Internet DRAFT - draft-guttman-nits-reqts

draft-guttman-nits-reqts



HTTP/1.1 200 OK
Date: Tue, 09 Apr 2002 00:13:22 GMT
Server: Apache/1.3.20 (Unix)
Last-Modified: Wed, 23 Jun 1999 14:35:00 GMT
ETag: "2e9a90-9c2c-3770f094"
Accept-Ranges: bytes
Content-Length: 39980
Connection: close
Content-Type: text/plain

Internet Engineering Task Force                               E. Guttman
INTERNET DRAFT                                                    Editor
                                                        Sun Microsystems
                                                           23 June  1999


                           NITS Requirements
                    draft-guttman-nits-reqts-00.txt


Status of This Memo

   This document is a submission by the NITS Working Group of the
   Internet Engineering Task Force (IETF).  Comments should be submitted
   to the nits@merit.edu mailing list.

   Distribution of this memo is unlimited.

   This document is an Internet-Draft and is in full conformance with
   all provisions of Section 10 of RFC2026.  Internet-Drafts are working
   documents of the Internet Engineering Task Force (IETF), its areas,
   and its working groups.  Note that other groups may also distribute
   working documents as Internet-Drafts.

   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."

   The list of current Internet-Drafts can be accessed at:
      http://www.ietf.org/ietf/1id-abstracts.txt

   The list of Internet-Draft Shadow Directories can be accessed at:
      http://www.ietf.org/shadow.html.


Abstract

   Networks in the Small (NITS) are a particular class of networks.
   These networks may be established in the home, in small offices or
   even on an impromptu basis.  NITS networks typically do not have and
   should not be expected to have network infrastructure such as DHCP
   servers, DNS and the like.  These networks may have only occasional
   connectivity to larger networks where these services exist.

   Typical users of NITS networks do not have the skills nor desire to
   install, configure, administer or maintain networking software.  To
   enable IP networking in the NITS environment, protocols for automatic
   configuration of hosts and discovery of services are needed which






Guttman, Editor             Expires 23 January 2000             [Page i]

Internet Draft   Requirements for Networks In The Small    23 June  1999


   require as little memory and processor resources on hosts which
   employ them as possible.

   This document presents requirements for host configuration and
   service discovery on NITS networks.  This document also addresses how
   hosts which implement NITS protocols must behave when they discover
   that they are on non-NITS networks.  This will ensure that NITS
   protocols do not disrupt normal network operations.


1. Contributing Authors

   The following authors (in alphabetical order) have contributed
   to this document:  Erik Guttman (Sun Microsystems), Myron Hattig
   (Intel), Brent Miller, Thomas Narten, Marcia Peters (IBM Corp.), Bob
   Quinn (Stardust Technologies) and John Tavs (IBM Corp.).




































Guttman, Editor            Expires 23 January 2000             [Page ii]

Internet Draft   Requirements for Networks In The Small    23 June  1999


2. Introduction

   This document will define and state the requirements for hosts
   operating in Networks in the Small (NITS). These are a particular
   class of networks which may be established in the home, in small
   offices or even on an impromptu basis.  NITS networks typically do
   not have and should not be expected to have network infrastructure
   such as DHCP servers, DNS and the like.  These networks may only
   have occasional connectivity to larger networks where these services
   exist.

   Typical users of NITS networks do not have the skills nor desire to
   install, configure, administer or maintain networking software.  For
   this reason, it is imperative that NITS networks not require any
   manual configuration for host parameters and that services may be
   discovered easily and with the minimum of required user interaction.

   Hosts present on NITS networks may have very reduced capabilities.
   For this reason, the protocols used in that setting must require as
   little memory and processor resources on hosts which employ them as
   possible.

   This document begins with a presentation of the terminology used
   later on.  The general goals of this document are briefly summarized
   in Section 4.

   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 [18].


3. Terminology

      Address Assignment

         Hosts require addresses in order to operate on IP networks.
         When hosts lack a fixed address, they must acquire their
         address via network protocols which assign it.

      Announcement

         This is a broadcast or multicast message which informs all
         (or some subset) of hosts on a network of something.  These
         messages are useful for passive automatic configuration.

      Client

         A client is a process which communicates with a server
         process running on the network.  The client may initiate this



Guttman, Editor             Expires 23 January 2000             [Page 1]

Internet Draft   Requirements for Networks In The Small    23 June  1999


         communication by sending messages over IP. Some client software
         obtains service without initiating it.  For example NTP clients
         may simply join a multicast group and listen for messages.

      Dynamic Configuration

         These are parameters which are obtained on demand through the
         use of network protocols such as DHCP [8] and MADCAP [21].

      Host

         An IP enabled device connected to a network which conforms
         to host requirements [15].  This need not be a 'computer,'
         the definition applies to any other device capable of network
         communication.

      Host Configuration

         These are parameters that a host needs in order to function
         on an IP network.  Specifically, a host needs to know its IP
         address, the subnet mask, the addresses of one or more gateway
         routers and the addresses of one or more DNS servers.  There
         may be no gateway router or DNS server present, but this is
         itself critical host configuration information.

      Interface

         A host's attachment point to a link.  It is possible (though
         unusual) for a host to have more than one interface to the
         same link.  Interfaces are uniquely identified by IP unicast
         addresses; a single interface may have more than one such
         address.

      Larger Network

         A larger network is in contrast to NITS. Larger networks deploy
         DHCP servers or hosts with static host configuration.  Larger
         networks deploy routers, DNS servers and require network
         administration.

      Link

         A communication facility or medium over which hosts can
         communicate at the link layer, i.e., the protocol layer
         immediately below IP.

      Neighboring

         Having an IP address belonging to the same subnet.



Guttman, Editor             Expires 23 January 2000             [Page 2]

Internet Draft   Requirements for Networks In The Small    23 June  1999


      Multicast interface

         An interface to a multicast link, that is, an interface to
         a link over which IP multicast or IP broadcast service is
         supported.

      Multicast link

         A link over which IP multicast or IP broadcast service is
         supported.

      Name Resolution

         The mapping of a device identifier to a (user-friendly)
         name, such as is done in DNS [1] (and can be done in other
         protocols).

      Router

         A host that forwards IP datagrams, as specified.

      Service

         A service is a process running on a host which receives and
         responds to messages transmitted over IP, using a particular
         transport protocol and port number.  It is possible that a
         service is not request/reply oriented.  For example if a
         service spontaneously multicasts to a particular group.

      Service Advertisement

         A service which employs a service discovery protocol to
         make itself known 'advertises' its services.  The service
         advertisement may be passive (in response to solicitations
         for service) or active (where the service announces itself)
         depending on the particulars of the service discovery
         protocol.  The service advertisement may or may not include the
         characteristics (attributes) of a service, depending on which
         protocol is used for service discovery.

      Service Characteristics

         Service characteristics may include server capabilities, status
         and operational information (such as contact information for
         repairs).  These characteristics may be defined so as to be
         humanly readable or to enable software only service discovery.
         If a client may use service characteristics as part of Service
         Discovery, the client may find a server which matches its




Guttman, Editor             Expires 23 January 2000             [Page 3]

Internet Draft   Requirements for Networks In The Small    23 June  1999


         capabilities or requirements.  If Service Discovery does not
         characteristics

      Service Discovery

         A client requires the location and other parameters of the
         service in order to contact it.  A client employs a service
         discovery protocol to acquire these parameters.

      Smaller Network

         A smaller network is 'autonomous.'  The network is not a
         logical part of a greater, centrally administered network.
         Implication:  One cannot assume the presence of services,
         planning of network address space, or any administrative
         configuration.  Also, the network may or may not have a gateway
         router.  A smaller network is also 'Limited in Scale.'  The
         network does not serve more than a certain size community, such
         as one would find in a home or small office.  There may be many
         services on the network, but few active clients.  Implication:
         NITS mechanisms may not work well if one attempts to use them
         on a campus or in a large office.  If these assumptions are not
         true, eventually scalability considerations will emerge and a
         Larger Network will be required.

      Solicitation

         This is a broadcast or multicast message which elicits replies
         from all hosts which recognize the request applies to them.
         Solicitations are useful for active automatic configuration.

      Static Configuration

         These are parameters which are entered manually or recorded in
         persistent storage.

      Subnet

         Either a single subnet of a subnetted IP network or a single
         non-subnetted IP network, i.e., the entity identified by an IP
         address logically ANDed with its assigned subnet mask.  More
         than one subnet may exist on the same link.

      User

         A person who interacts with client software.  Software
         processes with no human interface may also be a 'user' of
         client software.




Guttman, Editor             Expires 23 January 2000             [Page 4]

Internet Draft   Requirements for Networks In The Small    23 June  1999


4. General Goals

     - Create no new protocols unless they're needed.

     - Don't break existing networks.

     - Make IP networking easier to the point where newly added devices
       'just work.'


5. Scenarios and Applicability

   This section describes the applicability of the requirements defined
   in this document.


5.1. The home networking scenario

   To illustrate the problem statement that leads to requirements, a
   home networking scenario is presented.  Generically, this scenario
   considers the introduction of a new device into a home network.

   The specific scenario presented here is that of a new computer being
   introduced to an existing home network (other scenarios could also
   be developed).  The newly introduced computer, to be fully utilized,
   needs to become a participant in the "home area network".  It needs
   to make itself known to the other network participants, including
   services that it can offer to them (perhaps it is capable of router
   functions, or has specific software services that other devices
   can use).  The computer also needs to become aware of the other
   network participants, including services that they offer that may
   be of interest (perhaps another computer in the home network offers
   printing services or image capture services; the newly introduced
   computer also needs to find the network access point to the larger
   network).

   In a home environment, these functions need to be self-configuring
   (accomplished with minimal user setup and configuration).  This
   process should "just work", not requiring extensive user knowledge of
   the network nor onerous configuration or administration processes.

   This scenario assumes a home environment, although as noted above,
   the problem statement and requirements that arise from this scenario
   are likely to apply to other similar environments.  For this
   scenario, we assume that:

    1.  The home network is connected to a larger network (like the
        Internet), but only intermittently. The home network needs
        to be fully usable whether or not it is connected to the



Guttman, Editor             Expires 23 January 2000             [Page 5]

Internet Draft   Requirements for Networks In The Small    23 June  1999


        larger network.

    2.  The devices in the home network use IETF protocols, having
        at least IP support.  The devices share a physical layer
        connection.



5.2. Scalability

   In this scenario, there are not the same scalability requirements
   that would apply in larger networks.  It is possible to use multicast
   based discovery protocols that floods the entire network, for
   example, where in a corporate network that would not be acceptable.
   Still, there is a limit to the amount of bandwidth any operational
   protocol on NITS networks should claim.

   Small networks may have limited bandwidth available due to the
   performance of the underlying physical layer.

   There may also be many hosts on a home network.  It may be the case
   that a home network has hundreds or thousands embedded servers.  What
   distinguishes a small network from a large one is:

    1.  There are a limited number of users (that is people making use
        of client applications).  Though there may potentially be many
        services, few will be in use at any given time.

    2.  Small networks will lack network administration and
        infrastructure which provides scalable networking.  For
        examples, bridges and routers may not be deployed.


6. Automatic Configuration

   Automatic Configuration provides hosts with the ability to operate
   in the absence of static configuration and other services found in a
   larger network.


6.1. Host Configuration

   The fundamental configuration that a host needs to operate in a
   small network is host configuration.  Dynamic host configuration is
   obtained using DHCP [8].  It MUST also be possible to obtain host
   configuration in the absense of DHCP servers.






Guttman, Editor             Expires 23 January 2000             [Page 6]

Internet Draft   Requirements for Networks In The Small    23 June  1999


6.2. Service Discovery

   Clients on a NITS network MUST be able to obtain service
   configuration information (principally the location of services)
   through the use of a service discovery protocol.  This service
   discovery must be both interoperable with different services from
   different vendors and afford some security:  It MUST be possible to
   distinguish between a 'legitimate' service advertisement and one
   which is not.

   In larger networks, a 'centralized repository' or 'registry' for
   service information provides scalability for service discovery
   protocols.  The service discovery protocol SHOULD enable hosts to
   discover a registry automatically (if one exists), so that subsequent
   service lookup and advertising operations can be done in a scalable
   manner.

   In the past service discovery protocols which did not address these
   considerations caused scalability problems in larger networks (viz.
   SAP [6], NBP [7], SMB [5]).


6.3. Service Advertising

   Devices with services must be able to advertise themselves on the
   network.  This service advertisement must be both interoperable with
   different clients from different vendors and afford some security.
   The service advertisement MUST be made in such a way that the client
   can determine whether the service advertisement is legitimate or not.


6.4. Client interaction with Services

   The protocols which clients use to contact and communicate with
   services are out of scope of this specification.  NITS protocol
   requirements ensure that essential configuration is obtainable from
   the network.  Application layer protocols SHOULD then function
   equally well whether they are in a smaller or larger network.


6.5. Name resolution

   DNS name resolution is a fundamental service on IP networks.
   Although it is possible to perform some application layer operations
   using only IP addresses, hosts SHOULD be able to perform DNS name
   resolution even in the absense of DNS servers.

   If a DNS server is available it MUST be used in preference to a NITS
   DNS resolution mechanism.



Guttman, Editor             Expires 23 January 2000             [Page 7]

Internet Draft   Requirements for Networks In The Small    23 June  1999


7. Behavior when connected to a larger network

   Hosts in a NITS network which is connected or reconnected with a
   larger network MUST be reconfigured subsequently.

   This reconfiguration has at least two results:  Host configuration
   will change (as hosts become aware of a gateway router, etc.)
   Second, protocols which are host requirements on smaller networks
   but are not applicable on larger networks MUST NOT be used.  NITS
   protocols which are not appropriate for larger networks MUST include
   an applicability statement.

   Where possible, the same protocols SHOULD be requirements for use in
   smaller and larger networks as this will simplify the implementation
   and operation of networking software.


7.1. Scalability Considerations

   Note that in order to enable service discovery, device discovery
   using protocol layers 1-4 is required.  Before service discovery can
   be accomplished, it is probably necessary to learn a layer-2 address
   such as a MAC address or virtual circuit identifier, and/or a layer-3
   address like an IP address.  In considering the device discovery
   process, it is important to understand the scope of the message --
   how far it can travel and what blocks it.

   (Consider a layer-2 service such as Ethernet broadcast, a layer
   3 service such as IP multicast, and layer 4 services like a UDP
   datagram to a NDS or a NetBIOS Name Server or a WINS server.  Based
   upon the network topology, configuration and policies, broadcast or
   multicast messages might be locally confined, or they might traverse
   multiple routers, perhaps being limited by filters or hop counts).
   Scoping is important in discovering that a target device exists
   without unnecessarily broadcasting outside the required scope and
   potentially "flooding" larger and larger networks.  Furthermore,
   there may be multiple target devices of interest (and even invalid
   duplicate devices) that appear within different scopes.  Scoping is
   also important in the area of security.  In developing the device
   and service discovery requirements and potential solutions for home
   or similar environments, where "the network" may not resemble the
   Internet model, scoping is an important consideration.


7.2. Use of Multicast

   Global multicast addresses SHOULD NOT be used in NITS protocols
   since it is possible that messages intended only for the local
   unadministered network might escape and be routed to larger networks.



Guttman, Editor             Expires 23 January 2000             [Page 8]

Internet Draft   Requirements for Networks In The Small    23 June  1999


   Administrative scoped multicast addresses [22] are intended to
   localize multicast traffic and SHOULD be used in NITS protocols.

   Although NITS protocols are intended to be run on small networks
   they MUST not make use of multicast which would present scalability
   problems if they were connected to a larger network.  This could mean
   that NITS protocols always behave in a way that would be appropriate
   on a larger network or that hosts discover they are no longer on a
   smaller network and change their behavior.


8. Enumerated Requirements for NITS hosts

   The following list summarizes the requirements presented in previous
   sections.

    1. Hosts MUST be capable of dynamic configuration of host
       configuration, whether DHCP servers or relays are present or not.

    2. Hosts MUST be capable of being configured by DHCP so that they
       can be reconfigured if a NITS network is (re)attached to a larger
       network.

    3. Clients MUST be capable of obtaining service configuration
       information (principally the location of services) through the
       use of a service discovery protocol.

    4. Services MUST be discoverable by means of a service discovery
       protocol.

    5. Hosts SHOULD be able to resolve domain names even in the absense
       of a DNS server.

    6. Protocols specified for use in NITS networks which include an
       applicability statement which states they are not appropriate
       for use in larger networks MUST NOT be used in larger networks.
       There MUST be a way for hosts to detect when it is inappropriate
       for the protocol to be used.

    7. Global multicast addresses SHOULD NOT be used in NITS protocols.
       Administratively scoped multicast addresses SHOULD be used in
       NITS protocols.


9. Protocols which fulfill NITS requirements

   This document defines the requirements for NITS networking.  It does
   not enumerate the protocols which fulfill those requirements.




Guttman, Editor             Expires 23 January 2000             [Page 9]

Internet Draft   Requirements for Networks In The Small    23 June  1999


   It is worth noting, however, that for host configuration there
   are well understood protocols which will likely be part of a NITS
   networking host requirement document.  These include stateless
   address configuration for IPv4 [10] and IPv6 [12].


10. Open Issues

10.1. Is DNS resolution in NITS a requirement?

   Is DNS resolution on a NITS network really necessary?  If services
   can be discovered by means of a service discovery protocols, and
   locations transmitted in numerical IP addresses - why are names
   required?  There is no name space management in a NITS network.
   There is no hierarchy of domains to consider.  Assuming DNS
   resolution is possible on NITS in a decentralized fashion [3], how is
   the name space going to be kept unique?  What happens when two hosts
   claim the same name?

   What should we say about static configuration of DNS (if someone
   manually enters host to IP address tables)?


10.2. Multicast, Administrative Multicast and MADCAP

   Should multicast support be a requirement for hosts in NITS networks?
   Multicast seems to be required to support service discovery and other
   services like Multicast DNS [3] which are under discussion.

   Should MADCAP [21] be used so that hosts can allocate and use
   administrative scoped multicast addresses [22] from the smallest
   enclosing scope by default?  This would go a long way toward making
   NITS protocols more scalable and protecting larger networks from
   getting spill-over.  Note that a minimal MADCAP client is really
   quite small.

   It may also be worthwhile discussing how MASC-like [4] claim and
   defend for multicast addresses could be used in NITS when MADCAP
   servers are not present.


10.3. Behavior upon receiving router advertisements

   What should hosts do when they have employed stateless address
   configuration and see router advertisements [14]?  The hosts should
   have non-routable addresses.  Still, should they behave differently
   when they detect a transition (when a router appears or disappears?)





Guttman, Editor            Expires 23 January 2000             [Page 10]

Internet Draft   Requirements for Networks In The Small    23 June  1999


10.4. IP address autoconfiguration and Ad Hoc networking

   Stateless address configuration for IPv4 [10] and IPv6 [12] both
   involve claim and defend mechanisms.  These mechanisms will work
   poorly in Ad Hoc networks [23].  MANET protocols assume all hosts
   have unique addresses in order to establish routing.  Ad Hoc networks
   could benefit from NITS protocols in other respects as they by
   definition lack infrastructure and infrastructure and may be quite
   small.  If stateless address configuration is a requirement for hosts
   on NITS networks, a consequence may be that NITS may not apply to Ad
   Hoc networks.


10.5. NITS, NAT, VPN, Securing a small network

   Should NITS networking requirements include a specific list of VPN
   and NAT protocol and operational requirements?

   Should this document describe address sharing (a topic related to
   NAT)?

   What should be done about securing NITS services from abuse or
   illegitimate use from clients in larger networks to which the smaller
   network is connected?  What are the security requirements for clients
   and services on a NITS network?  These cannot be very sophisticated
   since a primary goal is ease of use and unskilled operators.

   A possible list of requirements was proposed:

    1. Allow communication from host to a corporate LAN via VPN.

    2. Allow communication from host to another home-network via VPN.

    3. Allow hosts in the home to simultaneously access the Internet
       using 1 or 2 globally unique IP addrs.

    4. Allow access from the Internet to a home-network server (e.g.
       household WEB server).

    5. Allow access from the Internet to multiple home servers of a
       single service (e.g.  personal WEB servers).

    6. Protect devices, resources, and communication in the home-network
       from Internet attacks.

    7. Protect communication when traversing the Internet to access
       corporate LAN or another home-network.





Guttman, Editor            Expires 23 January 2000             [Page 11]

Internet Draft   Requirements for Networks In The Small    23 June  1999


10.6. How big is small?

   The approximate scale of a NITS network has been discussed on the
   NITS mailing list.  The very rough consensus is that it is around
   10 users, 100s of hosts and 1000s of services, or smaller.  If the
   network has more than these numbers of entities it is not small.  Of
   course in small networks today we have fewer hosts and services, but
   we have to consider what will happen *if we are successful.* What if
   there are lots of devices enabled with the NITS protocols fulfilling
   the requirements in this document?

   Is it appropriate to include this discussion in the body of the
   requirements document?  It has been suggested that we focus on
   functionality and not debate quantities in preparing this document.


10.7. Networks in the Simple?

   Should we focus on 'Simple' instead of 'Small?'  The difference is
   instead of considering scaling issues, we are encouraged to look
   at a different kind of host and a different kind of network.  For
   instance, we could specifically define this as Home Networks so we
   could talk about Home Appliance requirements.  Should we focus on
   usability issues perhaps instead of how larger networks are to be
   protected from poorly scalable NITS mechanisms?

   Focusing on Small instead of Simple puts the focus on operation
   without infrastructure and with reduce number of active client hosts.


10.8. Do we need a 'Scenarios' or 'Applications' Section?

   There has been discussion of specific scenarios and applications of
   NITS protocols.  Would it help the clarity and motivation for the
   reader of this document to include this after the introduction?


10.9. Gateway requirements and small networks

   Should there be requirements on a gateway such that it automatically
   (a) provides either DHCP or DHCP relay service at the least, (b)
   ensures that hosts on the NITS network are configured so as to be
   able to reach hosts off of the smaller network.  Should this document
   define gateway requirements or only host requirements?








Guttman, Editor            Expires 23 January 2000             [Page 12]

Internet Draft   Requirements for Networks In The Small    23 June  1999


10.10. Is service discovery conclusive?

   When service discovery is successful, should the client be able to
   use the service, or do we assume that it may not be able to.  In the
   former case service discovery finds only those services which match
   the client's capabilities and requirements.  In the latter case the
   client may have to perform feature negociation with the service -
   the result of which may be that the client is not able to use the
   service.


10.11. Device identification and discovery

   Should the following terms be in the terminology section?  They are
   problematic to define and have confused most readers.

      Device Discovery

         This is the protocol allowing the resolution of an identifier
         to obtain the location of a device.  For example, a host with
         the destination IP address of another host may require its MAC
         address in order to send a datagram.  Protocols which enable
         device discovery in this sense include ARP [24] for IPv4 and
         Neighbor Discovery [13] for IPv6.

      Device Identification

         This is any unique identifier belonging to a host attached to
         a network.  An example of such an identifier is an IP address
         or a MAC layer address.  It is possible that other identifiers
         may be used in NITS protocols (ie.  a device supports a SNMP
         agent).

   It has been suggested that this section is an absolute must:

       If devices are not self descriptive then you need an external
       directory in which to look up an ID number (which will need to
       be relatively unique).  Having a common language for describing
       capabilities is the only way you can get interoperability of
       autonomous devices.


10.12. Browsing and human friendly identifiers

   Should this requirement document specify how human readable
   identifiers should be created to enable interoperability of browsers?
   This amounts to requirements for human interfaces to the service
   discovery protocol.  Some issues here are:




Guttman, Editor            Expires 23 January 2000             [Page 13]

Internet Draft   Requirements for Networks In The Small    23 June  1999


      Service type names:

         To browse all the speakers in my house the service
         advertisement of each speaker has to be consistent with some
         standards.  Only then can a service browser produce a listing
         of speakers from different vendors, interoperably.

      Internationalization:

         Any strings which are available from NITS protocols which are
         intended for human readability (ie.  in a browser) MUST be
         internationalizable.  What does that mean, though?  Are IDs
         passed around for which there are standardized translations?
         Should the protocols be able to support strings with different
         languages?

      Attributes:

         Browsers which offer directory (or directory-like) functions
         may provide attributes for viewing in a browser.  This is in
         contrast to providing only the name and descriptive text.
         Attributes also make sophisticated server selection possible
         as part of the human interface of the browser.  The problem
         is:  If attributes are part of service advertisements, they
         have to be standardized for interoperability.  Should the NITS
         requirements document describe how this is to be done?


11. IANA Considerations

   There are no known IANA considerations arising from NITS protocol
   requirements.


12. Internationalization Considerations

   This document describes interoperability requirements for hosts on
   NITS networks.  These include protocols which exchange human readable
   strings.  Requirements which involve human readable strings include
   considerations for internationalizing those strings.


13. Security Considerations

   NITS protocols will enable a great deal of automatic configuration.
   These present numerous opportunities for malicious masquerading as
   legitimate services.  NITS protocols must consider how to balance
   some protection from attackers who seek to misconfigure hosts and the
   zero-configuration ideal.



Guttman, Editor            Expires 23 January 2000             [Page 14]

Internet Draft   Requirements for Networks In The Small    23 June  1999


   Some of the requirements defined by this document provide some
   security properties to protect client hosts in NITS networks.  These
   include the requirement that the service discovery protocol enable
   a client to determine if a service advertisement is 'legitimate' or
   not.

   It is clear that this document needs to define the class of threats
   which are present when NITS protocols are employed.  Do these threats
   arise from adversaries *on* the NITS network?  Do we have to consider
   threats which arise from adversaries in a larger network which is
   impermanently connected to the smaller network?









































Guttman, Editor            Expires 23 January 2000             [Page 15]

Internet Draft   Requirements for Networks In The Small    23 June  1999


14. Full Copyright Statement

   Copyright (C) The Internet Society (1997).  All Rights Reserved.

   This document and translations of it may be copied and furnished to
   others, and derivative works that comment on or otherwise explain it
   or assist in its implementation may be prepared, copied, published
   and distributed, in whole or in part, without restriction of any
   kind, provided that the above copyright notice and this paragraph
   are included on all such copies and derivative works.  However,
   this document itself may not be modified in any way, such as by
   removing the copyright notice or references to the Internet Society
   or other Internet organizations, except as needed for the purpose
   of developing Internet standards in which case the procedures
   for copyrights defined in the Internet Standards process must be
   followed, or as required to translate it into languages other than
   English.

   The limited permissions granted above are perpetual and will not be
   revoked by the Internet Society or its successors or assigns.

   This document and the information contained herein is provided on an
   "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
   TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING
   BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION
   HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF
   MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE."


References

   [1] P. Mockapetris  Domain Names - Concepts and Facilities  RFC 1034,
       November 1987

   [2] A. Gulbrandsen, P. Vixie  A DNS RR for specifying the location of
       services (DNS SRV)  RFC 2052, October 1996.

   [3] B. Woodcock, B. Manning  Multicast Discovery of DNS Services
       draft-manning-multicast-dns-01.txt  December, 1998. A work in
       progress.

   [4] D. Estrin, et. al.  The Multicast Address-Set Claim (MASC)
       Protocol  draft-ietf-malloc-masc-01.txt  February, 1999. A work
       in progress.

   [5] Microsoft Networks  SMB File Sharing Protocol Extensions 3.0
       Document Version 1.09, November 1989





Guttman, Editor            Expires 23 January 2000             [Page 16]

Internet Draft   Requirements for Networks In The Small    23 June  1999


   [6] Novell, Inc.  IPX RIP and SAP Router Specification  Part Number
       107-000029-001, Version 1.30, May 1996

   [7] S. Gursharan, R. Andrews, and A. Oppenheimer  Inside AppleTalk
       Addison-Wesley, 1990.

   [8] R. Droms.  Dynamic Host Configuration Protocol.  RFC 2131, March
       1997.

   [9] S. Alexander, R. Droms  DHCP Options and BOOTP Vendor Extension
       RFC 2132, March 1997.

   [10] R. Troll  Automatically Choosing an IP Address in an Ad-Hoc IPv4
       Network  draft-ietf-dhc-ipv4-autoconfig-04.txt  April, 1999. A
       work in progress.

   [11] R. Troll  DHCP Option to Disable Stateless Auto-Configuration
       in IPv4 Clients  draft-ietf-autoconfig-04.txt  February, 1999. A
       work in progress.

   [12] S. Thomson, T. Narten  IPv6 Address Autoconfiguration  RFC 2462,
       December 1998.

   [13] T. Narten, E. Nordmark, W. Simpson  Neighbor Discovery for IP
       Version 6 (IPv6)  RFC 2461, December 1998.

   [14] S. Deering  ICMP Router Discovery Messages  RFC 1256, September
       1991.

   [15] R. Braden  Requirements for Internet Hosts -- Communication
       Layers  RFC 1122, October 1989

   [16] E. Guttman, C. Perkins, J. Veizades, and M. Day.  Service
       Location Protocol version 2.  RFC 2608, June 1999.

   [17] E. Guttman, C. Perkins, J. Kempf  Service Templates and service:
       Schemes  RFC 2609, June 1999.

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

   [19] H. Alvestrand  Language Tags  RFC 1776, July 1994.

   [20] F. Yergeau.  UTF-8, a transformation format of unicode and ISO
       10646.  RFC 2279, October 1998.

   [21] iS. Hanna, B. Patel, M. Shah  Multicast Address Dynamic Client
       Allocation Protocol (MADCAP)  draft-ietf-malloc-madcap-05.txt
       May 1999. A work in progress.



Guttman, Editor            Expires 23 January 2000             [Page 17]

Internet Draft   Requirements for Networks In The Small    23 June  1999


   [22] D. Meyer  Administratively Scoped Multicast Address  RFC 2365,
       July 1998.

   [23] IETF MANET WG  MANET WG Charter  http://www.ietf.org/

   [24] D. Plummer  An Ethernet Address Resolution Protocol  RFC 826,
       November 1982













































Guttman, Editor            Expires 23 January 2000             [Page 18]

Internet Draft   Requirements for Networks In The Small    23 June  1999


Editor's Address

   Questions about this memo can be directed to:

  Editor:

      Erik Guttman
      Sun Labs - Networking and Security - Advanced Development
      Sun Microsystems, Inc.
      Bahnstr. 2
      74915 Waibstadt
      Germany

      phone: +49 7263 911 701
      or:     +1 650 786 5992
      email: Erik.Guttman@Sun.Com




































Guttman, Editor            Expires 23 January 2000             [Page 19]