LWIG Working Group                                       C. Bormann, Ed.
Internet-Draft                                   Universitaet Bremen TZI
Intended status: Informational                         February 25, 2013
Expires: August 29, 2013


Guidance for Light-Weight Implementations of the Internet Protocol Suite
                      draft-ietf-lwig-guidance-03

Abstract

   Implementation of Internet protocols on small devices benefits from
   light-weight implementation techniques, which are often not
   documented in an accessible way.

   This document provides a first outline of and some initial content
   for the Light-Weight Implementation Guidance document planned by the
   IETF working group LWIG.

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 29, 2013.

Copyright Notice

   Copyright (c) 2013 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.  Code Components extracted from this document must
   include Simplified BSD License text as described in Section 4.e of



Bormann                 Expires August 29, 2013                 [Page 1]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   the Trust Legal Provisions and are provided without warranty as
   described in the Simplified BSD License.

Table of Contents

   1.  Introduction  . . . . . . . . . . . . . . . . . . . . . . . .   3
     1.1.  Objectives  . . . . . . . . . . . . . . . . . . . . . . .   4
     1.2.  Call for contributions  . . . . . . . . . . . . . . . . .   5
     1.3.  Terminology used in this document . . . . . . . . . . . .   5
     1.4.  Scope of the present document . . . . . . . . . . . . . .   6
     1.5.  Terminology boilerplate . . . . . . . . . . . . . . . . .   6
   2.  Drawing the Landscape . . . . . . . . . . . . . . . . . . . .   6
     2.1.  Design Objectives . . . . . . . . . . . . . . . . . . . .   6
     2.2.  Implementation Styles . . . . . . . . . . . . . . . . . .   7
     2.3.  Roles of nodes  . . . . . . . . . . . . . . . . . . . . .   8
     2.4.  Overview over the document  . . . . . . . . . . . . . . .   8
   3.  Data Plane Protocols  . . . . . . . . . . . . . . . . . . . .   8
     3.1.  Link Adaptation Layer . . . . . . . . . . . . . . . . . .   8
       3.1.1.  Fragmentation in a 6LoWPAN Route-Over Configuration .   8
         3.1.1.1.  Implementation Considerations for Not-So-
                   Constrained Nodes . . . . . . . . . . . . . . . .  10
     3.2.  Network Layer . . . . . . . . . . . . . . . . . . . . . .  10
     3.3.  Transport Layer . . . . . . . . . . . . . . . . . . . . .  10
       3.3.1.  TCP . . . . . . . . . . . . . . . . . . . . . . . . .  10
         3.3.1.1.  Absolutely required TCP behaviors for proper
                   functioning and interoperability  . . . . . . . .  11
         3.3.1.2.  Strongly encouraged, but non-essential, behaviors
                   of TCP  . . . . . . . . . . . . . . . . . . . . .  12
         3.3.1.3.  Experimental extensions that are not yet standard
                   practices . . . . . . . . . . . . . . . . . . . .  13
         3.3.1.4.  Others  . . . . . . . . . . . . . . . . . . . . .  13
     3.4.  Application Layer . . . . . . . . . . . . . . . . . . . .  14
       3.4.1.  General considerations about Application Programming
               Interfaces (APIs) . . . . . . . . . . . . . . . . . .  14
       3.4.2.  Constrained Application Protocol (CoAP) . . . . . . .  14
         3.4.2.1.  Message Layer Processing  . . . . . . . . . . . .  15
         3.4.2.2.  Message Parsing . . . . . . . . . . . . . . . . .  16
         3.4.2.3.  Storing Used Message IDs  . . . . . . . . . . . .  17
       3.4.3.  (Other Application Protocols...)  . . . . . . . . . .  20
   4.  Control Plane Protocols . . . . . . . . . . . . . . . . . . .  20
     4.1.  Link Layer Support  . . . . . . . . . . . . . . . . . . .  20
     4.2.  Network Layer . . . . . . . . . . . . . . . . . . . . . .  20
     4.3.  Routing . . . . . . . . . . . . . . . . . . . . . . . . .  21
     4.4.  Host Configuration and Lookup Services  . . . . . . . . .  21
     4.5.  Network Management  . . . . . . . . . . . . . . . . . . .  21
       4.5.1.  SNMP  . . . . . . . . . . . . . . . . . . . . . . . .  21
         4.5.1.1.  Background  . . . . . . . . . . . . . . . . . . .  21
         4.5.1.2.  Revisiting SNMP implementation for resource



Bormann                 Expires August 29, 2013                 [Page 2]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


                   constrained devices . . . . . . . . . . . . . . .  22
         4.5.1.3.  Proposed approach for building an memory
                   efficient SNMP agent  . . . . . . . . . . . . . .  22
         4.5.1.4.  Example . . . . . . . . . . . . . . . . . . . . .  23
         4.5.1.5.  Further improvements  . . . . . . . . . . . . . .  25
         4.5.1.6.  Conclusion  . . . . . . . . . . . . . . . . . . .  26
   5.  Security protocols  . . . . . . . . . . . . . . . . . . . . .  26
     5.1.  Cryptography for Constrained Devices  . . . . . . . . . .  26
     5.2.  Transport Layer Security  . . . . . . . . . . . . . . . .  26
     5.3.  Network Layer Security  . . . . . . . . . . . . . . . . .  26
     5.4.  Network Access Control  . . . . . . . . . . . . . . . . .  26
       5.4.1.  PANA  . . . . . . . . . . . . . . . . . . . . . . . .  26
         5.4.1.1.  PANA AVPs . . . . . . . . . . . . . . . . . . . .  27
         5.4.1.2.  PANA Phases . . . . . . . . . . . . . . . . . . .  27
         5.4.1.3.  PANA session state parameters . . . . . . . . . .  29
   6.  Wire-Visible Constraints  . . . . . . . . . . . . . . . . . .  31
   7.  Wire-Invisible Constraints  . . . . . . . . . . . . . . . . .  31
   8.  IANA Considerations . . . . . . . . . . . . . . . . . . . . .  32
   9.  Security Considerations . . . . . . . . . . . . . . . . . . .  32
   10. Acknowledgements  . . . . . . . . . . . . . . . . . . . . . .  32
     10.1.  Contributors . . . . . . . . . . . . . . . . . . . . . .  32
   11. References  . . . . . . . . . . . . . . . . . . . . . . . . .  33
     11.1.  Normative References . . . . . . . . . . . . . . . . . .  33
     11.2.  Informative References . . . . . . . . . . . . . . . . .  33
   Author's Address  . . . . . . . . . . . . . . . . . . . . . . . .  34

1.  Introduction

   Today's Internet is experienced by users as a set of applications,
   such as email, instant messaging, and social networks.  There are
   substantial differences in performance between the various end
   devices with these applications, but in general end devices
   participating in the Internet today are considered to have relatively
   high performance.

   More and more communications technology is being embedded into our
   environment.  Different types of devices in our buildings, vehicles,
   equipment and other objects have a need to communicate.  It is
   expected that most of these devices will employ the Internet Protocol
   suite.  The term "Internet of Things" denotes a trend where a large
   number of devices directly benefit from communication services that
   use Internet protocols.  Many of these devices are not primarily
   computing devices operated by humans, but exist as components in
   buildings, vehicles, and the environment.  There will be a lot of
   variation in the computing power, available memory, communications
   bandwidth, and other capabilities between different types of these
   devices.  With many low-cost, low-power and otherwise constrained
   devices, it is not always easy to embed all the necessary features.



Bormann                 Expires August 29, 2013                 [Page 3]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   Historically, there has been a trend to invent special "light-weight"
   _protocols_ to connect the most constrained devices.  However, much
   of this development can simply run on existing Internet protocols,
   provided some attention is given to achieving light-weight
   _implementations_.  In some cases the new, constrained environments
   can indeed benefit from protocol optimizations and additional
   protocols that help optimize Internet communications and lower the
   computational requirements.  Examples of IETF standardization efforts
   targeted for these environments include the "IPv6 over Low power WPAN
   (6LoWPAN)", "Routing Over Low power and Lossy networks (ROLL)", and
   "Constrained RESTful Environments (CoRE)" working groups.  More
   generally, however, techniques are required to implement both these
   optimized protocols as well as the other protocols of the Internet
   protocol suite in a way that makes them applicable to a wider range
   of devices.

1.1.  Objectives

   The present document, a product of the IETF Light-Weight
   Implementation Guidance (LWIG) Working Group, focuses on helping the
   implementers of the smallest devices.  The goal is to be able to
   build minimal yet interoperable IP-capable devices for the most
   constrained environments.

   Building a small implementation does not have to be hard.  Many small
   devices use stripped down versions of general purpose operating
   systems and their TCP/IP stacks.  However, there are implementations
   that go even further in minimization and can exist in as few as a
   couple of kilobytes of code, as on some devices this level of
   optimization is necessary.  Technical and cost considerations may
   limit the computing power, battery capacity, available memory, or
   communications bandwidth that can be provided.  To overcome these
   limitations the implementers have to employ the right hardware and
   software mechanisms.  For instance, certain types of memory
   management or even fixed memory allocation may be required.  It is
   also useful to understand what is necessary from the point of view of
   the communications protocols and the application employing them.  For
   instance, a device that only acts as a client or only requires one
   connection can simplify its TCP implementation considerably.

   The purpose of this document is to collect experiences from
   implementers of IP stacks in constrained devices.  The focus is on
   techniques that have been used in actual implementations and do not
   impact interoperability with other devices.  The techniques shall
   also not affect conformance to the relevant specifications.  We
   describe implementation techniques for reducing complexity, memory
   footprint, or power usage.




Bormann                 Expires August 29, 2013                 [Page 4]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   The topics for this working group will be chosen from Internet
   protocols that are in wide use today, such as IPv4 and IPv6; UDP and
   TCP; ICMPv4/v6, MLD/IGMP and ND; DNS and DHCPv4/v6; TLS, DTLS and
   IPsec; as well as from the optimized protocols that result from the
   work of the 6LoWPAN, RPL, and CoRE working groups.  This document
   will be helpful for the implementers of new devices or for the
   implementers of new general-purpose small IP stacks.  It is also
   expected that the document will increase our knowledge of what
   existing small implementations do, and will help in the further
   optimization of the existing implementations.  In areas where the
   considerations for small implementations have already been documented
   in an accessible way, we will refer to those documents instead of
   duplicating the material here.

   Generic hardware design advice and software implementation techniques
   are outside the scope of this document.  Protocol implementation
   experience, however, is the focus.  There is no intention to describe
   any new protocols or protocol behavior modifications beyond what is
   already allowed by existing RFCs, because it is important to ensure
   that different types of devices can work together.  For example,
   implementation techniques relating to security mechanisms are within
   scope, but mere removal of security functionality from a protocol is
   rarely an acceptable approach.

1.2.  Call for contributions

   The present draft of the document is an outline that will grow with
   the contributions received, which are expressly invited.  As this
   document focuses on experience from existing implementations, this
   requires implementer input; in particular, participation is required
   from the implementers of existing small IP stacks.  "Small" here is
   intended to be applicable approximately to what is described in
   Section 2 -\u002D where it is more important that the technique
   described is grounded in actual experience than that the experience
   is actually from a (very) constrained system.

   Only a few subsections are fleshed out in this initial draft;
   additional subsections will quickly be integrated from additional
   contributors.

1.3.  Terminology used in this document

   The present document has originally also been used to develop
   pertinent terminology.  This has been factored out into a separate
   document, [I-D.ietf-lwig-terminology], which is now a prerequisite to
   reading the present document.





Bormann                 Expires August 29, 2013                 [Page 5]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


1.4.  Scope of the present document

   Using this terminology, we can now more precisely define the scope of
   the present document:

   This document is about implementation techniques that enable
   constrained nodes to form constrained node networks.

   Delay-Tolerant Networks (DTNs) are out of scope.  (See Section 1.1
   above for a further list of non-goals.)

1.5.  Terminology boilerplate

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
   document are to be interpreted as described in RFC 2119.  As this is
   an informational document, the [RFC2119] keywords will only be used
   to underscore requirements where similar key words apply in the
   context of the specifications the light-weight implementation of
   which is being discussed.

   The term "byte" is used in its now customary sense as a synonym for
   "octet".

2.  Drawing the Landscape

   There is not a single kind of constrained, Internet-connected device.
   To the contrary, the trend is towards much more functional variety of
   such devices than is customary today in the Internet.  The
   terminology document [I-D.ietf-lwig-terminology] introduces a number
   of terms that will be used here to locate some of the technique
   described in the following sections within certain areas of
   applications.

2.1.  Design Objectives

   o  Consideration for design or implementation approaches for
      implementation of IP stacks for constrained devices will be
      impacted by the RAM usage for these designs.  Here the
      consideration is what is the best approach to minimize overhead.

   o  In addition, the impact on throughput in terms of IP protocol
      implementation must take into consideration the methods that
      minimize overhead but balance performance requirements for the
      light-weight constrained devices.






Bormann                 Expires August 29, 2013                 [Page 6]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   o  Protocol implementation must consider its impact on CPU
      utilization.  Here guidance will be provided on how to minimize
      tasks that require additional CPU execution time.

   How does the implementation of the IP stack effect the application
   both in terms of performance but also of those same attributes and
   requirements (RAM, CPU usage, etc.)  that we are examining for the IP
   protocol stack?

   From performing a synthesis of implementation experiences we will be
   able to understand and document the benefits and consequences of
   varied approaches.  Scaling code and selected approaches in terms of
   scaling from, say, a 8-bit micro to a 16-bit micro.  Such scaling for
   the approach will aid in the development of single code base when
   possible.

2.2.  Implementation Styles

   Compared to personal computing devices, constrained devices tend to
   make use of quite different classes of operating systems, if that
   term is even applicable.

   ...

   o  Single-threaded/giant mainloop

   o  Event-driven vs.  threaded/blocking

      *  The usual multi-threaded model blocks a thread on primitives
         such as connect(), accept() or read() until an external event
         takes place.  This model is often thought to consume too much
         RAM and CPU processing.

      *  The event driven model uses a non-blocking approach: E.g., when
         an application interface sends a message, the routine would
         return immediately (before the message is sent).  A call-back
         facility notifies the application or calling code when the
         desired processing is completed.  Here the benefit is that no
         thread context needs to be preserved for long periods of time.

   o  Single/multiple processing elements

   o  E.g., separate radio/network processor

   Introduce these briefly: Some techniques may be applicable only to
   some of these styles!





Bormann                 Expires August 29, 2013                 [Page 7]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


2.3.  Roles of nodes

   Constrained nodes are by necessity more specialized than general
   purpose computing devices; they may have a quite specific role.  Some
   implementation techniques may also

   o  Constrained nodes

   o  Nodes talking to constrained nodes

   o  Gateways/Proxies

   In all these cases, constrained nodes that are "sleepy" pose
   additional considerations.  (Explain sleepy...) E.g., a node talking
   to a sleepy node may need to make special arrangements; this is even
   more true where a gateway or proxy interfaces the general Internet

   o  Bandwidth/latency considerations

2.4.  Overview over the document

   The following sections will first go through a number of specific
   protocol layers, starting from layers of the data plane (link
   adaptation, network, transport, application), followed by control
   plane protocol layers (link layer support, network layer and routing,
   host configuration and lookup services).  We then look at security
   protocols (general cryptography considerations, transport layer
   security, network layer security, network access control).  Finally,
   we discuss some specific, cross-layer concerns, some "wire-visible",
   some of concern within a specific implementation.  Clearly, many
   topics could be discussed in more than one place in this structure.
   The objective is not to have something for each of the potential
   topics, but to document the most valuable experience that may be
   available.

3.  Data Plane Protocols

3.1.  Link Adaptation Layer

   6LoWPAN

3.1.1.  Fragmentation in a 6LoWPAN Route-Over Configuration

   Author: Carsten Bormann

   6LoWPAN [RFC4944] is an adaptation layer that maps IPv6 with its
   minimum MTU of 1280 bytes to IEEE 802.15.4, which has a physical
   layer MTU of only 127 bytes (some of which are taken by MAC layer and



Bormann                 Expires August 29, 2013                 [Page 8]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   adaptation layer headers).  Therefore, the adaptation layer provides
   a fragmentation and reassembly scheme that can fragment a single IPv6
   packet of up to 1280 bytes into multiple adaptation layer fragments
   of up to 127 bytes each (including MAC and adaptation layer
   overhead).

   In a route-over configuration, implementing this adaptation layer
   fragmentation scheme straightforwardly means that reassembly and then
   fragmentation are performed at each forwarding hop.  As fragments
   from several packets may be arriving interleaved with each other,
   this approach requires buffer space for multiple MTU-size IPv6
   packets.

   In a mesh-under configuration, adaptation layer fragments can be
   forwarded independently of each other.  It would be preferable if
   something similar were possible for route-over.  Complete
   independence in forwarding of adaptation layer fragments is not
   possible for route-over, however, as the layer-3 addresses needed for
   forwarding are in the initial bytes of the IPv6 header, which is
   present only in the first fragment of a larger packet.

   Instead of performing a full reassembly, implementations may be able
   to optimize this process by not keeping a full reassembly buffer, but
   just a runt buffer (called "virtual reassembly buffer" in [WEI]) for
   each IP packet.  This buffer caches only the datagram_tag field (as
   usual combined with the sender's link layer address, the
   destination's link layer address and the datagram_size field) and the
   IPv6 header including the relevant addresses.  Initial fragments are
   then forwarded independently (after header decompression/compression)
   and create a runt reassembly buffer.  Non-initial fragments (which
   don't require header decompression/compression in 6LoWPAN) are
   matched against the runt buffers by datagram_tag etc.  and forwarded
   if an IPv6 address is available.  (This simple scheme may be
   complicated a bit if header decompression/compression of the initial
   fragment causes an overflow of the physical MTU; in this case some
   overflow data may need to be stored in the runt buffers to be
   combined with further fragments or may simply be forwarded as a
   separate additional fragment.)

   If non-initial fragments arrive out of order before the initial
   fragment, a route-over router may want to keep the contents of the
   non-initial fragments until the initial fragment is available, which
   does need some buffer space.  If that is not available, a more
   constrained route-over router may simply discard out-of order non-
   initial fragments, possibly taking note that there is no point in
   forwarding any more fragments with the same combination of 6LoWPAN
   datagram_tag field, L2 addresses and datagram_size.




Bormann                 Expires August 29, 2013                 [Page 9]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   Runt buffers should time out like full reassembly buffers, and may
   either keep a map of fragments forwarded or they may simply be
   removed upon forwarding the final fragment, assuming that no out-of-
   order fragments will follow.

3.1.1.1.  Implementation Considerations for Not-So-Constrained Nodes

   [RFC4944] makes no explicit mandates about the order in which
   fragments should be sent.  Because it is heavily favored by the above
   implementation techniques, it is highly advisable for all
   implementations to always send adaptation layer fragments in natural
   order, i.e., starting with the initial fragment, continuing with
   increasing datagram_offset.

3.2.  Network Layer

   IPv4 and IPv6

3.3.  Transport Layer

   TCP and UDP

   Both TCP and UDP employ 16-bit one's-complement checksums to protect
   against transmission errors.  A number of RFCs discuss efficient
   implementation techniques for computing and updating Internet
   Checksums [RFC1071] [RFC1141] [RFC1624].  (Updating the Internet
   Checksum, as opposed to computing it from scratch, may be of interest
   where a pre-computed packet is provided, e.g., in Flash ROM, and a
   copy is made in RAM and updated with some current values, or when the
   actual transmitted packet is composed from pre-defined parts in ROM
   and new parts in RAM.)

3.3.1.  TCP

   Ed.  Note:

   The following outline of a section is an attempt to provide
   substructure for a future discussion of TCP-related issues based on
   the TCP Roadmap, [RFC4614].  The indented text, as well as the RFC
   citations, are copied (and redacted) from there; this certainly needs
   to be refined in a future version.  (Some additional adaptation of
   the material may also be required as RFC 2581 was since obsoleted by
   RFC 5681, and RFC 3782 was obsoleted by RFC 6582.)

   Author: Yuanchen Ma

   In [RFC4614], the TCP related RFCs are summarized.  Some RFCs
   describe absolutely required TCP behaviors for proper functioning and



Bormann                 Expires August 29, 2013                [Page 10]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   interoperability.  Further RFCs describe strongly encouraged, but
   non-essential, behaviors.  There are also experimental extensions
   that are not yet standard practices, but that potentially could be in
   the future.

   In this subsection, the influence of resource constrained nodes on
   TCP implementations are summarized according to the lists of
   [RFC4614].

3.3.1.1.  Absolutely required TCP behaviors for proper functioning and
          interoperability

   RFC 793 S: "Transmission Control Protocol", STD 7 (September 1981)

   In RFC793, the TCP state machine and event processing, and TCP's
   semantics for data transmission, reliability, flow control,
   multiplexing, and acknowledgment.  For this part, the constraint of
   memory will limit the multiplexing capability of TCP.  /_text needed
   for RFC793_/

   RFC 1122 S: "Requirements for Internet Hosts - Communication Layers"
   (October 1989)

   RFC 2460 S: "Internet Protocol, Version 6 (IPv6) Specification
   (December 1998)

   RFC 2873 S: "TCP Processing of the IPv4 Precedence Field" (June 2000)

      This document [RFC2873] removes from the TCP specification all
      processing of the precedence bits of the TOS byte of the IP
      header.

   These three RFCs mandate the support for IPv6 and TOS in IP header,
   which are a must for resource constrained node to implement.

   RFC 2581 S: "TCP Congestion Control" (April 1999)

      Although RFC 793 did not contain any congestion control
      mechanisms, today congestion control is a required component of
      TCP implementations.  This document [RFC2581] defines the current
      versions of Van Jacobson's congestion avoidance and control
      mechanisms for TCP, based on his 1988 SIGCOMM paper [Jac88].  RFC
      2001 was a conceptual precursor that was obsoleted by RFC 2581.

      A number of behaviors that together constitute what the community
      refers to as "Reno TCP" are described in RFC 2581.





Bormann                 Expires August 29, 2013                [Page 11]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


      RFC 1122 mandates the implementation of a congestion control
      mechanism, and RFC 2581 details the currently accepted mechanism.
      RFC 2581 differs slightly from the other documents listed in this
      section, as it does not affect the ability of two TCP endpoints to
      communicate; however, congestion control remains a critical
      component of any widely deployed TCP implementation and is
      required for the avoidance of congestion collapse and to ensure
      fairness among competing flows.

   RFC 2988 S: "Computing TCP's Retransmission Timer" (November 2000)

      Abstract: "This document defines the standard algorithm that
      Transmission Control Protocol (TCP) senders are required to use to
      compute and manage their retransmission timer.

3.3.1.2.  Strongly encouraged, but non-essential, behaviors of TCP

   RFC 1323 S: "TCP Extensions for High Performance" (May 1992)

      This document [RFC1323] defines TCP extensions for window scaling,
      timestamps, and protection against wrapped sequence numbers, for
      efficient and safe operation over paths with large bandwidth-delay
      products.

   RFC 2675 S: "IPv6 Jumbograms" (August 1999)

      IPv6 supports longer datagrams than were allowed in IPv4.

   RFC 3168 S: "The Addition of Explicit Congestion Notification (ECN)
   to IP" (September 2001)

3.3.1.2.1.  Congestion Control and Loss Recovery Extensions

   RFC 3042 S: "Enhancing TCP's Loss Recovery Using Limited Transmit"
   (January 2001)

      Abstract: "This document proposes Limited Transmit, a new
      Transmission Control Protocol (TCP) mechanism that can be used to
      more effectively recover lost segments when a connection's
      congestion window is small

   RFC 3390 S: "Increasing TCP's Initial Window" (October 2002)

      This document [RFC3390] updates RFC 2581 to permit an initial TCP
      window of three or four segments during the slow-start phase,
      depending on the segment size.





Bormann                 Expires August 29, 2013                [Page 12]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   RFC 3782 S: "The NewReno Modification to TCP's Fast Recovery
   Algorithm" (April 2004)

      This document [RFC3782] specifies a modification to the standard
      Reno fast recovery algorithm, whereby a TCP sender can use partial
      acknowledgments to make inferences determining the next segment to
      send in situations where SACK would be helpful but isn't
      available.

3.3.1.2.2.  SACK-Based Loss Recovery and Congestion Control

   RFC 2018 S: "TCP Selective Acknowledgment Options" (October 1996)

      This document [RFC2018] defines the basic selective acknowledgment
      (SACK) mechanism for TCP.

   RFC 2883 S: "An Extension to the Selective Acknowledgement (SACK)
   Option for TCP" (July 2000)

      This document [RFC2883] extends RFC 2018 to cover the case of
      acknowledging duplicate segments.

   RFC 3517 S: "A Conservative Selective Acknowledgment (SACK)-based
   Loss Recovery Algorithm for TCP" (April 2003)

3.3.1.2.3.  Dealing with Forged Segments

   RFC 1948 I: "Defending Against Sequence Number Attacks" (May 1996)

   RFC 2385 S: "Protection of BGP Sessions via the TCP MD5 Signature
   Option" (August 1998)

3.3.1.3.  Experimental extensions that are not yet standard practices

   The experimental extensions are not mature yet.  The contents need to
   be validated to be safe and logical behavior.  It is not recommended
   for the resource constrained node to implement.

3.3.1.4.  Others

   RFC 2923 I: "TCP Problems with Path MTU Discovery" (September 2000)

      From abstract: "This memo catalogs several known Transmission
      Control Protocol (TCP) implementation problems dealing with Path
      Maximum Transmission Unit Discovery (PMTUD), including the long-
      standing black hole problem, stretch acknowlegements (ACKs) due to
      confusion between Maximum Segment Size (MSS) and segment size, and
      MSS advertisement based on PMTU."  [RFC2923]



Bormann                 Expires August 29, 2013                [Page 13]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


3.4.  Application Layer

3.4.1.  General considerations about Application Programming Interfaces
        (APIs)

   Author: Carl Williams

   Constrained devices are not necessarily in a position to use APIs
   that would be considered "standard" for less constrained environments
   (e.g., Berkeley sockets or those defined by POSIX).

   When an API implements a protocol, this can be based on proxy methods
   for remote invocations that underneath rely on the communication
   protocol.  One of the roles of the API can be exactly to hide the
   detail of the transport protocol.

   Changes to the lower layers will be made to implement light-weight
   stacks so this impacts that implementation and inter-workings with
   the API.  Similar considerations such as RAM, CPU utilization and
   performance requirements apply to the API and its use of the lower
   layer resources (i.e., buffers).

   Considerations for the proper approach for a developer to request
   services from an application program need to be explored and
   documented.  Such considerations will allow the progression of a
   common consistent networking paradigm without inventing a new way of
   programming these devices.

   In addition, such considerations will take into account the inter-
   working of the API with the protocols.  Protocols are more complex to
   use as they are less direct and take a lot of serializing, de-
   serializing and dispatching type logic.

   So the connection of the API and the protocols on a constrained
   device becomes even more important to balance the requirements of
   RAM, CPU and performance.

   _** Here we will proceed to collect and document ... insert
   experiences from existing API on constrained devices (TBD) **_

3.4.2.  Constrained Application Protocol (CoAP)

   Author: Olaf Bergmann

   The Constrained Application Protocol [I-D.ietf-core-coap] has been
   designed specifically for machine-to-machine communication in
   networks with very constrained nodes.  Typical application scenarios
   therefore include building automation and the Internet of Things.



Bormann                 Expires August 29, 2013                [Page 14]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   The major design objectives have been set on small protocol overhead,
   robustness against packet loss, and high latency induced by small
   bandwidth shares or slow request processing in end nodes.  To
   leverage integration of constrained nodes with the world-wide
   Internet, the protocol design was led by the architectural style that
   accounts for the scalability and robustness of the Hypertext Transfer
   Protocol [RFC2616].

   Lightweight implementations benefit from this design in many
   respects: First, the use of Uniform Resource Identifiers (URIs) for
   naming resources and the transparent forwarding of their
   representations in a server-stateless request/response protocol make
   protocol-translation to HTTP a straightforward task.  Second, the set
   of protocol elements that are inevitable for the core protocol and
   thus must be implemented on every node has been kept very small to
   avoid unnecessary accumulation of optional features.  Options that
   -\u002D when present -\u002D are critical for message processing are
   explicitly marked as such to force immediate rejection of messages
   with unknown critical options.  Third, the syntax of protocol data
   units is easy to parse and is carefully defined to avoid creation of
   state in servers where possible.

   Although these features enable lightweight implementations of the
   Constrained Application Protocol, there is still a trade-off between
   robustness and latency of constrained nodes on one hand and resource
   demands (such as battery consumption, dynamic memory needs and static
   code-size) on the other.  This section gives some guidance on
   possible strategies to solve this trade-off for very constrained
   nodes (Class 1 in [I-D.ietf-lwig-terminology]).  The main focus is on
   servers as this is deemed the predominant case where CoAP
   applications are faced with tight resource constraints.

   Additional considerations for the implementation of CoAP on tiny
   sensors are given in [I-D.arkko-core-sleepy-sensors].

3.4.2.1.  Message Layer Processing

   For constrained nodes of Class 1 or even Class 2, limiting factors
   for (wireless) network communication usually are RAM size and battery
   lifetime.  Most applications therefore try to avoid dealing with
   fragmented packets on the network layer and minimize internal buffer
   space for both transmit and receive operations.  One of the most
   expensive operations hence is the retransmission of messages as it
   implies additional energy consumption for the (radio) network
   interface and occupied RAM storage for the send buffer.

   Where multi-threading is not an option at all because no full-fledged
   operating system is present, all operations are triggered by a big



Bormann                 Expires August 29, 2013                [Page 15]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   main loop in a send-receive-dispatch cycle.  To implement the packet
   retransmission, CoAP implementations at least need a separate send
   buffer and a decent notion of time, e.g.  as a strictly monotonic
   increasing tick counter.  For platforms that disable clock tick
   interrupts in sleep states, the application must take into
   consideration the clock deviation that occurs during sleep (or ensure
   to remain in idle state until the message has been acknowledged or
   the maximum number of retransmissions is reached).  Since CoAP allows
   up to four retransmissions with a binary exponential back-off it
   could take up to 45 seconds until the send operation is complete.
   Even in idle state, this means substantial energy consumption for
   low-power nodes.  Implementers therefore might choose a two-step
   strategy: First, do one or two retransmissions and then, in the later
   phases of back-off, go to sleep until the next retransmission is due.
   In the meantime, the node could check for new messages including the
   acknowledgement for any confirmable message to send.

   A similar strategy holds for confirmable messages with separate
   responses.  This concept entitles CoAP servers to return an empty
   acknowledgement to indicate that a confirmable request has been
   understood and is being processed.  Once a proper response has been
   generate to fulfill the request, it is sent back as a confirmable
   message as well.  The server implementation in this case must be able
   to map retransmissions of the original request to the ongoing
   operation and provide the client-selected Token to map between
   original request and the separate response.

   Depending on the number of requests that can be handled in parallel,
   an implementation might create a stub response filled with any option
   that has to be copied from the original request to the separate
   response, especially the Token option.  The drawback of this
   technique is that the server must be prepared to receive
   retransmissions of the previous (confirmable) request to which a new
   acknowledgement must be generated.  If memory is an issue, a single
   buffer can be used for both tasks: Only the message type and code
   must be updated, changing the message id is optional.  Once the
   resource representation is known, it is added as new payload at the
   end of the stub response.  Acknowledgements still can be sent as
   described before as long as no additional options are required to
   describe the payload.

3.4.2.2.  Message Parsing

   Both CoAP clients and servers must construct outgoing CoAP PDUs and
   parse incoming messages.  The basic message header consists of only
   four octets and thus can be mapped easily to an internal data
   structure, considering the actual byte order of the host.  Once the
   message is accepted for further processing, the set of options



Bormann                 Expires August 29, 2013                [Page 16]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   contained in the received message must be decoded to check for
   unknown critical options.  To avoid multiple passes through the
   option list, the option parser might maintain a bit-vector where each
   bit represents an option number that is present in the received
   request.  The delta-encoded option number indicates the number of
   left-shift operations to apply on a bit mask to set the corresponding
   bit.

   In addition, the byte index of every option is added to a sparse list
   (e.g.  a one-dimensional array) for fast retrieval.  This
   particularly enables efficient reduced-function handling of options
   that might occur more than once such as Uri-Path.  In this
   implementation strategy, the delta is zero for any subsequent path
   segment, hence the stored byte index for option 9 (Uri-Path) will be
   overwritten to hold a pointer to the last occurrence of that option,
   i.e., only the last path component actually matters.  (Of course,
   this requires choosing resource names where the combination of (final
   Uri-Path component, final Uri-Query component) is server-wide unique.

   Note:  Where skipping all but the last path segment is not feasible
      for some reason, resource identification could be ensured by some
      hash value calculated over the path segments.  For each segment
      encountered, the stored hash value is updated by the current
      option value.  This works if a cheap _perfect hashing_ scheme can
      be found for the resource names.

   Once the option list has been processed at least up to the highest
   option number that is supported by the application, any known
   critical option and all elective options can be masked out to
   determine if any unknown critical option was present.  If this is the
   case, this information can be used to create a 4.02 response
   accordingly.  (Note that the remaining options also must be processed
   to add further critical options included in the original request.)

3.4.2.3.  Storing Used Message IDs

   If CoAP is used directly on top of UDP (i.e., in NoSec mode), it
   needs to cope with the fact that the UDP datagram transport can
   reorder and duplicate messages.  (In contrast to UDP, DTLS has its
   own duplicate detection.)  CoAP has been designed with protocol
   functionality such that rejection of duplicate messages is always
   possible.  It is at the discretion of the receiver if it actually
   wants to make use of this functionality.  Processing of duplicate
   messages comes at a cost, but so does the management of the state
   associated with duplicate rejection.  Hence, a receiver may have good
   reasons to decide not to do the duplicate rejection.  If duplicate
   rejection is indeed necessary, e.g., for non-idempotent requests, it
   is important to control the amount of state that needs to be stored.



Bormann                 Expires August 29, 2013                [Page 17]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   Author: Esko Dijk

   CoAP's duplicate rejection functionality can be straightforwardly
   implemented in a CoAP end-point by storing, for each remote CoAP end-
   point ("peer") that it communicates with, a list of recently received
   CoAP Message IDs (MIDs) along with some timing information.  A CoAP
   message from a peer with a MID that is in the list for that peer can
   simply be discarded.

   The timing information in the list can then be used to time out
   entries that are older than the _expected extent of the re-ordering_,
   an upper bound for which can be estimated by adding the _potential
   retransmission window_ ([I-D.ietf-core-coap] section "Reliable
   Messages") and the time packets can stay alive in the network.

   Such a straightforward implementation is suitable in case other CoAP
   end-points generate random MIDs.  However, this storage method may
   consume substantial RAM in specific cases, such as:

   o  many clients are making periodic, non-idempotent requests to a
      single CoAP server;

   o  one client makes periodic requests to a large number of CoAP
      servers and/or requests a large number of resources; where servers
      happen to mostly generate separate CoAP responses (not piggy-
      backed);

   For example, consider the first case where the expected extent of re-
   ordering is 50 seconds, and N clients are sending periodic POST
   requests to a single CoAP server during a period of high system
   activity, each on average sending one client request per second.  The
   server would need 100 * N bytes of RAM to store the MIDs only.  This
   amount of RAM may be significant on a RAM-constrained platform.  On a
   number of platforms, it may be easier to allocate some extra program
   memory (e.g.  Flash or ROM) to the CoAP protocol handler process than
   to allocate extra RAM.  Therefore, one may try to reduce RAM usage of
   a CoAP implementation at the cost of some additional program memory
   usage and implementation complexity.

   Some CoAP clients generate MID values by a using a Message ID
   variable [I-D.ietf-core-coap] that is incremented by one each time a
   new MID needs to be generated.  (After the maximum value 65535 it
   wraps back to 0.)  We call this behavior "sequential" MIDs.  One
   approach to reduce RAM use exploits the redundancy in sequential MIDs
   for a more efficient MID storage in CoAP servers.

   Naturally such an approach requires, in order to actually reduce RAM
   usage in an implementation, that a large part of the peers follow the



Bormann                 Expires August 29, 2013                [Page 18]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   sequential MID behavior.  To realize this optimization, the authors
   therefore RECOMMEND that CoAP end-point implementers employ the
   "sequential MID" scheme if there are no reasons to prefer another
   scheme, such as randomly generated MID values.

   Security considerations might call for a choice for
   (pseudo)randomized MIDs.  Note however that with truly randomly
   generated MIDs the probability of MID collision is rather high in use
   cases as mentioned before, following from the Birthday Paradox.  For
   example, in a sequence of 52 randomly drawn 16-bit values the
   probability of finding at least two identical values is about 2
   percent.

   From here on we consider efficient storage implementations for MIDs
   in CoAP end-points, that are optimized to store "sequential" MIDs.
   Because CoAP messages may be lost or arrive out-of-order, a solution
   has to take into account that received MIDs of CoAP messages are not
   actually arriving in a sequential fashion, due to lost or reordered
   messages.  Also a peer might reset and lose its MID counter(s) state.
   In addition, a peer may have a single Message ID variable used in
   messages to many CoAP end-points it communicates with, which partly
   breaks sequentiality from the receiving CoAP end-point's perspective.
   Finally, some peers might use a randomly generated MID values
   approach.  Due to these specific conditions, existing sliding window
   bitfield implementations for storing received sequence numbers are
   typically not directly suitable for efficiently storing MIDs.

   Table 1 shows one example for a per-peer MID storage design: a table
   with a bitfield of a defined length _K_ per entry to store received
   MIDs (one per bit) that have a value in the range [MID_i + 1 , MID_i
   + K].

              +----------+----------------+-----------------+
              | MID base | K-bit bitfield | base time value |
              +----------+----------------+-----------------+
              | MID_0    | 010010101001   | t_0             |
              |          |                |                 |
              | MID_1    | 111101110111   | t_1             |
              |          |                |                 |
              | ... etc. |                |                 |
              +----------+----------------+-----------------+

        Table 1: A per-peer table for storing MIDs based on MID\_i

   The presence of a table row with base MID_i (regardless of the
   bitfield values) indicates that a value MID_i has been received at a
   time t_i.  Subsequently, each bitfield bit k (0...K-1) in a row i
   corresponds to a received MID value of MID_i + k + 1.  If a bit k is



Bormann                 Expires August 29, 2013                [Page 19]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   0, it means a message with corresponding MID has not yet been
   received.  A bit 1 indicates such a message has been received already
   at approximately time t_i.  This storage structure allows e.g.  with
   k=64 to store in best case up to 130 MID values using 20 bytes, as
   opposed to 260 bytes that would be needed for a non-sequential
   storage scheme.

   The time values t_i are used for removing rows from the table after a
   preset timeout period, to keep the MID store small in size and enable
   these MIDs to be safely re-used in future communications.  (Note that
   the table only stores one time value per row, which therefore needs
   to be updated on receipt of another MID that is stored as a single
   bit in this row.  As a consequence of only storing one time value per
   row, older MID entries typically time out later than with a simple
   per-MID time value storage scheme.  The end-point therefore needs to
   ensure that this additional delay before MID entries are removed from
   the table is much smaller than the time period after which a peer
   starts to re-use MID values due to wrap-around of a peer's MID
   variable.  One solution is to check that a value t_i in a table row
   is still recent enough, before using the row and updating the value
   t_i to current time.  If not recent enough, e.g.  older than N
   seconds, a new row with an empty bitfield is created.)  [Clearly,
   these optimizations would benefit if the peer were much more
   conservative about re-using MIDs than currently required in the
   protocol specification.]

   The optimization described is less efficient for storing randomized
   MIDs that a CoAP end-point may encounter from certain peers.  To
   solve this, a storage algorithm may start in a simple MID storage
   mode, first assuming that the peer produces non-sequential MIDs.
   While storing MIDs, a heuristic is then applied based on monitoring
   some "hit rate", for example, the number of MIDs received that have a
   Most Significant Byte equal to that of the previous MID divided by
   the total number of MIDs received.  If the hit rate tends towards 1
   over a period of time, the MID store may decide that this particular
   CoAP end-point uses sequential MIDs and in response improve
   efficiency by switching its mode to the bitfield based storage.

3.4.3.  (Other Application Protocols...)

4.  Control Plane Protocols

4.1.  Link Layer Support

   ARP, ND; 6LoWPAN-ND

4.2.  Network Layer




Bormann                 Expires August 29, 2013                [Page 20]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   ICMP, ICMPv6, IGMP/MLD

4.3.  Routing

   RPL, AODV/DYMO, OLSRv2

4.4.  Host Configuration and Lookup Services

   DNS, DHCPv4, DHCPv6

4.5.  Network Management

   SNMP, netconf?

4.5.1.  SNMP

   Author: Brinda M C

   This section describes an approach for developing a light-weight SNMP
   agent for resource constrained devices running the 6LoWPAN/RPL
   protocol stack.  The motivation for the work is driven by two major
   factors:

   o  SNMP plays a vital role in monitoring and managing any operational
      network; 6LoWPAN based WSN is no exception to this.

   o  There is a need for building a light-weight SNMP agent which
      consumes less memory and less computational resources.

   The following subsections are organized as follows:

   o  Section 4.5.1.1 provides some background.

   o  In Section 4.5.1.2, we revisit existing SNMP implementation in the
      context of memory constrained devices.

   o  In Section 4.5.1.3, we present our approach for building a memory
      efficient SNMP agent.

   o  Using a realistic example, in Section 4.5.1.4, we illustrate how
      the proposed method can be implemented.

   o  In Section 4.5.1.5, we explore a few ideas which can further help
      in improving the memory utilization.

4.5.1.1.  Background





Bormann                 Expires August 29, 2013                [Page 21]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   Our initial SNMP agent implementation was completely based on Net-
   SNMP, well-known open-source network monitoring and management
   software.  After porting the agent on to the TelosB mote, we observed
   that it occupies a text program memory of more than 8 KiB on TinyOS
   and Contiki OS platforms.  (Note that both these platforms already
   use compiler optimizations to minimize the memory footprint.)  8 KiB
   is already non-negligible given the 48 KiB program memory limit of
   TelosB.  Added to this, the memory taken up by 6LoWPAN and the
   related protocol stacks are ever growing, causing serious memory
   crunch in the resource constrained devices.  We reached a situation
   where we could not build an image on the TinyOS/Contiki OS platforms
   with our SNMP agent.

   We came across SNMPv1 agent implementations elsewhere in the
   literature which also report similar memory consumption.  This
   motivated us to have a re-look at the existing SNMP agent
   implementation, and explore the possibility of an alternate
   implementation using altogether a different approach.

4.5.1.2.  Revisiting SNMP implementation for resource constrained
          devices

   If we look at a typical SNMP agent implementation, we can see that
   much of the memory consuming code is pertaining to ASN.1 related SNMP
   PDU parsing and SNMP PDU build operations.  The SNMP parsing mainly
   recovers various fields from the incoming PDU, such as the OIDs,
   whereas the SNMP PDU build is the reverse operation of building the
   response PDU from the OIDs.

   The key observation is that, for a given MIB definition, an OID of
   interest contained in the incoming SNMP PDU is already available,
   albeit in an encoded form.  This enables identifying the OID from the
   packet in its "raw" form, simplifying parser operation.

   We also can make use of this observation while building the response
   SNMP PDU.  For a given MIB definition, we can think of statically
   having a pre-composed ASN.1 encoded version of OIDs, and use them
   while constructing the response SNMP PDU.

4.5.1.3.  Proposed approach for building an memory efficient SNMP agent

   As noted in the previous section, since an SNMP OID is already
   _contained_ in the incoming network PDU, we came up with a simple OID
   signature identification method performed directly on the network PDU
   through simple memory comparisons and table look-ups.  Once the OID
   has been identified from the packet "in situ", the corresponding per-
   OID processing is carried out.  Through this scheme we completely
   eliminated expensive SNMP parse operations.



Bormann                 Expires August 29, 2013                [Page 22]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   For the SNMP PDU build, we use _pre-encoded_ OID variables which can
   simply be plugged into the network SNMP response packet directly
   depending on the request OID.  Now that the expensive build operation
   is taken care, what remains is the construction of the overall SNMP
   pdu which can be built through simple logic.  Through this scheme we
   completely eliminated expensive SNMP build operations.

   Based on these ideas, we have re-architected our original SNMP agent
   implementation and with our new implementation we were able to bring
   down its text memory usage all the way down to 4 KiB from the native
   SNMP agent implementation which occupied 8 KiB.

4.5.1.3.1.  Discussion on memory usage

   With respect to the memory usage, while we have achieved major
   reduction in terms of text program memory, which occupies a major
   chunk of memory, a question might come to mind with regard to the
   static memory allocation for maintaining the tables.  We found that
   this is not very significant to start with.  Through an efficient
   table representation, we further optimized the memory consumption.
   We could do so because a typical OID description is mainly dominated
   by a fixed part of the hierarchy.  This enables us to define few
   static prefixes, each corresponding to a particular hierarchy level
   of the OID.  In the context of 6LoWPAN, it can be expected that the
   number of hierarchy levels will be small.

4.5.1.4.  Example

   This section illustrates the simplicity and practicality of our
   approach with an example.  Let us consider the fragment of a
   representative MIB definition depicted in Figure 1

   iso
   |
   org
   |
   dod
   |
   internet
   |
   mgmt.mib-2
   |
   lowpanMIB
   |
   +--lowpanPrimaryStatistics(10)
      |
      +--PrimeStatsEntry(1)
         |



Bormann                 Expires August 29, 2013                [Page 23]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


         +-- -R-- INTEGER   lowpanMoteBatteryVoltageP(1)
         +-- -R-- Counter   lowpanFramesReceivedP(2)
         +-- -R-- Counter   lowpanFramesSentP(3)
         +-- -R-- Counter   ipv6ForwardedMsgP(4)
         +-- -R-- Counter   OUTSolicitationP(5)
         +-- -R-- Counter   OUTAdvertisementP(6)

                  Figure 1: A fragment of a MIB hierarchy

4.5.1.4.1.  Optimized SNMP Parsing

   Let us consider a GET request for the OIDs lowpanMoteBatteryVoltageP
   and lowpanFramesSentP.  Corresponding to these OIDs, a C array dump
   of the network PDU of SNMP packet with two OIDs in a variable binding
   would look as in Figure 2.

   char snmp_get_req_pkt[] = {
       0x30, 0x81, 0x3d, 0x02, 0x01, 0x00, 0x04, 0x06,
       0x70, 0x75, 0x62, 0x6c, 0x69, 0x63, 0xa0, 0x30,
       0x02, 0x04, 0x28, 0x29, 0xe4, 0x5d, 0x02, 0x01,
       0x00, 0x02, 0x01, 0x00, 0x30, 0x22, 0x30, 0x0f,
       0x06, 0x0b, 0x2b, 0x06, 0x01, 0x02, 0x01, 0x83,
       0x90, 0x12, 0x0a, 0x01, 0x01, 0x05, 0x00, 0x30,
       0x0f, 0x06, 0x0b, 0x2b, 0x06, 0x01, 0x02, 0x01,
       0x83, 0x90, 0x12, 0x0a, 0x01, 0x03, 0x05, 0x00 };

                Figure 2: An SNMP packet, represented in C

   Inspecting the above packet, we see that the main components of the
   PDU are:

   1.  Version (SNMPv1): [0x02, 0x01, 0x00]

   2.  Community Name ("public"): [0x04, 0x06, 0x70, 0x75, 0x62, 0x6c,
       0x69, 0x63]

   3.  ASN.1 encoded OIDs for lowpanMoteBatteryVoltageP, and
       lowpanFramesReceivedP:

       *  [0x30, 0x0f, 0x06, 0x0b, 0x2b, 0x06, 0x01, 0x02, 0x01, 0x83,
          0x90, 0x12, 0x0a, 0x01, 0x01, 0x05, 0x00]

       *  [0x30, 0x0f, 0x06, 0x0b, 0x2b, 0x06, 0x01, 0x02, 0x01, 0x83,
          0x90, 0x12, 0x0a, 0x01, 0x03, 0x05, 0x00]

   There is a significant overlap between the two OIDs, which can be
   used to simplify the parsing process.  We can, for instance, define
   one statically initialized array containing elements common between



Bormann                 Expires August 29, 2013                [Page 24]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   these OIDs.  Using this notion of common prefix idea, we can come up
   with an optimized table and the OID identification then boils down to
   simple memory comparisons within this table.  The optimized table
   construction will also result in scalability.

4.5.1.4.2.  Optimized SNMP Build

   Extending the same approach as described above, we can build the GET
   response by plugging in pre-encoded OIDs into the response packets.
   So, corresponding to the GET request for the OIDs as given in section
   4.1, we can define C arrays containing pre-encoded OIDs which can go
   into the response packet as in Figure 3.

   pdu_batt_volt[] = {
       0x30, 0x11, 0x06, 0x0b, 0x2b, 0x06, 0x01, 0x02,
       0x01, 0x83, 0x90, 0x12, 0x0a, 0x01, 0x01, 0x02,
       0x02, 0x00, 0x00 };

   pdu_frames_sent[] = {
       0x30, 0x11, 0x06, 0x0b, 0x2b, 0x06, 0x01, 0x02,
       0x01, 0x83, 0x90, 0x12, 0x0a, 0x01, 0x03, 0x41,
       0x02, 0x00, 0x00 };

                        Figure 3: Pre-encoded OIDs

   Since the ASN.1 basic encoding rules are in TLV format, the offset
   within the encoded OID where the value needs to be filled-in can be
   obtained from the length field.

   The table size optimization discussed in the previous section can be
   applied here, too.

   Note: Though we have taken a simple example to illustrate the
   efficacy of the proposed approach, the ideas presented here can
   easily be extended to other scenarios as well.

4.5.1.5.  Further improvements

   A few simple methods can reduce the code size as well as generate
   computationally inexpensive code.  These methods might sound obvious
   and trivial but are important for constrained devices.

   o  If possible, avoid using memory consuming data types such as
      floating point while representing a monitored variable when an
      equivalent representation of the same that occupies less memory is
      adequate.  For example, while a battery voltage indication could
      take a fractional value between 0 and 3 V, opt for an 8-bit
      quantized value.



Bormann                 Expires August 29, 2013                [Page 25]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   o  Using meta data in the MIB definition instead of absolute numbers
      can bring down the memory and processing significantly and can
      improve scalability too especially for a large scale WSN
      deployments.  Using the same example of battery voltage, one might
      think of an OID which represents fewer levels of the battery
      voltage signifying high, medium, low, very low.

   o  While a multi-level hierarchy for MIB definition might improve OID
      segregation the flip side is that it increases the overall length
      of the OID and results in extra memory and processing overhead.
      One may have to make a judicious choice while coming up with the
      MIB.

4.5.1.6.  Conclusion

   This subsection proposes a simple SNMP packet processing based
   approach for building a light-weight SNMP agent.  While there is
   scope for further improvement, we believe that the proposed method
   can be a reasonably good starting point for resource constrained
   6LoWPAN based networks.

5.  Security protocols

5.1.  Cryptography for Constrained Devices

5.2.  Transport Layer Security

   TLS, DTLS, ciphersuites, certificates

5.3.  Network Layer Security

   IPsec, IKEv2, transforms

   Advice for a minimal implementation of IKEv2 can be found in
   [I-D.kivinen-ipsecme-ikev2-minimal].

5.4.  Network Access Control

   (PANA, EAP, EAP methods)

5.4.1.  PANA

   Author: Mitsuru Kanda

   PANA [RFC5191] provides network access authentication between clients
   and access networks.  The PANA protocol runs between a PANA Client
   (PaC) and a PANA Authentication Agent (PAA).  PANA carries UDP
   encapsulated EAP [RFC3748] and includes various operational options.



Bormann                 Expires August 29, 2013                [Page 26]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   From the point of view of minimal implementation, some of these are
   not necessary for constrained devices.  This section describes a
   minimal PANA implementation for these devices.

   The minimization objective for this implementation mainly targets
   PaCs because constrained devices often are installed as network
   clients, such as sensors, metering devices, etc.

5.4.1.1.  PANA AVPs

   Each PANA message can carry zero or more AVPs (Attribute-Value Pairs)
   within its payload.  [RFC5191] specifies nine types of AVPs (AUTH,
   EAP-Payload, Integrity-Algorithm, Key-Id, Nonce, PRF-Algorithm,
   Result-Code, Session-Lifetime, and Termination-Cause).  All of them
   are required by all minimal implementations.  But there are some
   notes.

   Integrity-Algorithm AVP and PRF-Algorithm AVP:

   All PANA implementations MUST support AUTH_HMAC_SHA1_160 for PANA
   message integrity protection and PRF_HMAC_SHA1 for pseudo-random
   function (PRF) specified in [RFC5191].  Both of these are based on
   SHA-1, which therefore needs to be implemented in a minimal
   implementation.

   Nonce AVP:

   As the basic hash function is SHA-1, including a nonce of 20 bytes in
   the Nonce AVP is appropriate ([RFC5191], section 8.5).

5.4.1.2.  PANA Phases

   A PANA session consists of four phases -\u002D Authentication and
   authorization phase, Access phase, Re-Authentication phase, and
   Termination phase.

   Authentication and authorization phase:

   There are two types of PANA session initiation, PaC-initiated session
   and PAA-initiated session.  The minimal implementation must support
   PaC-initiated session and does not need to support PAA-initiated
   session.  Because a PaC (a constrained device) which may be a
   sleeping device, can not receive an unsolicited PANA-Auth-Request
   message from a PAA (PAA-initiated session).

   EAP messages can be carried in PANA-Auth-Request and PANA-Auth-Answer
   messages.  In order to reduce the number of messages, "Piggybacking
   EAP" is useful.  Both the PaC and PAA should include EAP-Payload AVP



Bormann                 Expires August 29, 2013                [Page 27]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   in each of PANA-Auth-Request and PANA-Auth-Answer messages as much as
   possible.  Figure 4 shows an example "Piggybacking EAP" sequence of
   the Authentication and authorization phase.

   PaC      PAA  Message(sequence number)[AVPs]
   ---------------------------------------------------------------------
      ----->     PANA-Client-Initiation(0)
      <-----     PANA-Auth-Request(x)[PRF-Algorithm,Integrity-Algorithm]
                                              // The 'S' (Start) bit set
      ----->     PANA-Auth-Answer(x)[PRF-Algorithm, Integrity-Algorithm]
                                              // The 'S' (Start) bit set
      <-----     PANA-Auth-Request(x+1)[Nonce, EAP-Payload]
      ----->     PANA-Auth-Answer(x+1)[Nonce, EAP-Payload]
      <-----     PANA-Auth-Request(x+2)[EAP-Payload]
      ----->     PANA-Auth-Answer(x+2)[EAP-Payload]
      <-----     PANA-Auth-Request(x+3)[Result-Code, EAP-Payload,
                                        Key-Id, Session-Lifetime, AUTH]
                                           // The 'C' (Complete) bit set
      ----->     PANA-Auth-Answer(x+3)[Key-Id, AUTH]
                                           // The 'C' (Complete) bit set

    Figure 4: Example sequence of the Authentication and authorization
       phase for a PaC-initiated session (using "Piggybacking EAP")

   Note: It is possible to include an EAP-Payload in both the PANA-Auth-
   Request and PANA-Auth-Answer messages with the 'S' bit set.  But the
   PAA should not include an EAP-Payload in the PANA-Auth-Request
   message with the 'S' bit set in order to stay stateless in response
   to a PANA-Client-Initiation message.

   Access phase:

   After Authentication and authorization phase completion, the PaC and
   PAA share a PANA Security Association (SA) and move Access phase.
   During Access phase, [RFC5191] describes both the PaC and PAA can
   send a PANA-Notification-Request message with the 'P' (Ping) bit set
   for the peer's PANA session liveness check (a.k.a "PANA ping").  From
   the minimal implementation point of view, the PAA should not send a
   PANA-Notification-Request message with the 'P' (Ping) bit set to
   initiate PANA ping since the PaC may be sleeping.  The PaC does not
   need to send a PANA-Notification-Request message with the 'P' (Ping)
   bit set for PANA ping to the PAA periodically and may omit the PANA
   ping feature itself if the PaC can detect the PANA session failure by
   other methods, for example, network communication failure.  In
   conclusion, the PaC does not need to implement the periodic liveness
   check feature sending PANA ping but a PaC that is awake should
   respond to a incoming PANA-Notification-Request message with the 'P'
   (Ping) bit set for PANA ping as possible.



Bormann                 Expires August 29, 2013                [Page 28]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   Re-Authentication phase:

   Before PANA session lifetime expiration, the PaC and PAA MUST re-
   negotiate to keep the PANA session.  This means that the PaC and PAA
   enter Re-Authentication phase.  Also in the Authentication and
   authorization phase, there are two types of re-authentication.  The
   minimal implementation must support PaC-initiated re-authentication
   and does not need to support PAA-initiated re-authentication (again
   because the PaC may be a sleeping device).  "Piggybacking EAP" is
   also useful here and should be used as well.  Figure 5 shows an
   example "Piggybacking EAP" sequence of the Re-Authentication phase.

   PaC      PAA  Message(sequence number)[AVPs]
   ---------------------------------------------------------------------
     ----->     PANA-Notification-Request(q)[AUTH]
                              // The 'A' (re-Authentication) bit set
     <-----     PANA-Notification-Answer(q)[AUTH]
                              // The 'A' (re-Authentication) bit set
     <-----     PANA-Auth-Request(p)[EAP-Payload, Nonce, AUTH]
     ----->     PANA-Auth-Answer(p)[EAP-Payload, Nonce, AUTH]
     <-----     PANA-Auth-Request(p+1)[EAP-Payload, AUTH]
     ----->     PANA-Auth-Answer(p+1)[EAP-Payload, AUTH]
     <-----     PANA-Auth-Request(p+2)[Result-Code, EAP-Payload,
                                       Key-Id, Session-Lifetime, AUTH]
                                       // The 'C' (Complete) bit set
     ----->     PANA-Auth-Answer(p+2)[Key-Id, AUTH]
                                       // The 'C' (Complete) bit set

   Figure 5: Example sequence of the Re-Authentication phase for a PaC-
               initiated session (using "Piggybacking EAP")

   Termination Phase:

   The PaC and PAA should not send a PANA-Termination-Request message
   except for explicitly terminating a PANA session within the lifetime.
   Both the PaC and PAA know their own PANA session lifetime expiration.
   This means the PaC and PAA should not send a PANA-Termination-Request
   message when the PANA session lifetime expired because of reducing
   message processing cost.

5.4.1.3.  PANA session state parameters

   All PANA implementations internally keep PANA session state
   information for each peer.  At least, all minimal implementations
   need to keep PANA session state parameters below (in the second
   column storage sizes are given in bytes):

   +----------------------+------------+-------------------------------+



Bormann                 Expires August 29, 2013                [Page 29]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   | State Parameter      | Size       | Comment                       |
   +----------------------+------------+-------------------------------+
   | PANA Phase           | 1          | Used for recording the        |
   | Information          |            | current PANA phase.           |
   |                      |            |                               |
   | PANA Session         | 4          |                               |
   | Identifier           |            |                               |
   |                      |            |                               |
   | PaC's IP address and | 6 or 18    | IP Address length (4 bytes    |
   | UDP port number      |            | for IPv4 and 16 bytes for     |
   |                      |            | IPv6) plus 2 bytes for UDP    |
   |                      |            | port number.                  |
   |                      |            |                               |
   | PAA's IP address and | 6 or 18    | IP Address length (4 bytes    |
   | UDP port number      |            | for IPv4 and 16 bytes for     |
   |                      |            | IPv6) plus 2 bytes for UDP    |
   |                      |            | port number.                  |
   |                      |            |                               |
   | Outgoing message     | 4          | Next outgoing request message |
   | sequence number      |            | sequence number.              |
   |                      |            |                               |
   | Incoming message     | 4          | Next expected incoming        |
   | sequence number      |            | request message sequence      |
   |                      |            | number.                       |
   |                      |            |                               |
   | A copy of the last   | variable   | Necessary to be able to       |
   | sent message payload |            | retransmit the message        |
   |                      |            | (unless it can be             |
   |                      |            | reconstructed on the fly).    |
   |                      |            |                               |
   | Retransmission       | 4          |                               |
   | interval             |            |                               |
   |                      |            |                               |
   | PANA Session         | 4          |                               |
   | lifetime             |            |                               |
   |                      |            |                               |
   | PaC nonce            | 20         | Generated by PaC and carried  |
   |                      |            | in the Nonce AVP.             |
   |                      |            |                               |
   | PAA nonce            | 20         | Generated by PAA and carried  |
   |                      |            | in the Nonce AVP.             |
   |                      |            |                               |
   | EAP MSK Identifier   | 4          |                               |
   |                      |            |                               |
   | EAP MSK value        | *)         | Generated by EAP method and   |
   |                      |            | used for generating           |
   |                      |            | PANA_AUTH_KEY.                |
   |                      |            |                               |



Bormann                 Expires August 29, 2013                [Page 30]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   | PANA_AUTH_KEY        | 20         | Necessary for PANA message    |
   |                      |            | protection.                   |
   |                      |            |                               |
   | PANA PRF algorithm   | 4          | Used for generating           |
   | number               |            | PANA_AUTH_KEY.                |
   |                      |            |                               |
   | PANA Integrity       | 4          | Necessary for PANA message    |
   | algorithm number     |            | protection.                   |
   +----------------------+------------+-------------------------------+


   *) (Storage size depends on the key derivation algorithm.)

   Note: EAP parameters except for MSK have not been listed here.  These
   EAP parameters are not used by PANA and depend on what EAP method you
   choose.

6.  Wire-Visible Constraints

   o  Checksum

   o  MTU

   o  Fragmentation and reassembly

   o  Options -\u002D implications of leaving some out

   o  Simplified TCP optimized for LLNs

   o  Out-of-order packets

7.  Wire-Invisible Constraints

   o  Buffering

   o  Memory management

   o  Timers

   o  Energy efficiency

   o  API

   o  Data structures

   o  Table sizes (somewhat wire-visible)

   o  Improved error handling due to resource overconsumption



Bormann                 Expires August 29, 2013                [Page 31]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


8.  IANA Considerations

   This document makes no requirements on IANA.  (This section to be
   removed by RFC editor.)

9.  Security Considerations

   (TBD.)

10.  Acknowledgements

   Much of the text of the introduction is taken from the charter of the
   LWIG working group and the invitation to the IAB workshop on
   Interconnecting Smart Objects with the Internet.  Thanks to the
   numerous contributors.  Angelo Castellani provided comments that led
   to improved text.

10.1.  Contributors

   The RFC guidelines no longer allow RFCs to be published with a large
   number of authors.  As there are many authors that have contributed
   to the sections of this document, their names are listed in the
   individual section headings as well as alphabetically listed with
   their affiliations below.

   +------------+--------------------+---------------------------------+
   | Name       | Affiliation        | Contact                         |
   +------------+--------------------+---------------------------------+
   | Brinda M C | Indian Institute   | brinda@ece.iisc.ernet.in        |
   |            | of Science         |                                 |
   |            |                    |                                 |
   | Carl       | MCSR Labs          | carlw@mcsr-labs.org             |
   | Williams   |                    |                                 |
   |            |                    |                                 |
   | Carsten    | Universitaet       | cabo@tzi.org                    |
   | Bormann    | Bremen TZI         |                                 |
   |            |                    |                                 |
   | Esko Dijk  | Philips Research   | esko.dijk@philips.com           |
   |            |                    |                                 |
   | Mitsuru    | Toshiba            | mitsuru.kanda@toshiba.co.jp     |
   | Kanda      |                    |                                 |
   |            |                    |                                 |
   | Olaf       | Universitaet       | bergmann@tzi.org                |
   | Bergmann   | Bremen TZI         |                                 |
   |            |                    |                                 |
   | Yuanchen   | Hitachi (China)    | ycma@hitachi.cn                 |
   | Ma         | R&D Corporation    |                                 |
   |            |                    |                                 |



Bormann                 Expires August 29, 2013                [Page 32]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   | ...        | ...                |                                 |
   +------------+--------------------+---------------------------------+


11.  References

11.1.  Normative References

   [I-D.ietf-lwig-terminology]
              Bormann, C. and M. Ersue, "Terminology for Constrained
              Node Networks", draft-ietf-lwig-terminology-00 (work in
              progress), February 2013.

   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119, March 1997.

   [RFC4944]  Montenegro, G., Kushalnagar, N., Hui, J., and D. Culler,
              "Transmission of IPv6 Packets over IEEE 802.15.4
              Networks", RFC 4944, September 2007.

11.2.  Informative References

   [I-D.arkko-core-sleepy-sensors]
              Arkko, J., Rissanen, H., Loreto, S., Turanyi, Z., and O.
              Novo, "Implementing Tiny COAP Sensors", draft-arkko-core-
              sleepy-sensors-01 (work in progress), July 2011.

   [I-D.ietf-6lowpan-btle]
              Nieminen, J., Savolainen, T., Isomaki, M., Patil, B.,
              Shelby, Z., and C. Gomez, "Transmission of IPv6 Packets
              over BLUETOOTH Low Energy", draft-ietf-6lowpan-btle-12
              (work in progress), February 2013.

   [I-D.ietf-core-coap]
              Shelby, Z., Hartke, K., Bormann, C., and B. Frank,
              "Constrained Application Protocol (CoAP)", draft-ietf-
              core-coap-13 (work in progress), December 2012.

   [I-D.kivinen-ipsecme-ikev2-minimal]
              Kivinen, T., "Minimal IKEv2", draft-kivinen-ipsecme-
              ikev2-minimal-01 (work in progress), October 2012.

   [I-D.mariager-6lowpan-v6over-dect-ule]
              Mariager, P. and J. Petersen, "Transmission of IPv6
              Packets over DECT Ultra Low Energy", draft-mariager-
              6lowpan-v6over-dect-ule-02 (work in progress), May 2012.





Bormann                 Expires August 29, 2013                [Page 33]


Internet-Draft Guidance for Light-Weight Implementations   February 2013


   [RFC1071]  Braden, R., Borman, D., Partridge, C., and W. Plummer,
              "Computing the Internet checksum", RFC 1071, September
              1988.

   [RFC1141]  Mallory, T. and A. Kullberg, "Incremental updating of the
              Internet checksum", RFC 1141, January 1990.

   [RFC1624]  Rijsinghani, A., "Computation of the Internet Checksum via
              Incremental Update", RFC 1624, May 1994.

   [RFC2616]  Fielding, R., Gettys, J., Mogul, J., Frystyk, H.,
              Masinter, L., Leach, P., and T. Berners-Lee, "Hypertext
              Transfer Protocol -- HTTP/1.1", RFC 2616, June 1999.

   [RFC3748]  Aboba, B., Blunk, L., Vollbrecht, J., Carlson, J., and H.
              Levkowetz, "Extensible Authentication Protocol (EAP)", RFC
              3748, June 2004.

   [RFC4614]  Duke, M., Braden, R., Eddy, W., and E. Blanton, "A Roadmap
              for Transmission Control Protocol (TCP) Specification
              Documents", RFC 4614, September 2006.

   [RFC5191]  Forsberg, D., Ohba, Y., Patil, B., Tschofenig, H., and A.
              Yegin, "Protocol for Carrying Authentication for Network
              Access (PANA)", RFC 5191, May 2008.

   [WEI]      Shelby, Z. and C. Bormann, "6LoWPAN: the Wireless Embedded
              Internet", ISBN 9780470747995, 2009.

Author's Address

   Carsten Bormann (editor)
   Universitaet Bremen TZI
   Postfach 330440
   Bremen  D-28359
   Germany

   Phone: +49-421-218-63921
   Email: cabo@tzi.org











Bormann                 Expires August 29, 2013                [Page 34]