al.            Informational                    [Page 12]

RFC 7452        Smart Object Architectural Considerations     March 2015


   sense to limit communications to on-link or to specific media.
   Design your applications so that the participating devices can easily
   interact with multiple other applications.

4.  The Deployed Internet Matters



   Despite the applicability of Internet protocols for smart objects,
   picking the specific protocols for a particular use case can be
   tricky.  As the Internet has evolved, certain protocols and protocol
   extensions have become the norm, and others have become difficult to
   use in all circumstances.

   Taking into account these constraints is particularly important for
   smart objects, as there is often a desire to employ specific features
   to support smart object communication.  For instance, from a pure
   protocol-specification perspective, some transport protocols may be
   more desirable than others.  These constraints apply both to the use
   of existing protocols as well as designing new ones on top of the
   Internet protocol stack.

   The following list illustrates a few of those constraints, but every
   communication protocol comes with its own challenges.

   In 2005, Fonseca, et al.  [IPoptions] studied the usage of IP
   options-enabled packets in the Internet and found that overall,
   approximately half of Internet paths drop packets with options,
   making extensions using IP options "less ideal" for extending IP.

   In 2010, Honda, et al.  [HomeGateway] tested 34 different home
   gateways regarding their packet dropping policy of UDP, TCP, the
   Datagram Congestion Control Protocol (DCCP), the Stream Control
   Transmission Protocol (SCTP), ICMP, and various timeout behavior.
   For example, more than half of the tested devices do not conform to
   the IETF-recommended timeouts for UDP, and for TCP the measured
   timeouts are highly variable, ranging from less than 4 minutes to
   longer than 25 hours.  For NAT traversal of DCCP and SCTP, the
   situation is poor.  None of the tested devices, for example, allowed
   establishing a DCCP connection.

   In 2011, the behavior of networks with regard to various TCP
   extensions was tested in [TCPextensions]: "From our results we
   conclude that the middleboxes implementing layer 4 functionality are
   very common -- at least 25% of paths interfered with TCP in some way
   beyond basic firewalling."

   Extending protocols to fulfill new uses and to add new functionality
   may range from very easy to difficult, as [RFC6709] explains in great
   detail.  A challenge many protocol designers are facing is to ensure



Tschofenig, et al.            Informational                    [Page 13]

RFC 7452        Smart Object Architectural Considerations     March 2015


   incremental deployability and interoperability with incumbent
   elements in a number of areas.  In various cases, the effort it takes
   to design incrementally deployable protocols has not been taken
   seriously enough at the outset.  RFC 5218 on "What Makes For a
   Successful Protocol" [RFC5218] defines wildly successful protocols as
   protocols that are widely deployed beyond their envisioned use cases.

   As these examples illustrate, protocol architects have to take
   developments in the greater Internet into account, as not all
   features can be expected to be usable in all environments.  For
   instance, middleboxes [RFC3234] complicate the use of extensions in
   basic IP protocols and transport layers.

   RFC 1958 [RFC1958] considers this aspect and says "... the community
   believes that the goal is connectivity, the tool is the Internet
   Protocol, and the intelligence is end to end rather than hidden in
   the network."  This statement is challenged more than ever with the
   perceived need to develop intermediaries interacting with less
   intelligent end devices.  However, RFC 3724 [RFC3724] has this to say
   about this crucial aspect: "One desirable consequence of the
   end-to-end principle is protection of innovation.  Requiring
   modification in the network in order to deploy new services is still
   typically more difficult than modifying end nodes."  Even this
   statement will become challenged, as large numbers of devices are
   deployed, and it indeed might be the case that changing those devices
   will be hard.  But RFC 4924 [RFC4924] adds that a network that does
   not filter or transform the data that it carries may be said to be
   "transparent" or "oblivious" to the content of packets.  Networks
   that provide oblivious transport enable the deployment of new
   services without requiring changes to the core.  It is this
   flexibility that is perhaps both the Internet's most essential
   characteristic as well as one of the most important contributors to
   its success.

5.  Design for Change



   How to embrace rapid innovation and at the same time accomplish a
   high level of interoperability is one of the key aspects for
   competing in the marketplace.  RFC 1263 [RFC1263] points out that
   "protocol change happens and is currently happening at a very
   respectable clip...We simply propose [for engineers developing the
   technology] to explicitly deal with the changes rather [than] keep
   trying to hold back the flood."

   In [Tussles], Clark, et al. suggest to "design for variation in
   outcome, so that the outcome can be different in different places,
   and the tussle takes place within the design, not by distorting or
   violating it.  Do not design so as to dictate the outcome.  Rigid



Tschofenig, et al.            Informational                    [Page 14]

RFC 7452        Smart Object Architectural Considerations     March 2015


   designs will be broken; designs that permit variation will flex under
   pressure and survive."  The term "tussle" refers to the process
   whereby different parties, which are part of the Internet milieu and
   have interests that may be adverse to each other, adapt their mix of
   mechanisms to try to achieve their conflicting goals, and others
   respond by adapting the mechanisms to push back.

   In order to accomplish this, Clark, et al. suggest to:

   1.  Break complex systems into modular parts, so that one tussle does
       not spill over and distort unrelated issues.

   2.  Design for choice to permit the different players to express
       their preferences.  Choice often requires open interfaces.

   The main challenge with the suggested approach is predicting how
   conflicts among the different players will evolve.  Since tussles
   evolve over time, there will be changes to the architecture, too.  It
   is certainly difficult to pick the right set of building blocks and
   to develop a communication architecture that will last a long time,
   and many smart object deployments are envisioned to be rather long
   lived.

   Luckily, the design of the system does not need to be cast in stone
   during the design phase.  It may adjust dynamically since many of the
   protocols allow for configurability and dynamic discovery.  But,
   ultimately, software update mechanisms may provide the flexibility
   needed to deal with more substantial changes.

   A solid software update mechanism is needed not only for dealing with
   the changing Internet communication environment and for
   interoperability improvements but also for adding new features and
   for fixing security bugs.  This approach may appear to be in conflict
   with classes of severely restricted devices since, in addition to a
   software update mechanism, spare flash and RAM capacity is needed.
   It is, however, a trade-off worth thinking about since better product
   support comes with a price.

   As technology keeps advancing, the constraints that technology places
   on devices evolve as well.  Microelectronics have become more capable
   as time goes by, often making it possible for new devices to be both
   less expensive and more capable than their predecessors.  This trend
   can, however, be in some cases offset by the desire to embed
   communications technology in even smaller and cheaper objects.  But
   it is important to design communications technology not just for
   today's constraints but also for tomorrow's.  This is particularly
   important since the cost of a product is not only determined by the




Tschofenig, et al.            Informational                    [Page 15]

RFC 7452        Smart Object Architectural Considerations     March 2015


   cost of hardware but also by the cost of not reusing already-
   available protocol stacks and software libraries by developing custom
   solutions.

   Software updates are common in operating systems and application
   programs today.  Without them, most devices would pose a latent risk
   to the Internet at large.  Arguably, the JavaScript-based web employs
   a very rapid software update mechanism with code being provided by
   many different parties (e.g., by websites loaded into the browser or
   by smartphone apps).

6.  Security Considerations



   Security is often even more important for smart objects than for more
   traditional computing systems, since interacting directly with the
   physical world can present greater dangers, and smart objects often
   operate autonomously without any human interaction for a long time
   period.  The problem is compounded by the fact that there are often
   fewer resources available in constrained devices to actually
   implement security (e.g., see the discussion of "Class 0 devices" in
   Section 3 of [RFC7228]).  As such, it is critical to design for
   security, taking into account a number of key considerations:

   o  A key part of any smart object design is the problem of how to
      establish trust for a smart object.  Typically, bootstrapping
      trust involves giving the device the credentials it needs to
      operate within a larger network of devices or services.

   o  Smart objects will, in many cases, be deployed in places where
      additional physical security is difficult or impossible.
      Designers should take into account that any such device can and
      will be compromised by an attacker with direct physical access.
      Thus, trust models should distinguish between devices susceptible
      to physical compromise and devices with some level of physical
      security.  Physical attacks, such as timing, power analysis, and
      glitching, are commonly applied to extract secrets
      [PhysicalAttacks].

   o  Smart objects will, in many cases, be deployed as collections of
      identical or near identical devices.  Protocols should be designed
      so that a compromise of a single device does not result in
      compromise of the entire collection, especially since the
      compromise of a large number of devices can enable additional
      attacks such as a distributed denial of service.  Sharing secret
      keys across an entire product family is, therefore, also
      problematic since compromise of a single device might leave all
      devices from that product family vulnerable.




Tschofenig, et al.            Informational                    [Page 16]

RFC 7452        Smart Object Architectural Considerations     March 2015


   o  Smart objects will, in many cases, be deployed in ways that the
      designer never considered.  Designers should either seek to
      minimize the impact of misuse of their systems and devices or
      implement controls to prevent such misuse where applicable.

   o  It is anticipated that smart objects will be deployed with a long
      (e.g., 5-40 years) life cycle.  Any security mechanism chosen at
      the outset may not be "good enough" for the full lifespan of the
      device.  Thus, long-lived devices should start with good security
      and provide a path to deploy new security mechanisms over the
      lifetime of the device.

   o  Security protocols often rely on random numbers, and offering
      randomness in embedded devices is challenging.  For this reason,
      it is important to consider the use of hardware-based random
      number generators during early states of the design process.

   A more detailed security discussion can be found in the "Report from
   the Smart Object Security Workshop" [RFC7397] that was held prior to
   the IETF meeting in Paris, March 2012, and in the report from the
   National Science Foundation's "Cybersecurity Ideas Lab" workshop
   [NSF] that was held in February 2014.  For example, [NSF] includes,
   among other recommendations, these recommendations specific to the
   Internet of Things:

      Enhance the Security of the Internet of Things by Identifying
      Enclaves: The security challenges posed by the emerging Internet
      of Things should be addressed now, to prepare before it is fully
      upon us.  By identifying specific use segments, or "enclaves",
      Internet of Things infrastructure stakeholders can address the
      security requirements and devise event remediations for that
      enclave.

      Create a Framework for Managing Software Updates: The Internet of
      Things will challenge our current channels for distributing
      security updates.  An environment must be developed for
      distributing security patches that scales to a world where almost
      everything is connected to the Internet and many "things" are
      largely unattended.

   Finally, we reiterate that use of standards that have gotten wide
   review can often avoid a number of security issues that could
   otherwise arise.  Section 3.3 of [RFC6574] reminds us about the IETF
   work style regarding security:







Tschofenig, et al.            Informational                    [Page 17]

RFC 7452        Smart Object Architectural Considerations     March 2015


      In the development of smart object applications, as with any other
      protocol application solution, security has to be considered early
      in the design process.  As such, the recommendations currently
      provided to IETF protocol architects, such as RFC 3552 [RFC3552],
      and RFC 4101 [RFC4101], apply also to the smart object space.

   In the IETF, security functionality is incorporated into each
   protocol as appropriate, to deal with threats that are specific to
   them.  It is extremely unlikely that there is a one-size-fits-all
   security solution given the large number of choices for the 'right'
   protocol architecture (particularly at the application layer).  For
   this purpose, [RFC6272] offers a survey of IETF security mechanisms
   instead of suggesting a preferred one.

7.  Privacy Considerations



   This document mainly focuses on an engineering audience, i.e., those
   who are designing smart object protocols and architectures.  Since
   there is no value-free design, privacy-related decisions also have to
   be made, even if they are just implicit in the reuse of certain
   technologies.  RFC 6973 [RFC6973] and the threat model in
   [CONFIDENTIALITY] were written as guidance specifically for that
   audience and are also applicable to the smart object context.

   For those looking at privacy from a deployment point of view, the
   following additional guidelines are suggested:

   Transparency:  Transparency of data collection and processing is key
      to avoid unpleasant surprises for owners and users of smart
      objects.  Users and impacted parties must be put in a position to
      understand what items of personal data concerning them are
      collected and stored, as well for what purposes they are sought.

   Data Collection / Use Limitation:  Smart objects should only store
      personal data that is adequate, relevant, and not excessive in
      relation to the purpose(s) for which they are processed.  The use
      of anonymized data should be preferred wherever possible.

   Data Access:  Before deployment starts, it is necessary to consider
      who can access personal data collected by smart objects and under
      which conditions.  Appropriate and clear procedures should be
      established in order to allow data subjects to properly exercise
      their rights.








Tschofenig, et al.            Informational                    [Page 18]

RFC 7452        Smart Object Architectural Considerations     March 2015


   Data Security:   Standardized data security measures to prevent
      unlawful access, alteration, or loss of smart object data need to
      be defined and deployed.  Robust cryptographic techniques and
      proper authentication frameworks have to be used to limit the risk
      of unintended data transfers or unauthorized access.

   A more detailed treatment of privacy considerations that extend
   beyond engineering can be found in a publication from the Article 29
   Working Party [WP223].

8.  Informative References



   [CONFIDENTIALITY]
              Barnes, R., Schneier, B., Jennings, C., Hardie, T.,
              Trammell, B., Huitema, C., and D. Borkmann,
              "Confidentiality in the Face of Pervasive Surveillance: A
              Threat Model and Problem Statement", Work in Progress,
              draft-iab-privsec-confidentiality-threat-04, March 2015.

   [Gamma]    Gamma, E., "Design Patterns: Elements of Reusable Object-
              Oriented Software", 1995.

   [HomeGateway]
              Eggert, L., "An Experimental Study of Home Gateway
              Characteristics", In Proceedings of the 10th annual
              Internet Measurement Conference, 2010,
              <http://eggert.org/papers/2010-imc-hgw-study.pdf>.

   [IPoptions]
              Fonseca, R., Porter, G., Katz, R., Shenker, S., and I.
              Stoica, "IP options are not an option", Technical Report
              UCB/EECS2005-24, 2005,
              <http://citeseer.ist.psu.edu/viewdoc/
              summary?doi=10.1.1.123.4251>.

   [NSF]      National Science Foundation, "Interdisciplinary Pathways
              towards a More Secure Internet", A report on the NSF-
              sponsored Cybersecurity Ideas Lab held in Arlington,
              Virginia, February 2014, <http://www.nsf.gov/cise/news/
              CybersecurityIdeasLab_July2014.pdf>.

   [PhysicalAttacks]
              Koeune, F. and F. Standaert, "A Tutorial on Physical
              Security and Side-Channel Attacks", in Foundations of
              Security Analysis and Design III: FOSAD 2004/2005 Tutorial
              Lectures; Lecture Notes in Computer Science, Vol. 3655,
              pp. 78-108, September 2005,
              <http://link.springer.com/chapter/10.1007%2F11554578_3>.



Tschofenig, et al.            Informational                    [Page 19]

RFC 7452        Smart Object Architectural Considerations     March 2015


   [RFC1263]  O'Malley, S. and L. Peterson, "TCP Extensions Considered
              Harmful", RFC 1263, October 1991,
              <http://www.rfc-editor.org/info/rfc1263>.

   [RFC1958]  Carpenter, B., "Architectural Principles of the Internet",
              RFC 1958, June 1996,
              <http://www.rfc-editor.org/info/rfc1958>.

   [RFC3234]  Carpenter, B. and S. Brim, "Middleboxes: Taxonomy and
              Issues", RFC 3234, February 2002,
              <http://www.rfc-editor.org/info/rfc3234>.

   [RFC3238]  Floyd, S. and L. Daigle, "IAB Architectural and Policy
              Considerations for Open Pluggable Edge Services", RFC
              3238, January 2002,
              <http://www.rfc-editor.org/info/rfc3238>.

   [RFC3444]  Pras, A. and J. Schoenwaelder, "On the Difference between
              Information Models and Data Models", RFC 3444, January
              2003, <http://www.rfc-editor.org/info/rfc3444>.

   [RFC3552]  Rescorla, E. and B. Korver, "Guidelines for Writing RFC
              Text on Security Considerations", BCP 72, RFC 3552, July
              2003, <http://www.rfc-editor.org/info/rfc3552>.

   [RFC3724]  Kempf, J., Austein, R., and IAB, "The Rise of the Middle
              and the Future of End-to-End: Reflections on the Evolution
              of the Internet Architecture", RFC 3724, March 2004,
              <http://www.rfc-editor.org/info/rfc3724>.

   [RFC4101]  Rescorla, E. and IAB, "Writing Protocol Models", RFC 4101,
              June 2005, <http://www.rfc-editor.org/info/rfc4101>.

   [RFC4924]  Aboba, B. and E. Davies, "Reflections on Internet
              Transparency", RFC 4924, July 2007,
              <http://www.rfc-editor.org/info/rfc4924>.

   [RFC5218]  Thaler, D. and B. Aboba, "What Makes For a Successful
              Protocol?", RFC 5218, July 2008,
              <http://www.rfc-editor.org/info/rfc5218>.

   [RFC6272]  Baker, F. and D. Meyer, "Internet Protocols for the Smart
              Grid", RFC 6272, June 2011,
              <http://www.rfc-editor.org/info/rfc6272>.

   [RFC6347]  Rescorla, E. and N. Modadugu, "Datagram Transport Layer
              Security Version 1.2", RFC 6347, January 2012,
              <http://www.rfc-editor.org/info/rfc6347>.



Tschofenig, et al.            Informational                    [Page 20]

RFC 7452        Smart Object Architectural Considerations     March 2015


   [RFC6540]  George, W., Donley, C., Liljenstolpe, C., and L. Howard,
              "IPv6 Support Required for All IP-Capable Nodes", BCP 177,
              RFC 6540, April 2012,
              <http://www.rfc-editor.org/info/rfc6540>.

   [RFC6574]  Tschofenig, H. and J. Arkko, "Report from the Smart Object
              Workshop", RFC 6574, April 2012,
              <http://www.rfc-editor.org/info/rfc6574>.

   [RFC6709]  Carpenter, B., Aboba, B., and S. Cheshire, "Design
              Considerations for Protocol Extensions", RFC 6709,
              September 2012, <http://www.rfc-editor.org/info/rfc6709>.

   [RFC6749]  Hardt, D., "The OAuth 2.0 Authorization Framework", RFC
              6749, October 2012,
              <http://www.rfc-editor.org/info/rfc6749>.

   [RFC6973]  Cooper, A., Tschofenig, H., Aboba, B., Peterson, J.,
              Morris, J., Hansen, M., and R. Smith, "Privacy
              Considerations for Internet Protocols", RFC 6973, July
              2013, <http://www.rfc-editor.org/info/rfc6973>.

   [RFC7228]  Bormann, C., Ersue, M., and A. Keranen, "Terminology for
              Constrained-Node Networks", RFC 7228, May 2014,
              <http://www.rfc-editor.org/info/rfc7228>.

   [RFC7252]  Shelby, Z., Hartke, K., and C. Bormann, "The Constrained
              Application Protocol (CoAP)", RFC 7252, June 2014,
              <http://www.rfc-editor.org/info/rfc7252>.

   [RFC7397]  Gilger, J. and H. Tschofenig, "Report from the Smart
              Object Security Workshop", RFC 7397, December 2014,
              <http://www.rfc-editor.org/info/rfc7397>.

   [TCPextensions]
              Honda, M., Nishida, Y., Greenhalgh, A., Handley, M., and
              H. Tokuda, "Is it Still Possible to Extend TCP?", In
              Proceedings of the ACM Internet Measurement Conference
              (IMC), Berlin, Germany, November 2011,
              <http://conferences.sigcomm.org/imc/2011/docs/p181.pdf>.

   [Tussles]  Clark, D., Wroclawski, J., Sollins, K., and R. Braden,
              "Tussle in Cyberspace: Defining Tomorrow's Internet", In
              Proceedings of ACM SIGCOMM, 2002,
              <http://conferences.sigcomm.org/sigcomm/2002/papers/
              tussle.html>.





Tschofenig, et al.            Informational                    [Page 21]

RFC 7452        Smart Object Architectural Considerations     March 2015


   [WP223]    Article 29 Data Protection Working Party, "Opinion 8/2014
              on the Recent Developments on the Internet of Things", 14/
              EN, WP 223, September 2014, <http://ec.europa.eu/justice/
              data-protection/article-29/documentation/
              opinion-recommendation/files/2014/wp223_en.pdf>.














































Tschofenig, et al.            Informational                    [Page 22]

RFC 7452        Smart Object Architectural Considerations     March 2015


Appendix A.  IAB Members at the Time of Approval



   Jari Arkko
   Mary Barnes
   Marc Blanchet
   Joel Halpern
   Ted Hardie
   Joe Hildebrand
   Russ Housley
   Eliot Lear
   Xing Li
   Erik Nordmark
   Andrew Sullivan
   Dave Thaler
   Brian Trammell

Acknowledgements



   We would like to thank the participants of the IAB Smart Object
   workshop for their input to the overall discussion about smart
   objects.

   Furthermore, we would like to thank Mike St. Johns, Jan Holler,
   Patrick Wetterwald, Atte Lansisalmi, Hannu Flinck, Bernard Aboba,
   Markku Tuohino, Wes George, Robert Sparks, S.  Moonsesamy, Dave
   Crocker, and Steve Crocker in particular for their review comments.

























Tschofenig, et al.            Informational                    [Page 23]

RFC 7452        Smart Object Architectural Considerations     March 2015


Authors' Addresses



   Hannes Tschofenig
   ARM Ltd.
   6060 Hall in Tirol
   Austria

   EMail: Hannes.Tschofenig@gmx.net
   URI:   http://www.tschofenig.priv.at


   Jari Arkko
   Jorvas  02420
   Finland

   EMail: jari.arkko@piuha.net


   Dave Thaler
   One Microsoft Way
   Redmond, WA  98052
   United States

   EMail: dthaler@microsoft.com


   Danny McPherson
   12061 Bluemont Way
   Reston, VA  20190
   United States

   EMail: dmcpherson@verisign.com



















Tschofenig, et al.            Informational                    [Page 24]