Extensible Authentication Protocol                              J. Arkko
Internet-Draft                                                  Ericsson
Expires: July 10, 2004                                    B. Aboba, Eds.
                                                        January 10, 2004

                Network Discovery and Selection Problem

Status of this Memo

   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 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://

   The list of Internet-Draft Shadow Directories can be accessed at

   This Internet-Draft will expire on July 10, 2004.

Copyright Notice

   Copyright (C) The Internet Society (2004).  All Rights Reserved.


   The so called network discovery and selection problem affects network
   access, particularly in the presence of multiple available wireless
   accesses and roaming.  This problem has been the subject of
   discussions in various standards bodies.  This document summarizes
   the discussion held about this problem in the Extensible
   Authentication Protocol (EAP) working group at the IETF.  The problem
   is defined and divided into subproblems, and some constraints for
   possible solutions are outlined.  The document presents also some
   existing mechanisms which help solve at least parts of the problem,
   and gives some suggestions on how to proceed for the rest.

Arkko & Aboba, Eds.      Expires July 10, 2004                  [Page 1]

Internet-Draft    Network Discovery and Selection Problem   January 2004

Table of Contents

   1.    Introduction . . . . . . . . . . . . . . . . . . . . . . . .  3
   2.    Problem Definition . . . . . . . . . . . . . . . . . . . . .  4
         2.1   Access Network Discovery . . . . . . . . . . . . . . .  4
               2.1.1 Issues with Access Network Discovery . . . . . .  5
         2.2   Identity selection . . . . . . . . . . . . . . . . . .  5
         2.3   AAA routing  . . . . . . . . . . . . . . . . . . . . .  7
               2.3.1 Issues with AAA Routing  . . . . . . . . . . . .  8
         2.4   Payload Routing  . . . . . . . . . . . . . . . . . . . 10
               2.4.1 Issues with Payload Routing  . . . . . . . . . . 10
         2.5   Discovery, Decision, and Selection . . . . . . . . . . 11
   3.    Design Constrains  . . . . . . . . . . . . . . . . . . . . . 13
   4.    Existing Work  . . . . . . . . . . . . . . . . . . . . . . . 14
         4.1   IETF . . . . . . . . . . . . . . . . . . . . . . . . . 14
         4.2   IEEE . . . . . . . . . . . . . . . . . . . . . . . . . 15
         4.3   3GPP . . . . . . . . . . . . . . . . . . . . . . . . . 16
         4.4   Other  . . . . . . . . . . . . . . . . . . . . . . . . 17
   5.    Conclusions  . . . . . . . . . . . . . . . . . . . . . . . . 18
   6.    Security Considerations  . . . . . . . . . . . . . . . . . . 21
         Normative References . . . . . . . . . . . . . . . . . . . . 22
         Informative References . . . . . . . . . . . . . . . . . . . 23
         Authors' Addresses . . . . . . . . . . . . . . . . . . . . . 24
   A.    Contributors . . . . . . . . . . . . . . . . . . . . . . . . 25
         Intellectual Property and Copyright Statements . . . . . . . 26

Arkko & Aboba, Eds.      Expires July 10, 2004                  [Page 2]

Internet-Draft    Network Discovery and Selection Problem   January 2004

1. Introduction

   The so called network discovery and selection problem affects network
   access and wireless access networks in particular.  This problem
   comes relevant when any of the following conditions are true:

   o  There is more than one available network attachment point, and the
      different points may have different characteristics.

   o  The user has multiple sets of credentials.  For instance, the user
      could have one set of credentials from a public service provider
      and another set from his company.

   o  There is more than one way to provide roaming between the access
      and home network, and service parameters or pricing differs
      between them.  For instance, the access network could have both a
      direct relationship with the home network and an indirect
      relationship through a roaming consortium.

   o  The roaming relationships between access and home networks are so
      complicated that current AAA protocols can not route the requests
      to the home network unaided, just based on the domain in the given
      Network Access Identifier (NAI) [4].

   o  Payload packets get routed or tunneled differently, based on which
      particular roaming relationship variation is used.  This may have
      an impact on the available services or their pricing.

   o  Providers share the same infrastructure, such as wireless access

   The network discovery and selection problem spans multiple protocol
   layers and has been the subject of discussions in IETF, 3GPP, and
   IEEE 802.11.  This document summarizes the discussion held about this
   problem in the Extensible Authentication Protocol working group at

   In Section 2 the problem is defined and divided into subproblems, and
   some constraints for possible solutions are outlined in Section 3.
   Section 4 discusses existing mechanisms which help solve at least
   parts of the problem.  Section 5 gives some suggestions on how to
   proceed for the rest.

Arkko & Aboba, Eds.      Expires July 10, 2004                  [Page 3]

Internet-Draft    Network Discovery and Selection Problem   January 2004

2. Problem Definition

   There are four somewhat orthogonal problems being discussed under the
   rubric of "network discovery and selection".

   o  First, there is the problem of "Access Network discovery".  This
      is the problem of discovering access networks available in the
      vicinity, and the points of presence (POPs) associated with those

   o  Second, there is the problem of "Identifier selection".  This is
      the problem of selecting which identity (and credentials) to use
      to authenticate to a given POP.

   o  Thirdly, there is the problem of "AAA routing" which involves
      figuring out how to route the authentication conversation
      originating from the selected identity back to the home realm.

   o  Finally, there is the the problem of "Payload routing" which
      involves figuring how the payload packets are routed, where more
      advanced mechanisms than destination-based routing is needed.

   Alternatively, the problem can be divided to the discovery, decision,
   and the selection components.  The AAA routing problem, for instance,
   involves all components: discovery (which mediating networks are
   available?), decision (choose the "best" one), and selection (this is
   the chosen network) components.

2.1 Access Network Discovery

   The Access Network Discovery problem has been extensively studied,
   see for instance the results of the IETF Seamoby WG, IEEE
   specifications on 802.11 wireless LAN beaconing and probing process,
   studies (such as [29]) on the effectiveness of these mechanisms, and
   so on.

   Traditionally, the problem of discovering available networks has been
   handled as a part of the link layer attachment procedures, or through
   out-of-band mechanisms.

   RFC 2194 [3] describes the pre-provisioning of dialup roaming
   clients, which typically included a list of potential phone numbers,
   updated by the provider(s) with which the client had a contractual
   relationship.  RFC 3017 [8] describes the IETF Proposed Standard for
   the Roaming Access XML DTD.  This covers not only the attributes of
   the Points of Presence (POPs) and Internet Service Providers (ISPs),
   but also hints on the appropriate NAI to be used with a particular
   POP.  The RFC supports dial-in and X.25 access, but has extensible

Arkko & Aboba, Eds.      Expires July 10, 2004                  [Page 4]

Internet-Draft    Network Discovery and Selection Problem   January 2004

   address and media type fields.

   In IEEE 802.11 WLANs, the Beacon/Probe Request/Response mechanism
   provides a way for Stations to discover Access Points (APs), as well
   as the capabilities of those APs.  Among the Information Elements
   (IEs) included within the Beacon and Probe Response is the SSID, a
   non-unique identifier of the network to which an Access Point is
   attached.  By combining network identification along with
   capabilities discovery, the Beacon/Probe facility provides the
   information required for both network discovery and roaming decisions
   within a single mechanism.

2.1.1 Issues with Access Network Discovery

   As noted in [28], the IEEE 802.11 Beacon mechanism does not scale
   well; with a Beacon interval of 100ms, and 10 APs in the vicinity,
   approximately 32 percent of an 802.11b AP's capacity is used for
   beacon transmission.  In addition, since Beacon/Probe Response frames
   are sent by each AP over the wireless medium, stations can only
   discover APs within range, which implies substantial coverage overlap
   for roaming to occur without interruption.

   A number of enhancements have been proposed to the Beacon/Probe
   Response mechanism in order to improve scalability and roaming
   performance.  These include allowing APs to announce capabilities of
   neighbor APs as well as their own, as proposed in IEEE 802.11k;
   propagation of discovery information over IP, as handled in the CARD
   protocol developed within the IETF SEAMOBY WG, etc.

   Typically scalability enhancement mechanisms attempt to get around
   Beacon/Probe Response restrictions by sending advertisements at the
   higher rates which are enabled one the station has associated with an
   AP.  Since these mechanisms run over IP, they can utilize IP
   facilities such as fragmentation, which the link layer mechanisms may
   not always be able to do.  For instance, in IEEE 802.11, Beacon
   frames cannot use fragmentation because they are multicast frames,
   and multicast frames are not acknowledged in 802.11.

2.2 Identity selection

   As networks proliferate, it becomes more and more likely that a given
   EAP peer may have multiple identities and credential sets, available
   for use in different situations.  For example, the EAP peer may have
   an account with one or more Public WLAN providers;  a corporate WLAN;
   one or more wireless WAN providers.  As a result, the EAP peer has to
   decide which credential set to use when presented with a given set of
   potential EAP authenticators.

Arkko & Aboba, Eds.      Expires July 10, 2004                  [Page 5]

Internet-Draft    Network Discovery and Selection Problem   January 2004

   Figure 1 illustrates a situation where the user does not know whether
   he is connected to access network 1, which only serves the ISP,
   access network 3, which only serves the company, or access network 2,
   which serves both.

                                     +---------+       +---------+
                                     | Access  |       |         |
                               _---->| Network |------>| isp.com |
                              /      |    1    |    _->|         |
                              |      +---------+   /   +---------+
                              |                   /
                              |      +---------+ /
   User "subscriber@          |      | Access  |/
   isp.com" also known as --- ? ---->| Network |
   "employee123@corp.com"     |      |    2    |\
                              |      +---------+ \
                              |                   \
                              |      +---------+   \_  +---------+
                              \_     | Access  |     ->|         |
                                ---->| Network |------>| corp.com|
                                     |   3     |       |         |
                                     +---------+       +---------+

         Figure 1: Two credentials, three possible access links

   Traditionally, hints useful in identity selection have also been
   provided out-of-band.  For example, via the RFC 3017 XML DTD [8], a
   client can select between potential POPs, and then based on
   information provided in the DTD, determine the appropriate NAI to use
   with the selected POP.

   Perhaps the most typical case is a link layer that provides some
   information about the network before EAP is initiated.  For instance,
   in IEEE 802.11 provides the SSID, though in some cases the client may
   not learn about all the SSIDs supported by the given access point.
   In IKEv2 [15], the identity of the responder (typically the security
   gateway) is provided in the same packet as the EAP Identity Request
   is transported.  In order to use this information in deciding the
   right identity to use, the provided information has to either match
   with one of the client's home networks, or the client has to have
   some other knowledge that enables to link the advertised network and
   the home network.  For instance, the client may be aware that his
   home network has a roaming contract with a given access network.

   It is also possible for hints to be embedded within credentials.  In
   [11], usage hints are provided within certificates used for wireless

Arkko & Aboba, Eds.      Expires July 10, 2004                  [Page 6]

Internet-Draft    Network Discovery and Selection Problem   January 2004

   authentication.  This involves extending the client's certificate to
   include the SSIDs with which the certificate can be used.

   Finally, some EAP implementations use the space after the NUL
   character in an EAP Identity Request to communicate some parameters
   relating to the network requesting EAP authentication.  However,
   there is no standard interpretation of the field beyond the NUL

2.3 AAA routing

   Once the identity has been selected, it is necessary for the
   authentication conversation to be routed back to the home realm.
   This is typically done today through the use of the Network Access
   Identifier (NAI), RFC 2486 [4], and the ability of the AAA network to
   route requests to the domain indicated in the NAI.

   Within the IETF ROAMOPS WG, a number of additional approaches were
   considered for this, including source routing techniques based on the
   NAI, and techniques relying on the AAA infrastructure.  Given the
   relative simplicity of the roaming implementations described in RFC
   2194 [3], static routing mechanisms appeared adequate for the task
   and it was not deemed necessary to develop dynamic routing protocols.

   As noted in RFC 2607 [5], RADIUS proxies are deployed not only for
   routing purposes, but also to mask a number of inadequacies in the
   RADIUS protocol design, such as the lack of standardized
   retransmission behavior and the need for shared secret provisioning.

   By removing many of the protocol inadequacies, introducing new AAA
   agent types such as Redirects, providing support for
   certificate-based authentication as well as inter and intra-domain
   service discovery, Diameter allows a NAS to directly open a Diameter
   connection to the home realm without having to utilize a network of
   proxies.  For instance, the Redirect feature could be used to provide
   a centralized routing function for AAA, without having to know all
   home network names in all access networks.

   This is somewhat analogous to the evolution of email delivery.  Prior
   to the widespread proliferation of the Internet, it was necessary to
   gateway between SMTP-based mail systems and alternative delivery
   technologies, such as UUCP and FidoNet, and email-address based
   source-routing was used to handle this.  However, as mail could
   increasingly be delivered directly, the use of source routing

   As with the selection of certificates by stations, a Diameter client
   wishing to authenticate with a Diameter server may have a choice of

Arkko & Aboba, Eds.      Expires July 10, 2004                  [Page 7]

Internet-Draft    Network Discovery and Selection Problem   January 2004

   available certificates, and therefore it may need to choose between

2.3.1 Issues with AAA Routing

   No dynamic routing protocols are in use in AAA infrastructure today.
   This implies that there has to be a device (such as a proxy) within
   the access network that knows how to route to different domains, even
   if they are further than one hop away, as shown in Figure 2.  In this
   figure, the user "joe@corp3.com" has to be authenticated through ISP
   2, since the domain "corp3.com" is served by it.

                 +---------+          +---------+
                 |         |          |         |
   User "joe     | Access  |--------->|  ISP 1  |-------> "corp1.com"
   @corp3.com"-->| Network |          |         |
                 |         |          +---------+
                    |         |--------> "corp2.com"
                    |  ISP 2  |
                    |         |--------> "corp3.com"

                Figure 2: AAA routing problem

   A related issue is that the roaming relationship graph may have
   ambiguous routes, as shown in Figure 3.  As billing is based on AAA
   and pricing may be based on the used intermediaries, it is necessary
   to select which route is used.  For instance, in Figure 3, access
   through the roaming group 1 may be cheaper, than if roaming group 2
   is used.  For commercial reasons, intermediaries want to participate
   the AAA transaction.

Arkko & Aboba, Eds.      Expires July 10, 2004                  [Page 8]

Internet-Draft    Network Discovery and Selection Problem   January 2004

                                     |         |---------> "isp1.com"
                                     | Roaming |
                 +---------+         | Group 1 |
                 |         |-------->|         |---------> "isp2.com"
   User "joe     | Access  |         +---------+
   @isp1.com"--->| Network |
                 |         |         +---------+
                 |         |-------->|         |---------> "isp1.com"
                 +---------+         | Roaming |
                                     | Group 2 |
                                     |         |---------> "isp3.com"

                Figure 3: Ambiguous AAA routing

   Currently planned networks include one level with a small number of
   intermediaries, just a few now and perhaps up to 50 as a maximum.
   However, multiple levels and higher number of alternative networks
   are possible in theory.

   There has been requests to place credential and AAA route selection
   under user control, as the user is affected by the pricing and other
   differences.  Optionally, automatic tools could make the selection
   based on the user's preferences.  On the other hand, user control is
   similar to source routing, and as discussed earlier, network-based
   routing mechanisms have traditionally won over source routing-based

   If users can control the selection of intermediaries, such
   intermediaries still have to be legitimate AAA proxies.  That is, an
   access network should not send a request to an unknown intermediary.
   If it has a business relationship with three intermediaries int1.com,
   int2.com, and int3.com, it will route your request through one of
   them, even if you tried to request routing through mitm.org.  Thus,
   NAI-based source routing is not source routing in the classic sense.
   It is merely suggesting preferences among already established routes.
   If the route does not already exist, or is not feasible, then
   NAI-based source routing cannot establish it.

   An additional issue is that even if the intermediaries are
   legitimate, they could be switched.  For instance, an access network
   could advertise that it has a deal with cheapintermediary.net, and
   then switch the user's selection to priceyintermediary.com instead.
   To make this relevant, the pricing would have to be based on the
   intermediary.  Even if it were possible to secure this selection, it
   would not be possible to guarantee that QoS or other properties

Arkko & Aboba, Eds.      Expires July 10, 2004                  [Page 9]

Internet-Draft    Network Discovery and Selection Problem   January 2004

   claimed by the network were indeed provided.  As a result, it may be
   useful to think of the intermediary selection only as a hint.

   Only a limited amount of information about AAA routes can be
   dynamically communicated.  It is necessary to retrieve network and
   intermediary names, but quality of service or pricing information is
   clearly something that would need to be pre-provisioned, or perhaps
   just available via the web.  Similarly, dynamic communication of
   network names can not be expected to provide all possible home
   network names, as their number can be quite large globally.

2.4 Payload Routing

   The access network, mediating AAA infrastructure, and the home
   network may all have a desire to affect the kind of treatment payload
   packets get.  This includes filtering, prioritization, routing paths,
   and mandatory tunneling.

   Traditionally, the involved entities have been able to provide some
   control of this through AAA protocols such as RADIUS [6] and Diameter
   [9].  RFC 2868 [7] defines tunneling attributes that the home and
   mediating networks can use to establish mandatory tunneling at the
   access network.  RFC 3588 [9] defines a filter syntax which can be
   used to install per-session filters under the control of AAA.

2.4.1 Issues with Payload Routing

   In an attack described by Michael Richardson, a rogue hotspot
   operator (but with a legitimate roaming relationship to a home
   network) steals revenues from local hotspot operator by spoofing its
   identity.  The rogue operator places a node with two interfaces in
   the coverage area of the local operator, and attaches to the Internet
   via this operator using a flat fee -based account.  It then starts to
   advertise itself as a hotspot operator on the other interface, luring
   unsuspecting clients to use this node rather the than the local
   operator.  The users authenticate via this node and its roaming
   relationship.  All AAA and payload traffic goes via the local
   hotspot, suitably NATted by the rogue node.  As a result, the rogue
   operator gets the roaming service fees for a number of clients,
   whereas the local operator gets just one client.

   Due to the way that the IEEE 802.11, IETF protocols, and common EAP
   methods have been designed, the rogue operator can actually advertise
   the same identity (SSID) as the local operator; the parameters
   advertised by the access point information are not authenticated
   end-to-end to the home network.  EAP methods that support channel
   bindings [10] do not have this problem, but this support is not
   present in commonly used methods.  Rogue access point can present a

Arkko & Aboba, Eds.      Expires July 10, 2004                 [Page 10]

Internet-Draft    Network Discovery and Selection Problem   January 2004

   different set of parameters to the client and to the home network.
   The current network access mechanisms enable us to have
   authentication, and link layer protection.  They do not, however,
   guarantee anything about the delivery of the actual payload packets.
   In particular, there is no guarantee that the payload packets are
   delivered through a right route, or NATed only up to some specific
   number of times.

   We call this the "payload route binding problem".  In this problem,
   authentication and link layer support do not guarantee that the
   packets are actually routed through the path that appears to have
   been offered.  Basically, if the "rogue" access point has a flat fee
   account and a contract with a clearing house, it can offer access to
   anyone with a per-byte account, NAT their packets, and send the
   traffic forward on its own account, and generate income.  The local
   ISP would not be able to detect this by looking at the traffic
   stream.  The user would not be able to detect it either, unless an
   EAP method with channel binding support is used.  And even if it is,
   the user may not care about the identity of the access point enough
   to look at it; channel bindings can only ensure that all parties
   agree about the parameters, not that they make sense.

   The working group did not come to a conclusion whether this attack
   needs to be prevented.  Some of the proposed solutions include the
   use of EAP EMSK in the authentication exchange with the DHCP server,
   or the use of EAP to provide the certificates that SEND requires for
   the authentication of Router Advertisements.  Either approach means
   that we are sure we are speaking at layer 3 to the services that we
   authenticated at layer two.  However, this does not prevent an
   attacker from offering such services, and then performing a NAT on
   the packets later.  However, IPsec could be used to detect the
   presence of such NATs, even if NAT traversal is in use.

2.5 Discovery, Decision, and Selection

   An alternative decomposition of the problem is to consider the
   discovery, decision, and selection aspects separately.  Discovery
   consists of discovering access networks and associated POPs,
   discovering what identities the access networks will accept (either
   directly or through roaming relationships), and discovering which
   potential AAA intermediaries or routes exist.

   Selection consists of attaching to the "right" access network and
   POP, offering an identity through EAP Identity Response, and
   providing a hint about the desired AAA intermediary.  The selection
   of the AAA intermediary, along with the home and access networks,
   determines also the treatment of payload packets.

Arkko & Aboba, Eds.      Expires July 10, 2004                 [Page 11]

Internet-Draft    Network Discovery and Selection Problem   January 2004

   Decision can be either manual selection or automatic.  Most likely,
   automatic mechanisms are preferred, even if manual selection should
   be retained as a fallback.  The type of the decision also places
   additional requirements on the type of information that the discovery
   phase must provide.  Just knowing which choices are available is
   probably enough for manual selection.  Unfortunately, automatic
   selection based on a list of choices is by itself not possible:

   o  Some access networks may be preferred over others.  For instance,
      the user's private corporate network may be preferred over a
      public network due to cost and efficiency reasons.

   o  Similarly, some credentials may be preferred over others.

   o  Use of an access network with direct connection to home network
      may be preferred over using mediating networks.

   o  Some mediating networks may be preferred to others, most likely
      based on cost.  Note that optimizing cost is not a trivial
      problem, because the total cost may be a combination of a fixed
      fee, per-minute, per-megabyte, volume discounts, and so on.

   o  Preferences may come from the user, his or her employer (who's
      paying the bill), home network, or access network.

   Different discovery mechanisms can accommodate such preferences in
   various ways.  Some user input or perhaps a pre-provisioned database
   seems inevitable.

Arkko & Aboba, Eds.      Expires July 10, 2004                 [Page 12]

Internet-Draft    Network Discovery and Selection Problem   January 2004

3. Design Constrains

   All solutions to the network selection and discovery problem must
   satisfy the following additional constraints:

   o  AAA routing mechanisms have to work for requests, responses, as
      well as server-initiated messages.

   o  Solution is compatible with all AAA protocols.

   o  Does not prevent the introduction of new AAA or access network
      features, such as link-state AAA routing protocols or fast

   o  Does not consume a significant amount of resources, such as
      bandwidth or increase network attachment time.

   o  Does not cause a problem with limited packet sizes of current

   o  Where new protocol mechanisms are required, it should be possible
      to deploy the solution without requiring changes to the largest
      base of installed devices -- network access servers, wireless
      access points, and clients.

   o  Similarly, new solutions should allow interoperability with
      clients, access networks, AAA proxies, and AAA servers that have
      not been modified to support network discovery and selection.

Arkko & Aboba, Eds.      Expires July 10, 2004                 [Page 13]

Internet-Draft    Network Discovery and Selection Problem   January 2004

4. Existing Work

4.1 IETF

   There has already been a lot of past work in this area, including a
   number of IETF Proposed Standards generated by the ROAMOPS WG.  The
   topic of roaming was considered different enough from both AAA and
   access protocols such as PPP that it deserved its own WG.

   In addition to work on ROAMOPS directly relating to the problem,
   there has been work in SEAMOBY relating to scaling of network
   discovery mechanisms; work in PKIX relating to identity and
   credential selection; and work in AAA WG relating to access routing.

   The PANA protocol [14] has a mechanism to advertise and select "ISPs"
   through the exchange of the ISP-Information AVP in its initial

   Adrangi et al [16] discuss the use of the EAP-Request/Identity for
   network discovery.  As noted in [10] Section 3.1, the minimum EAP MTU
   is 1020 octets, so that an EAP-Request/Identity is only guaranteed to
   be able to include 1015 octets within the Type-Data field.  Since RFC
   1035 [1] enables FQDNs to be up to 255 octets in length, this may not
   enable the announcement of very many networks, although if SSIDs are
   used, the maximum length of 32 octets per SSID may provide
   substantially better scaling.  The use of other network identifiers
   than domain names is also possible, for instance the PANA protocol
   uses an a free form string and an SMI Network Management Private
   Enterprise Code [14], or Mobile Network Codes could be used as hinted
   in [16].

   As noted in [30], the use of the EAP-Request/Identity for network
   discovery has substantial negative impact on handoff latency, since
   this may result in a station needing to initiate an EAP conversation
   with each Access Point in order to receive an EAP-Request/Identity
   describing which networks are supported.  Since IEEE 802.11-1999 does
   not support use of Class 1 data frames in State 1 (unauthenticated,
   unassociated) within an ESS, this implies either that the APs must
   support 802.1X pre-authentication (optional in IEEE 802.11i) or that
   the station must associate with each AP prior to sending an
   EAPOL-Start to initiate EAP.  This will dramatically increase handoff

   The effects to handoff latency depend also on the specific protocol
   design, and the expected likelihood of having to provide
   advertisements and initiate scanning of several access points.  The
   use of advertisements only as a last resort when the AAA routing has
   failed is a better approach than the use of advertisement - scanning

Arkko & Aboba, Eds.      Expires July 10, 2004                 [Page 14]

Internet-Draft    Network Discovery and Selection Problem   January 2004

   procedure on every attachment.

   Furthermore, if the AP has not been updated to present an up to date
   set of networks in the EAP-Requests/Identity, after associating to
   candidate APs and then choosing one, it is possible that the station
   will find that the chosen network is not supported after all.  In
   this case, the station's EAP-Response/Identity may be answered with
   an updated EAP-Request/Identity, adding yet more latency.

4.2 IEEE

   There has been work in IEEE 802.11 and 802.1 relating to network
   discovery enhancements.

   Some recent contributions in this space include the following:

   o  [18] defines the Beacon and Probe Response mechanisms used with
      IEEE 802.11.  Unfortunately, Beacons are only sent at the lowest
      supported rate.  Studies such as [31] have identified MAC layer
      performance problems, and [28] have identified scaling issues
      resulting from a lowering of the Beacon interval.

   o  [21] discusses the evolution of authentication models in WLANs,
      and the need for the network to migrate from existing models to
      new ones, based on either EAP layer indications or through the use
      of SSIDs to represent more than the local network.  It notes the
      potential need for management or structuring of the SSID space.

      The paper also notes that virtual APs have scalability issues.  It
      does not analyze these scalability issues in relation to those
      existing in other alternative solutions, however.

   o  [22] discusses requirements for differentiation in the way that
      the user's payload traffic is routed, based on home network
      control.  Such requirements have come up, for instance, in the
      context of 3GPP.

   o  [19] discusses mechanisms currently used to provide "Virtual AP"
      capabilities within a single physical access point.  A "Virtual
      AP" appears at the MAC and IP layers to be distinct physical AP.
      As noted in the paper, full compatibility with existing 802.11
      station implementations can only be maintained if each virtual AP
      uses a distinct MAC address (BSSID) for use in Beacons and Probe
      Responses.  This draft does not discuss scaling issues in detail,
      but recommends that only a limited number of virtual APs be
      supported by a single physical access point.  The simulations
      presented in [28] appear to confirm this conclusion; with a Beacon
      interval of 100 ms, once more than 8 virtual APs are supported on

Arkko & Aboba, Eds.      Expires July 10, 2004                 [Page 15]

Internet-Draft    Network Discovery and Selection Problem   January 2004

      a single channel, more than 20% of bandwidth is used for Beacons
      alone.  This would indicate a limit of approximately 20 virtual
      APs per physical AP.

4.3 3GPP

   The 3GPP technical specification [23] covers the interworking of WLAN
   networks with 2G and 3G networks.  This specification discusses also
   network discovery and selection issues.

   The specification requires that Access Network Discovery is performed
   as specified in the standards for the relevant WLAN link layer
   technology.  An early version of the technical specification required
   the use of a 3GPP-specific SSID, but that has since then been
   abandoned; access network or local 3GPP network based SSIDs may be
   used instead.  It has not been decided whether some conventions on
   the format of these SSIDs is required by 3GPP.

   In addition to Access Network Discovery, it is necessary to select
   intermediary networks for the purposes of AAA Routing.  In 3GPP,
   these networks are called Visited Public Land-based Mobile Networks
   (VPLMNs), and it is assumed that WLAN networks may have a contract
   with more than one VPLMN.  GSM/UMTS roaming mechanisms are then
   employed for routing AAA requests from the VPLMN to the home network.

   In order to select the VPLMN, the following is required:

   o  User can choose the desired VPLMN.

   o  AAA message are routed according to the NAI.

   o  Existing EAP mechanisms are used where possible.

   o  Extensibility is desired, to allow the advertisement of other
      parameters later.

   The referenced 3GPP technical specification is a so called stage 2
   specification, and contains only the principles of operation, leaving
   detailed protocol work for later.  Nevertheless, the specification
   states that advertisement information shall be provided only when the
   access network is unable to route the request using normal AAA
   routing means, such as when it sees an unknown NAI domain.  It is
   also stated that where VPLMN control is required, the necessary
   information is added to a NAI.

   The security properties related to different mediating network
   selection mechanisms have been discussed in the 3GPP contribution

Arkko & Aboba, Eds.      Expires July 10, 2004                 [Page 16]

Internet-Draft    Network Discovery and Selection Problem   January 2004

   [24], which concludes that both SSID- and EAP-based mechanisms have
   roughly similar (and very limited) security properties, and that, as
   a result, network advertisement should be considered only as hints.

   Ahmavaara, Haverinen, and Pichna [26] discuss the new network
   selection requirements that 3G-WLAN roaming introduces.  It is
   necessary to support automatic network selection, and not just manual
   selection by the user.  There may be multiple levels of networks, the
   hotspot owner may have a contract with a provider who in turn has a
   contract with one 3G network, and this 3G network has a roaming
   capability with a number of other networks.

4.4 Other

   [27] discusses the need for network selection in a situation where
   there is more than one available access network with a roaming
   agreement to the home network.  It also lists EAP-level, SSID-based,
   and PEAP-based mechanisms as potential solutions to the network
   selection problem.

   Eijk et al [25] discussed the general issue of network selection.
   They concentrated primarily on the Access Network Discovery problem,
   based on various criteria, and did not consider the other aspects of
   the network selection problem.  Nevertheless, they mention that one
   of the network selection problems is that the information about
   accessibility and roaming relationships is not stored in one
   location, but rather spread around the network.

Arkko & Aboba, Eds.      Expires July 10, 2004                 [Page 17]

Internet-Draft    Network Discovery and Selection Problem   January 2004

5. Conclusions

   The issues surrounding the network discovery and selection problem
   have been summarized.

   In the opinion of the editors of this document, the main findings

   o  There is a clear need for access network discovery, identifier
      selection, AAA routing, and payload routing.

   o  Existing mechanisms appear sufficient for the control of payload
      routing, but there appears to be justification for enhanced
      mechanisms relating to access network discovery, identifier
      selection, and AAA routing.

   o  Nevertheless, many of the problems discussed in this draft are
      very hard when one considers them in an environment that requires
      a potentially large number of networks, fast handoffs, and
      automatic decisions.

   o  The proliferation of multiple competing network discovery
      technologies within IEEE 802, IETF, and 3GPP appears to a
      significant problem going forward.  In the absence of a clearly
      defined solution to the problem it is likely that any or all of
      these solutions will be utilized, resulting in industry
      fragmentation and lack of interoperability.

      In order to avoid this fate, it is strongly suggested that a
      discussion be initiated between IETF and IEEE 802 in order to work
      out the roles of the each organization in solving this problem,
      and to invite 3GPP participation so that their requirements can be
      fulfilled by the planned solutions.

   o  New link layers should be designed with facilities that enable the
      efficient distribution of network advertisement information.

   o  Solving all problems with current link layers and existing network
      access devices may not be possible.  It may be useful to consider
      a phased approach where only certain functions are provided now,
      and the full functionality is provided when extensions to current
      link layers become available.

   We will briefly comment on the specific mechanisms related to network
   discovery and selection:

   o  As noted in studies such as [31] and [28], the IEEE 802.11 Beacon/
      Probe Response mechanism has substantial scaling issues, and as a

Arkko & Aboba, Eds.      Expires July 10, 2004                 [Page 18]

Internet-Draft    Network Discovery and Selection Problem   January 2004

      result a single physical access point is in practice limited to
      less than a dozen virtual APs on each channel with IEEE 802.11b.

      The situation is improved substantially with successors such as
      IEEE 802.11a which enable additional channels, thus potentially
      increasing the number of potential virtual APs.

      However, even these enhancements it is not feasible to advertise
      more than 50 different networks using existing mechanisms, and
      probably significantly less in most circumstances.

      As a result, there appears to be justification for enhancing the
      scalability of network advertisements.

   o  Work is already underway in IEEE 802.1 to provide enhanced
      discovery functionality.  For example, IEEE 802.1ab enables
      network devices to announce themselves and provide information on
      their capabilities.  Similarly, the IEEE 802.1af has discussed the
      idea of supporting network discovery within a future revision to
      IEEE 802.1X.  However, neither IEEE 801.ab nor IEEE 802.1af is
      likely to address the transport of large quantities of data where
      fragmentation would be a problem.

   o  Given that EAP does not support fragmentation of EAP-Request/
      Identity packets, and that use of EAP for network selection on all
      attachments will have a very substantial adverse impact on roaming
      performance without appropriate lower layer support (such as
      support for Class 1 data frames within IEEE 802.11), the use of
      EAP is at best limited.  Long-term, it makes more sense for the
      desired functionality to be handled either within IEEE 802 or at
      the IP layer.

   o  In the IETF, a potential alternative is use of the SEAMOBY CARD
      protocol [13], which enables advertisement of network device
      capabilities over IP.  Another alternative is the recently
      proposed Device Discovery Protocol (DDP) [12], which provides
      functionality equivalent to IEEE 802.1ab using ASN.1 encoded
      advertisements sent to a link-local scope multicast address.

      A limitation of these IP layer solutions is that they can only
      work as a means to speed up the attachment procedures when moving
      from one location to another; when a node starts up, it needs to
      be able to attach to a network before IP communications are
      available.  This is fine for optimizations, but precludes the use
      in a case where the discovery information is mandatory before
      successful attachment can be accomplished, for instance when the
      access network is unable to route the AAA request unaided.

Arkko & Aboba, Eds.      Expires July 10, 2004                 [Page 19]

Internet-Draft    Network Discovery and Selection Problem   January 2004

   o  "Phone-book" based approaches such as RFC 3017 appear attractive
      due to their ability to provide sufficient information for
      automatic selection decisions.  However, there is no experience on
      applying such approaches to wireless access.  The number of WLAN
      access points is significantly higher than the number of dial-in
      POPs; the distributed nature of the access network has created a
      more complicated business and roaming structure, and the expected
      rate of change in the information is high.

   Finally, to address some of the security concerns that have come up
   during this work, the IETF should in any case initiate work that
   enables support for channel bindings in methods.  Preferably, popular
   methods should be updated, ensuring compatibility with existing
   deployments.  The representation of link layer parameters within EAP
   should utilize a common framework, to make it easier to define new
   link layers and keep the selection of EAP methods independent of the
   link layer.

Arkko & Aboba, Eds.      Expires July 10, 2004                 [Page 20]

Internet-Draft    Network Discovery and Selection Problem   January 2004

6. Security Considerations

   All aspects of the network discovery and selection problem are
   security related.  The security issues and requirements have been
   discussed in the previous sections.

   The security requirements for network discovery depend on the type of
   information being discovered.  Some of the parameters may have a
   security impact, such as the claimed name of the network the user
   tries to attach to.  Unfortunately, current EAP methods do not always
   make the verification of such parameters possible.

   The security requirements for network selection depend on whether the
   selection is considered as a command or a hint.  For instance, the
   selection that the user provided may be ignored if it relates to AAA
   routing and the access network can route the AAA traffic to the
   correct home network using other means in any case.

Arkko & Aboba, Eds.      Expires July 10, 2004                 [Page 21]

Internet-Draft    Network Discovery and Selection Problem   January 2004

Normative References

   [1]   Mockapetris, P., "Domain names - implementation and
         specification", STD 13, RFC 1035, November 1987.

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

   [3]   Aboba, B., Lu, J., Alsop, J., Ding, J. and W. Wang, "Review of
         Roaming Implementations", RFC 2194, September 1997.

   [4]   Aboba, B. and M. Beadles, "The Network Access Identifier", RFC
         2486, January 1999.

   [5]   Aboba, B. and J. Vollbrecht, "Proxy Chaining and Policy
         Implementation in Roaming", RFC 2607, June 1999.

   [6]   Rigney, C., Willens, S., Rubens, A. and W. Simpson, "Remote
         Authentication Dial In User Service (RADIUS)", RFC 2865, June

   [7]   Zorn, G., Leifer, D., Rubens, A., Shriver, J., Holdrege, M. and
         I. Goyret, "RADIUS Attributes for Tunnel Protocol Support", RFC
         2868, June 2000.

   [8]   Riegel, M. and G. Zorn, "XML DTD for Roaming Access Phone
         Book", RFC 3017, December 2000.

   [9]   Calhoun, P., Loughney, J., Guttman, E., Zorn, G. and J. Arkko,
         "Diameter Base Protocol", RFC 3588, September 2003.

   [10]  Blunk, L., "Extensible Authentication Protocol (EAP)",
         draft-ietf-eap-rfc2284bis-07 (work in progress), December 2003.

   [11]  Housley, R. and T. Moore, "Certificate Extensions and
         Attributes Supporting Authentication in PPP and  Wireless LAN",
         draft-ietf-pkix-wlan-extns-04 (work in progress), December

Arkko & Aboba, Eds.      Expires July 10, 2004                 [Page 22]

Internet-Draft    Network Discovery and Selection Problem   January 2004

Informative References

   [12]  Enns, R., Marques, P. and D. Morrell, "Device Discovery
         Protocol (DDP)", draft-marques-ddp-00 (work in progress), May

   [13]  Liebsch, M., "Candidate Access Router Discovery",
         draft-ietf-seamoby-card-protocol-05 (work in progress),
         November 2003.

   [14]  Forsberg, D., "Protocol for Carrying Authentication for Network
         Access (PANA)", draft-ietf-pana-pana-02 (work in progress),
         October 2003.

   [15]  Kaufman, C., "Internet Key Exchange (IKEv2) Protocol",
         draft-ietf-ipsec-ikev2-07 (work in progress), April 2003.

   [16]  Adrangi, F., "Network Discovery and Selection within the EAP
         draft-adrangi-eap-network-discovery-and-selection-00 (work in
         progress), October 2003.

   [17]  Institute of Electrical and Electronics Engineers, "Local and
         Metropolitan Area Networks: Port-Based Network Access Control",
         IEEE Standard 802.1X, September 2001.

   [18]  Institute of Electrical and Electronics Engineers, "Wireless
         LAN Medium Access Control (MAC) and Physical Layer (PHY)
         Specifications", IEEE Standard 802.11, 1999.

   [19]  Aboba, B., "Virtual Access Points", IEEE Contribution
         11-03-154r1, May 2003.

   [20]  Mishra, A., "Improving the latnecy of the Probe Phase during
         802.11 Handoff", IEEE Contribution 11-03-417r2, May 2003.

   [21]  Hepworth, E., "Co-existence of Different Authentication
         Models", IEEE Contribution 11-03-0827 2003.

   [22]  Hong, C. and T. Yew, "Interworking - WLAN Control", IEEE
         Contribution 11-03-0843 2003.

   [23]  3GPP, "3GPP System to Wireless Local Area Network (WLAN)
         interworking; System Description; Release 6", 3GPP Draft
         Technical Specification 23.234 v 2.2.0, December 2003.

   [24]  Ericsson, "Security of EAP and SSID based network
         advertisements", 3GPP Contribution S3-030736, November 2003.

Arkko & Aboba, Eds.      Expires July 10, 2004                 [Page 23]

Internet-Draft    Network Discovery and Selection Problem   January 2004

   [25]  Eijk, R., Brok, J., Bemmel, J. and B. Busropan, "Access Network
         Selection in a 4G Environment and the Role of Terminal and
         Service Platform", 10th WWRF, New York, October 2003.

   [26]  Ahmavaara, K., Haverinen, H. and R. Pichna, "Interworking
         Architecture between WLAN and 3G Systems", IEEE Communications
         Magazine, November 2003.

   [27]  Intel, "Wireless LAN (WLAN) End to End Guidelines for
         Enterprises and Public Hotspot Service Providers", November

   [28]  Velayos, H. and G. Karlsson, "Techniques to Reduce IEEE 802.11b
         MAC Layer Handover Time", Laboratory for Communication
         Networks, KTH, Royal Institute of Technology, Stockholm,
         Sweden, TRITA-IMIT-LCN R 03:02, April 2003.

   [29]  Judd, G. and P. Steenkiste, "Fixing 802.11 Access Point
         Selection", Sigcomm Poster Session 2002.

   [30]  Eronen, P., "Network Selection Issues", presentation to EAP WG
         at IETF 58, November 2003.

   [31]  Heusse, M., "Performance Anomaly of 802.11b", LSR-IMAG
         Laboratory, Grenoble, France, IEEE Infocom 2003.

Authors' Addresses

   Jari Arkko

   Jorvas  02420

   EMail: jari.arkko@ericsson.com

   Bernard Aboba
   One Microsoft Way
   Redmond, WA  98052

   EMail: aboba@internaut.com

Arkko & Aboba, Eds.      Expires July 10, 2004                 [Page 24]

Internet-Draft    Network Discovery and Selection Problem   January 2004

Appendix A. Contributors

   This draft is based on the discussion held on the EAP WG mailing list
   in December 2003, and on a number of input documents such as [16].
   The editors of this document would like to especially acknowledge the
   contributions of Farid Adrangi, Farooq Bari, Michael Richardson, Pasi
   Eronen, Mark Watson, Mark Grayson, Johan Rune, and Tomas

Arkko & Aboba, Eds.      Expires July 10, 2004                 [Page 25]

Internet-Draft    Network Discovery and Selection Problem   January 2004

Intellectual Property Statement

   The IETF takes no position regarding the validity or scope of any
   intellectual property or other rights that might be claimed to
   pertain to the implementation or use of the technology described in
   this document or the extent to which any license under such rights
   might or might not be available; neither does it represent that it
   has made any effort to identify any such rights.  Information on the
   IETF's procedures with respect to rights in standards-track and
   standards-related documentation can be found in BCP-11.  Copies of
   claims of rights made available for publication and any assurances of
   licenses to be made available, or the result of an attempt made to
   obtain a general license or permission for the use of such
   proprietary rights by implementors or users of this specification can
   be obtained from the IETF Secretariat.

   The IETF invites any interested party to bring to its attention any
   copyrights, patents or patent applications, or other proprietary
   rights which may cover technology that may be required to practice
   this standard.  Please address the information to the IETF Executive

Full Copyright Statement

   Copyright (C) The Internet Society (2004).  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

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

   This document and the information contained herein is provided on an

Arkko & Aboba, Eds.      Expires July 10, 2004                 [Page 26]

Internet-Draft    Network Discovery and Selection Problem   January 2004



   Funding for the RFC Editor function is currently provided by the
   Internet Society.

Arkko & Aboba, Eds.      Expires July 10, 2004                 [Page 27]