idnits 2.17.1 draft-ietf-mif-current-practices-09.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (March 28, 2011) is 4768 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Outdated reference: A later version (-15) exists of draft-ietf-mif-problem-statement-11 -- Obsolete informational reference (is this intentional?): RFC 3484 (Obsoleted by RFC 6724) -- Duplicate reference: RFC3484, mentioned in 'RFC3484', was also mentioned in 'ANDROID-RFC3484'. -- Obsolete informational reference (is this intentional?): RFC 3484 (Obsoleted by RFC 6724) -- Duplicate reference: RFC3484, mentioned in 'WNDS-RFC3484', was also mentioned in 'RFC3484'. -- Obsolete informational reference (is this intentional?): RFC 3484 (Obsoleted by RFC 6724) Summary: 0 errors (**), 0 flaws (~~), 2 warnings (==), 6 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Internet Engineering Task Force M. Wasserman 3 Internet-Draft Painless Security, LLC 4 Intended status: Informational P. Seite 5 Expires: September 29, 2011 France Telecom - Orange 6 March 28, 2011 8 Current Practices for Multiple Interface Hosts 9 draft-ietf-mif-current-practices-09 11 Abstract 13 An increasing number of hosts are operating in multiple-interface 14 environments, where different network interfaces are providing 15 unequal levels of service or connectivity. This document summarizes 16 current practices in this area, and describes in detail how some 17 common operating systems cope with these challenges. 19 Status of this Memo 21 This Internet-Draft is submitted in full conformance with the 22 provisions of BCP 78 and BCP 79. 24 Internet-Drafts are working documents of the Internet Engineering 25 Task Force (IETF). Note that other groups may also distribute 26 working documents as Internet-Drafts. The list of current Internet- 27 Drafts is at http://datatracker.ietf.org/drafts/current/. 29 Internet-Drafts are draft documents valid for a maximum of six months 30 and may be updated, replaced, or obsoleted by other documents at any 31 time. It is inappropriate to use Internet-Drafts as reference 32 material or to cite them other than as "work in progress." 34 This Internet-Draft will expire on September 29, 2011. 36 Copyright Notice 38 Copyright (c) 2011 IETF Trust and the persons identified as the 39 document authors. All rights reserved. 41 This document is subject to BCP 78 and the IETF Trust's Legal 42 Provisions Relating to IETF Documents 43 (http://trustee.ietf.org/license-info) in effect on the date of 44 publication of this document. Please review these documents 45 carefully, as they describe your rights and restrictions with respect 46 to this document. Code Components extracted from this document must 47 include Simplified BSD License text as described in Section 4.e of 48 the Trust Legal Provisions and are provided without warranty as 49 described in the Simplified BSD License. 51 Table of Contents 53 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 54 2. Summary of Current Approaches . . . . . . . . . . . . . . . . 3 55 2.1. Centralized Connection Management . . . . . . . . . . . . 3 56 2.2. Per Application Connection Settings . . . . . . . . . . . 4 57 2.3. Stack-Level Solutions to Specific Problems . . . . . . . . 4 58 2.3.1. DNS Resolution Issues . . . . . . . . . . . . . . . . 5 59 2.3.2. Routing . . . . . . . . . . . . . . . . . . . . . . . 5 60 2.3.3. Address Selection Policy . . . . . . . . . . . . . . . 5 61 3. Current Practices in Some Operating Systems . . . . . . . . . 6 62 3.1. Mobile Handset Operating Systems . . . . . . . . . . . . . 6 63 3.1.1. Nokia S60 3rd Edition, Feature Pack 2 . . . . . . . . 7 64 3.1.2. Microsoft Windows Mobile and Windows Phone 7 . . . . . 9 65 3.1.3. RIM BlackBerry . . . . . . . . . . . . . . . . . . . . 10 66 3.1.4. Google Android . . . . . . . . . . . . . . . . . . . . 11 67 3.1.5. Qualcomm Brew . . . . . . . . . . . . . . . . . . . . 12 68 3.1.6. Leadcore Tech. Arena . . . . . . . . . . . . . . . . . 13 69 3.2. Desktop Operating Systems . . . . . . . . . . . . . . . . 13 70 3.2.1. Microsoft Windows . . . . . . . . . . . . . . . . . . 14 71 3.2.1.1. Routing . . . . . . . . . . . . . . . . . . . . . 14 72 3.2.1.2. Outbound and Inbound Addresses . . . . . . . . . . 14 73 3.2.1.3. DNS Configuration . . . . . . . . . . . . . . . . 14 74 3.2.2. Linux and BSD-based Operating Systems . . . . . . . . 15 75 3.2.2.1. Routing . . . . . . . . . . . . . . . . . . . . . 16 76 3.2.2.2. Outbound and Inbound Addresses . . . . . . . . . . 16 77 3.2.2.3. DNS Configuration . . . . . . . . . . . . . . . . 17 78 3.3. Focus on access network selection . . . . . . . . . . . . 18 79 4. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 19 80 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 20 81 6. Security Considerations . . . . . . . . . . . . . . . . . . . 20 82 7. Contributors . . . . . . . . . . . . . . . . . . . . . . . . . 20 83 8. References . . . . . . . . . . . . . . . . . . . . . . . . . . 21 84 8.1. Normative References . . . . . . . . . . . . . . . . . . . 21 85 8.2. Informative References . . . . . . . . . . . . . . . . . . 21 86 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 23 88 1. Introduction 90 Multiple-interface hosts face several challenges not faced by single- 91 interface hosts, some of which are described in the MIF problem 92 statement, [I-D.ietf-mif-problem-statement]. This document 93 summarizes how current implementations deal with the problems 94 identified in the MIF problem statement. 96 Publicly-available information about the multiple-interface solutions 97 implemented in some widely used operating systems, including both 98 mobile handset and desktop operating systems, is collected in this 99 document, including: Nokia S60 [S60], Microsoft Windows Mobile 100 [WINDOWSMOBILE], Blackberry [BLACKBERRY], Google Android [ANDROID], 101 Microsoft Windows, Apple Mac OS X, Linux and BSD-based operating 102 systems. 104 2. Summary of Current Approaches 106 This section summarizes current approaches that are used to resolve 107 the multi-interface issues described in the Multiple Interface 108 Problem Statement [I-D.ietf-mif-problem-statement]. These approaches 109 can be broken down into three major categories: 111 o Centralized connection management 113 o Per-application connection settings 115 o Stack-level solutions to specific problems 117 2.1. Centralized Connection Management 119 It is a common practice for mobile handset operating systems to use a 120 centralized connection manager that performs network interface 121 selection based on application or user input. The information used 122 by the connection manager may be programmed into an application or 123 provisioned on a handset-wide basis. When information is not 124 available to make an interface selection, the connection manager will 125 query the user to choose between available choices. 127 Routing tables are not typically used for network interface selection 128 when a connection manager is in use, as the criteria for network 129 selection is not strictly IP-based but is also dependent on other 130 properties of the interface (cost, type, etc.). Furthermore, 131 multiple overlapping private IPv4 address spaces are often exposed to 132 a multiple-interface host, making it difficult to make interface 133 selection decisions based on prefix matching. 135 2.2. Per Application Connection Settings 137 In mobile handsets, applications are often involved in choosing what 138 interface and related configuration information should be used. In 139 some cases, the application selects the interface directly, and in 140 other cases the application provides more abstract information to a 141 connection manager that makes the final interface choice. 143 2.3. Stack-Level Solutions to Specific Problems 145 In most desktop operating systems, multiple interface problems are 146 dealt with in the stack and related components, based on system- 147 level configuration information, without the benefit of input from 148 applications or users. These solutions tend to map well to the 149 problems listed in the problem statement: 151 o DNS resolution issues 153 o Routing 155 o Address selection policy 157 The configuration information for desktop systems comes from one of 158 three sources: DHCP, proprietary configuration systems or manual 159 configuration. While these systems universally accept IP address 160 assignment on a per-interface basis, they differ in what set of 161 information can be assigned on a per-interface basis and what can be 162 configured only on a per-system basis. 164 When choosing between multiple sets of information provided, these 165 systems will typically give preference to information received on the 166 "primary" interface. The mechanism for designating the "primary" 167 interface differs by system. 169 There is very little commonality in how desktop operating systems 170 handle multiple sets of configuration information, with notable 171 variations between different versions of the same operating system 172 and/or within different software packages built for the same 173 operating system. Although these systems differ widely, it is not 174 clear that any of them provide a completely satisfactory user 175 experience in multiple-interface environments. 177 The following sections discuss some of the solutions used in each of 178 the areas raised in the MIF problem statement. 180 2.3.1. DNS Resolution Issues 182 There is very little commonality in how desktop operating systems 183 handle the DNS server list. Some systems support per-interface DNS 184 server lists, while others only support a single system-wide list. 186 On hosts with per-interface DNS server lists, different mechanisms 187 are used to determine which DNS server is contacted for a given 188 query. In most cases, the first DNS server listed on the "primary" 189 interface is queried first, with back off to other servers if an 190 answer is not received. 192 Systems that support a single system-wide list differ in how they 193 select which DNS server to use in cases where they receive more than 194 one DNS server list to configure (e.g. from DHCP on multiple 195 interfaces). Some accept the information received on the "primary" 196 interface, while others use either the first or last set DNS server 197 list configured. 199 2.3.2. Routing 201 Routing information is also handled differently on different desktop 202 operating systems. While all systems maintain some sort of routing 203 cache, to handle redirects and/or statically configured routes, most 204 packets are routed based on configured default gateway information. 206 Some systems do allow the configuration of different default router 207 lists for different interfaces. These systems will always choose the 208 default gateway on the interface with the lowest routing metric, with 209 different behavior when two or more interfaces have the same routing 210 metric. 212 Most systems do not allow the configuration of more than one default 213 router list, choosing instead to use the first or last default router 214 list configured and/or the router list configured on the "primary" 215 interface. 217 2.3.3. Address Selection Policy 219 There is somewhat more commonality in how desktop hosts handle 220 address selection. Applications typically provide the destination 221 address for an outgoing packet, and the IP stack is responsible for 222 picking the source address. 224 IPv6 specifies a specific source address selection mechanism in 225 [RFC3484], and several systems implement this mechanism with similar 226 support for IPv4. However, many systems do not provide any mechanism 227 to update this default policy, and there is no standard way to do so. 229 In some cases, the routing decision (including which interface to 230 use) is made before source address selection is performed, and a 231 source address is chosen from the outbound interface. In other 232 cases, source address selection is performed before, or independently 233 from outbound interface selection. 235 3. Current Practices in Some Operating Systems 237 The following sections briefly describe the current multiple- 238 interface host implementations on some widely-used operating systems. 239 Please refer to the References section for pointers to original 240 documentation on most of these systems, including further details. 242 3.1. Mobile Handset Operating Systems 244 Cellular devices typically run a variety of applications in parallel, 245 each with different requirements for IP connectivity. A typical 246 scenario is shown in figure 1, where a cellular device is utilizing 247 WLAN access for web browsing and GPRS access for transferring 248 multimedia messages (MMS). Another typical scenario would be a real- 249 time VoIP session over one network interface in parallel with best 250 effort web browsing on another network interface. Yet another 251 typical scenario would be global Internet access through one network 252 interface and local (e.g. corporate VPN) network access through 253 another. 255 Web server MMS Gateway 256 | | 257 -+--Internet---- ----Operator network--+- 258 | | 259 +-------+ +-------+ 260 |WLAN AP| | GGSN | 261 +-------+ +-------+ 262 | +--------+ | 263 +--------|Cellular|--------+ 264 |device | 265 +--------+ 267 A cellular device with two network interfaces 269 Figure 1 271 Different network access technologies require different settings. 272 For example, WLAN requires Service Set Identifier (SSID) and the GPRS 273 network requires the Access Point Name (APN) of the Gateway GPRS 274 Support Node (GGSN), among other parameters. It is common that 275 different accesses lead to different destination networks (e.g. to 276 "Internet", "intranet", cellular network services, etc.). 278 3.1.1. Nokia S60 3rd Edition, Feature Pack 2 280 S60 is a software platform for mobile devices running on the Symbian 281 OS. S60 uses the concept of an Internet Access Point (IAP) [S60] 282 that contains all information required for opening a network 283 connection using a specific access technology. A device may have 284 several IAPs configured for different network technologies and 285 settings (multiple WLAN SSIDs, GPRS APNs, dial-up numbers, and so 286 forth). There may also be 'virtual' IAPs that define parameters 287 needed for tunnel establishment (e.g. for VPN). 289 For each application, a correct IAP needs to be selected at the point 290 when the application requires network connectivity. This is 291 essential, as the wrong IAP may not be able to support the 292 application or reach the desired destination. For example, MMS 293 application must use the correct IAP in order to reach the MMS 294 Gateway, which typically is not accessible from the public Internet. 295 As another example, an application might need to use the IAP 296 associated with its corporate VPN in order to reach internal 297 corporate servers. Binding applications to IAPs avoids several 298 problems, such as choosing the correct DNS server in the presence of 299 split DNS (as an application will use the DNS server list from its 300 bound IAP), and overlapping private IPv4 address spaces used for 301 different interfaces (as each application will use the default routes 302 from its bound IAP). 304 If multiple applications utilize the same IAP, the underlying network 305 connection can typically be shared. This is often the case when 306 multiple Internet-using applications are running in parallel. 308 The IAP for an application can be selected in multiple ways: 310 o Statically: e.g. from a configuration interface, via client 311 provisioning/device management system, or at build-time. 313 o Manually by the user: e.g. each time an application starts the 314 user may be asked to select the IAP to use. This may be needed, 315 for example, if a user sometimes wishes to access his corporate 316 intranet and other times would prefer to access the Internet 317 directly. 319 o Automatically by the system: after the destination network has 320 been selected statically or dynamically. 322 The static approach is fine for certain applications, like MMS, for 323 which configuration can be provisioned by the network operator and 324 does not change often. Manual selection works, but may be seen as 325 troublesome by the user. An automatic selection mechanism needs to 326 have some way of knowing which destination network the user, or an 327 application, is trying access. 329 S60 3rd Edition, Feature Pack 2, introduces a concept of Service 330 Network Access Points (SNAPs) that group together IAPs that lead to 331 the same destination. This enables static or manual selection of the 332 destination network for an application and leaves the problem of 333 selecting the best of the available IAPs within a SNAP to the 334 operating system. 336 When SNAPs are used, it is possibly for the operating system to 337 notify applications when a preferred IAP, leading to the same 338 destination, becomes available (for example, when a user comes within 339 range of his home WLAN access point), or when the currently used IAP 340 is no longer available and applications have to reconnect via another 341 IAP (for example, when a user goes out of range of his home WLAN and 342 must move to the cellular network). 344 In S60 3.2 does not support RFC 3484 for source address selection 345 mechanisms. Applications are tightly bound the network interface 346 selected for them or by them. E.g. an application may be connected 347 to IPv6 3G connection, IPv4 3G connection, WLAN connection, or VPN 348 connection. The application can change between the connections, but 349 uses only one at a time. If the interface happens to be dual-stack, 350 then IPv4 is preferred over IPv6. 352 DNS configuration is per-interface; an application bound to an 353 interface will always use the DNS settings for that interface. Hence 354 the device itself remembers these pieces of information for each 355 interface separately. 357 The S60 3.2 manages with totally overlapping addresses spaces. Each 358 interface can even have same IPv4 address configured on it without 359 issues. This is so because interfaces are kept totally separate from 360 each other. This also implies that the interface selection has to be 361 done at application layer, as from network layer point of view device 362 is not multihomed in the IP-sense. 364 Please see the source documentation for more details and screenshots: 365 [S60]. 367 3.1.2. Microsoft Windows Mobile and Windows Phone 7 369 Microsoft Windows Mobile leverages on a Connection Manager 370 [WINDOWSMOBILE] to handle multiple network connections. This 371 architecture centralizes and automates network connection 372 establishment and management, and makes it possible to automatically 373 select a connection, to dial-in automatically or by user initiation, 374 and to optimize connection and shared resource usage. Connection 375 Manager periodically re-evaluates the validity of the connection 376 selection. The Connection Manager uses various attributes such as 377 cost, security, bandwidth, error rate, and latency in its decision 378 making. 380 The Connection Manager selects the best possible connection for the 381 application based on the destination network the application wishes 382 to reach. The selection is made between available physical and 383 virtual connections (e.g. VPN, GPRS, WLAN, and wired Ethernet) that 384 are known to provide connectivity to the destination network, and the 385 selection is based on the costs associated with each connection. 386 Different applications are bundled to use the same network connection 387 when possible, but in conflict situations when a connection cannot be 388 shared, higher priority applications take precedence, and the lower 389 priority applications lose connectivity until the conflict situation 390 clears. 392 During operation, Connection Manager opens new connections as needed, 393 and also disconnects unused or idle connections. 395 To optimize resource use, such as battery power and bandwidth, 396 Connection Manager enables applications to synchronize network 397 connection usage by allowing applications to register their 398 requirements for periodic connectivity. An application is notified 399 when a suitable connection becomes available for its use. 401 In comparison to Windows Mobile connection management, Windows phone 402 7 updates the routing functionality in the case where the terminal 403 can be attached simultaneously to several interfaces. Windows Phone 404 7 routes the traffic through a preferred interface, which has a lower 405 metric. When there are multiple interfaces, the applications system 406 will, by default, choose from an ordered list of available 407 interfaces. The default connection policy will prefer wired over 408 wireless and WLAN over cellular. Hence, if an application wants to 409 use cellular 3G as the active interface when WLAN is available, the 410 application needs to override the default connection mapping policy. 411 An application specific mapping policy can be set via a microsoft API 412 or provisioned by the Mobile Operator. The application, in 413 compliance with the security model, can request connection type by 414 interface (WLAN, cellular), by minimum interface speed (x kbps, y 415 mbps), or by name (Access Point Name). 417 In dual-stack systems, Windows mobile and Windows phone 7 implement 418 adress selection rules as per [WNDS-RFC3484]. An administrator can 419 configure a policy table that can override the default behavior of 420 the selection algorithms. It is reminded that the policy table 421 specifies precedence values and preferred source prefixes for 422 destination prefixes (see [RFC3484], section 2.1, for details). If 423 the system has not been configured, then the default policy table 424 specified in [RFC3484] is used. 426 3.1.3. RIM BlackBerry 428 Depending on the network configuration, applications in reasearch In 429 Motion (RIM) BlackBerry devices [BLACKBERRY] can use can use direct 430 TCP/IP connectivity or different application proxys to establish 431 connections over the wireless network. For instance, some wireless 432 service providers provide an Internet gateway to offer direct TCP/IP 433 connectivity to the Internet while some others can provide a WAP 434 gateway that allows HTTP connections to occur over the WAP (Wireless 435 Application Protocol) protocol. It is also possible to use the 436 BlackBerry Enterprise Server [BLACKBERRY] as a network gateway, The 437 BlackBerry Enterprise Server provides an HTTP and TCP/IP proxy 438 service to allow the application to use it as a secure gateway for 439 managing HTTP and TCP/IP connections to the intranet or the Internet. 440 An application connecting to the Internet, can use either the 441 BlackBerry Internet Service or the Internet gateway of the wireless 442 server provider or direct Internet connectivity over WLAN to manage 443 connections. The problem of gateway selection is supposed to be 444 managed independently by each application. For instance, an 445 application can be designed to always use the default Internet 446 gateway, while another application can be designed to use a preferred 447 proxy when available. 449 A BlackBerry device [BLACKBERRY] can be attached to multiple networks 450 simultaneously (wireless/wired). In this case, Multiple network 451 interfaces can be associated to a single IP stack or multiple IP 452 stacks. The device, or the application, can select the network 453 interface to be used in various ways. For instance, the device can 454 always map the applications to the default network interface (or the 455 default access network). When muliple IP stacks are associated to 456 multiple interfaces, the application can select the source address 457 correponding to the preferred network interface. Per-interface IP 458 stacks also allow to manage overlapping addresses spaces. When 459 multiple network interfaces are aggregated into a single IP stack, 460 the device associates each application to the more appropriate 461 network interface. The selection can be based on cost, type-of- 462 service and/or user preference. 464 The BlackBerry uses per-interface DNS configuration; applications 465 bound to a specific interface will use the DNS settings for that 466 interface. 468 3.1.4. Google Android 470 Android is based on a Linux kernel and, in many situations, behaves 471 like a Linux device as described in Section 3.2.2. As per Linux, 472 Android can manage multiple routing tables and rely on policy based 473 routing associated with packet filtering capabilities (see 474 Section 3.2.2.1 for details). Such a framework can be used to solve 475 complex routing issue brought by multiple interfaces terminals, e.g. 476 address space overlapping. 478 For incoming packets, Android implements the weak host model 479 [RFC1122] on both IPv4 and IPv6. However, Android can also be 480 configured to support the strong host model. 482 Regarding DNS configuration, Android does not list the DNS servers in 483 the file /etc/resolv.conf, used by Linux. However, as per Linux, DNS 484 configuration is node-scoped, even if DNS configuration can rely on 485 the DHCP client. For instance, the udhcp client [UDHCP], which is 486 also available for Linux, can be used on Android. Each time new 487 configuration data is received by the host from a DHCP server, 488 regardless of which interface it is received on, the DHCP client 489 rewrites the global configuration data with the most recent 490 information received. 492 Actually, the main difference between Linux and Android is on the 493 address selection mechanism. Android version prior to 2.2 simply 494 prefers IPv6 connectivity over IPv4. However, it should be noted 495 that, at the time of writing, IPv6 is available only on WiFi and 496 virtual interfaces, but not on the cellular interface (without IPv6 497 in IPv4 encapsulation). Android 2.2 has been updated with 498 [ANDROID-RFC3484], which implements some of the address selection 499 rules defined in [RFC3484]. All RFC3484 rules are supported, except 500 rule 3 (avoid deprecated addresses), 4 (prefer home addresses) and 7 501 (prefer native transport). Also, rule 9 (use longest matching 502 prefix) has been modified so it does not sort IPv4 addresses. 504 The Android reference documentation describes the android.net package 505 [ANDROID] and the ConnectivityManager class that applications can use 506 to request a route to a specified destination address via a specified 507 network interface (3GPP or WLAN). Applications also ask Connection 508 Manager for permission to start using a network feature. The 509 Connectivity Manager monitors changes in network connectivity and 510 attempts to failover to another network if connectivity to an active 511 network is lost. When there are changes in network connectivity, 512 applications are notified. Applications are also able to ask for 513 information about all network interfaces, including their 514 availability, type and other information. 516 3.1.5. Qualcomm Brew 518 This section describes how multi-interface support is handled by 519 Advanced Mobile Station Software (AMSS) that comes with Brew OS for 520 all Qualcomm chipsets (e.g., MSM, Snapdragon etc). AMSS is a low 521 level connectivity platform, on top of which manufacturers can build 522 to provide the necessary connectivity to applications. The 523 interaction model between AMSS, the Operating System, and the 524 applications is not unique and depend on the design chosen by the 525 manufacturer. The Mobile OS can let an application invoke the AMSS 526 directly (via API), or provide its own connection manager that will 527 request connectivity to the AMSS based on applications needs. The 528 interaction between the OS connection manager and the applications is 529 OS dependent. 531 AMSS supports a concept of netpolicy which allows each application to 532 specify the type of network connectivity desired. The netpolicy 533 contains parameters such as access technology, IP version type and 534 network profile. Access technology could be a specific technology 535 type such as CDMA or WLAN or could be a group of technologies, such 536 as ANY_Cellular or ANY_Wireless. IP version could be one of IPv4, 537 IPv6 or Default. The network profile identifies a type of network 538 domain or service within a certain network technology, such as 3GPP 539 APN or Mobile IP Home Agent. It also specifies all the mandatory 540 parameters required to connect to the domain such authentication 541 credentials and other optional parameters such as QoS attributes. 542 Network Profile is technology specific and set of parameters 543 contained in the profile could vary for different technologies. 545 Two models of network usage are supported: 547 o Applications requiring network connectivity specify an appropriate 548 netpolicy in order to select the desired network. The netpolicy 549 may match one or more network interfaces. AMSS system selection 550 module selects the best interface out of the ones that match the 551 netpolicy based on various criteria such as cost, speed or other 552 provisioned rules. Application explicitly starts the selected 553 network interface and, as a result, the application also gets 554 bound to the corresponding network interface. All outbound 555 packets from this application are always routed over this bound 556 interface using the source address of the interface. 558 o Applications may rely on a separate connection manager to control 559 (e.g. start/stop) the network interface. In this model, 560 applications are not necessarily bound to any one interface. All 561 outbound packets from such applications are routed on one of the 562 interfaces that match its netpolicy. The routing decision is made 563 individually for each packet and selects the best interface based 564 on the criteria described above and the destination address. 565 Source address is always that assigned to the interface used to 566 transmit the packet. 568 All of the routing/interface selection decisions are based on the 569 netpolicy and not just on the destination address to avoid 570 overlapping private IPv4 address issue. This also allows multiple 571 interfaces to be configured with the same IP address, for example, to 572 handle certain tunnelling scenarios. Applications that do not 573 specify a netpolicy are routed by AMSS to the best possible interface 574 using the default netpolicy. Default netpolicy could be pre-defined 575 or provisioned by the administrator or operator. Hence default 576 interface could vary from device to device and also depends upon the 577 available networks at any given time. 579 AMSS allows each interface to be configured with its own set of DNS 580 configuration parameters (e.g. list of DNS servers, domain names 581 etc.). Interface selected to make a DNS resolution is the one to 582 which application making the DNS query is bound. Applications can 583 also specify a different netpolicy as part of DNS request to select 584 another interface for DNS resolution. Regardless, all the DNS 585 queries are sent only over this selected interface using the DNS 586 configuration from the interface. DNS resolution is first attempted 587 with the primary server configured in the interface. If a response 588 is not received, the queries are sent to all the other servers 589 configured in the interface in a sequential manner using a backoff 590 mechanism. 592 3.1.6. Leadcore Tech. Arena 594 Arena, a mobile OS based on Linux, provides a Connection Manager, 595 which is described in [I-D.zhang-mif-connection-manager-arena] and 596 [I-D.yang-mif-connection-manager-impl-req]. The arena connection 597 manager provides a means for applications to register their 598 connectivity requirement. The Connection Manager can then choose an 599 interface that matches the application's needs while considering 600 other factors such as availability, cost and stability. Also, the 601 Connection Manager can handle multiple-interfaces issues such as 602 connection sharing. 604 3.2. Desktop Operating Systems 606 Multi-interface issues also occur in desktop environments in those 607 cases where a desktop host has multiple (logical or physical) 608 interfaces connected to networks with different reachability 609 properties, such as one interface connected to the global Internet, 610 while another interface is connected to a corporate VPN. 612 3.2.1. Microsoft Windows 614 The multi-interface functionality currently implemented in Microsoft 615 Windows operation systems is described in more detail in 616 [I-D.montenegro-mif-multihoming]. 618 3.2.1.1. Routing 620 It is possible, although not often desirable, to configure default 621 routers on more than one Windows interface. In this configuration, 622 Windows will use the default route on the interface with the lowest 623 routing metric (i.e. the fastest interface). If multiple interfaces 624 share the same metric, the behavior will differ based on the version 625 of Windows in use. Prior to Windows Vista, the packet would be 626 routed out of the first interface that was bound to the TCP/IP stack, 627 the preferred interface. In Windows vista, host-to-router load 628 sharing [RFC4311] is used for both IPv4 and IPv6. 630 3.2.1.2. Outbound and Inbound Addresses 632 If the source address of the outgoing packet has not been determined 633 by the application, Windows will choose from the addresses assigned 634 to its interfaces. Windows implements [RFC3484] for source address 635 selection in IPv6 and, in Windows Vista, for IPv4. Prior to Windows 636 Vista, IPv4 simply chose the first address on the outgoing interface. 638 For incoming packets, Windows will check if the destination address 639 matches one of the addresses assigned to its interfaces. Windows has 640 implemented the weak host model [RFC1122] on IPv4 in Windows 2000, 641 Windows XP and Windows Server 2003. The strong host model became the 642 default for IPv4 in Windows Vista and Windows server 2008, however 643 the weak host model is available via per-interface configuration. 644 IPv6 has always implemented the strong host model. 646 3.2.1.3. DNS Configuration 648 Windows largely relies on suffixes to solve DNS resolution issues. 649 Suffixes are used for four different purposes that are reminded 650 hereafter: 652 1. DNS Suffix Search List (aka domain search list): suffix is added 653 to non-FQDN names. 655 2. Interface-specific suffix list, which allows sending different 656 DNS queries to different DNS servers. 658 3. Suffix to control Dynamic DNS Updates: determine which DNS server 659 will receive a dynamic update for a name with a certain suffix. 661 4. Suffix in the Name Resolution Policy Table [NRPT] to aid in 662 identifying a Namespace that requires special handling (feature 663 available only after Windows 7 and its server counterpart, 664 Windows Server 2008 R2). 666 However, this section focuses on the interface-specific suffix list 667 since it is the only suffix usage in the scope of this document. 669 DNS configuration information can be host-wide or interface specific. 670 Host-wide DNS configuration is input via static configuration or, in 671 sites that use Active Directory, Microsoft's Group Policy. Interface 672 specific DNS configuration can be input via static configuration or 673 via DHCP. 675 The host-wide configuration consists of a primary DNS suffix to be 676 used for the local host, as well as a list of suffix that can be 677 appended to names being queried. Before Windows Vista and Windows 678 Server 2008, there was also a host-wide DNS server list that took 679 precedent over per-interface DNS configuration. 681 The interface-specific DNS configuration comprises an interface- 682 specific suffix list and a list of DNS server IP addresses. 684 Windows uses a host-wide "effective" server list for an actual query, 685 where the effective server list may be different for different names. 686 In the list of DNS server addresses, the first server is considered 687 the "primary" server, with all other servers being secondary. 689 When a DNS query is performed in Windows, the query is first sent to 690 the primary DNS server on the preferred interface. If no response is 691 received in one second, the query is sent to the primary DNS servers 692 on all interfaces under consideration. If no response is received 693 for 2 more seconds, the DNS server sends the query to all of the DNS 694 servers on the DNS server lists for all interfaces under 695 consideration. If the host still doesn't receive a response after 4 696 seconds, it will send to all of the servers again and wait 8 seconds 697 for a response. 699 3.2.2. Linux and BSD-based Operating Systems 700 3.2.2.1. Routing 702 In addition to the two commonly used routing tables (the local and 703 main routing tables), the kernel can support up to 252 additional 704 routing tables which can be added in the file /etc/iproute2/ 705 rt_tables. A routing table can contain an arbitrary number of 706 routes, the selection of route is classically made according to the 707 destination address of the packet. Linux also provides more flexible 708 routing selection based on the Type of Service, scope, output 709 interface. In addition, since kernel version 2.2, Linux supports 710 policy based routing using the multiple routing tables capability and 711 a routing policy database. This database contains routing rules used 712 by the kernel. Using policy based routing, the source address, the 713 ToS flags, the interface name and an "fwmark" (a mark carried through 714 added in the data structure representing the packet) can be used as 715 route selectors. 717 Policy based routing can be used in addition to Linux packet 718 filtering capabilities, e.g provided by the "iptables" tool. In a 719 multiple interfaces context, this tool can be used to mark the 720 packets, i.e assign a number to fwmark, in order to select the 721 routing rule according to the type of traffic. This mark can be 722 assigned according to parameters like protocol, source and/or 723 destination addresses, port number and so on. 725 Such a routing management framework allows to deal with complex 726 situation such as address space overlapping. In this situation, the 727 administrator can use packet marking and policy based routing to 728 select the correct interface. 730 3.2.2.2. Outbound and Inbound Addresses 732 By default, source address selection follows the following basics 733 rules: the initial source address for an outbound packet can be 734 chosen by the application using the bind() call. Without information 735 from the application, the kernel chooses the first address configured 736 on the interface which belongs to the same subnet than the 737 destination address or the nexthop router. 739 Linux also implements [RFC3484] for source address selection for IPv6 740 and dual-stack configurations. However, the address sorting rules 741 from [RFC3484] are not always adequate. For this reason, Linux 742 allows the system administrator to dynamically change the sorting. 743 This can be achieved with the /etc/gai.conf file. 745 For incoming packets, Linux checks if the destination address matches 746 one of the addresses assigned to its interfaces then, processes the 747 packet according the configured host model. By default, Linux 748 implements the weak host model [RFC1122] on both IPv4 and IPv6. 749 However, Linux can also be configured to support the strong host 750 model. 752 3.2.2.3. DNS Configuration 754 Most BSD and Linux distributions rely on their DHCP client to handle 755 the configuration of interface-specific information (such as an IP 756 address and netmask), and a set of system-wide configuration 757 information, (such a DNS server list, an NTP server list and default 758 routes). Users of these operating systems have the choice of using 759 any DHCP client available for their platform, with an operating 760 system default. This section discusses the behavior of several DHCP 761 clients that may be used with Linux and BSD distributions. 763 The Internet Systems Consortium (ISC) DHCP Client [ISCDHCP] and its 764 derivative for OpenBSD [OPENBSDDHCLIENT] can be configured with 765 specific instructions for each interface. However, each time new 766 configuration data is received by the host from a DHCP server, 767 regardless of which interface it is received on, the DHCP client 768 rewrites the global configuration data, such as the default routes 769 and the DNS server list (in /etc/resolv.conf) with the most recent 770 information received. Therefore, the last configured interface 771 always become the primary one. The ISC DHCPv6 client behaves 772 similarly. 774 The Phystech dhcpcd client [PHYSTECHDHCPC] behaves similarly to the 775 ISC client. It replaces the DNS server list in /etc/resolv.conf and 776 the default routes each time new DHCP information is received on any 777 interface. However, the -R flag can be used to instruct the client 778 to not replace the DNS servers in /etc/resolv.conf. However, this 779 flag is a global flag for the DHCP server, and is therefore 780 applicable to all interfaces. When dhcpd is called with the -R flag, 781 the DNS servers are never replaced. 783 The pump client [PUMP] also behaves similarly to the ISC client. It 784 replaces the DNS servers in /etc/resolv.conf and the default routes 785 each time new DHCP information is received on any interface. 786 However, the nodns and nogateway options can be specified on a per 787 interface basis, enabling the user to define which interface should 788 be used to obtain the global configuration information. 790 The udhcp client [UDHCP] is often used in embedded platforms based on 791 busybox. The udhcp client behaves similarly to the ISC client. It 792 rewrites default routes and the DNS server list each time new DHCP 793 information is received. 795 Redhat-based distributions, such as Redhat, Centos and Fedora have a 796 per-interface configuration option (PEERDNS) that indicates that the 797 DNS server list should not be updated based on configuration received 798 on that interface. 800 The most configurable DHCP clients can be set to define a primary 801 interface to use only that interface for the global configuration 802 data. However, this is limited, since a mobile host might not always 803 have the same set of interfaces available. Connection managers may 804 help in this situation. 806 Some distributions also have a connection manager. However, most 807 connection managers serve as a GUI to the DHCP client, therefore not 808 changing the functionality described above. 810 3.3. Focus on access network selection 812 This section describes behaviors of connection managers in presence 813 of multiple points of attachment for a same interface. In order to 814 illustrate different practices, a set of representative handsets 815 considered: LG Pathfinder, Android/HTC magic, RIM BlackBerry and 816 iPhone (3G and 3GS). The section focuses on WLAN access technology, 817 it is described how does the connection manager deal with the list of 818 preferred SSID and how does it select the access point for 819 attachment. Desktops are not covered since many different connection 820 managers can be easily installed, thus making hard to report a common 821 behaviour. This section only focuses on a specific use-case and 822 current implementations; however further considerations on network 823 discovery and selection can be found in [RFC5113]. [RFC5113] 824 describes the network discovery and selection and discusses 825 limitations and constraints on potential solutions. 827 When the terminal is under coverage of different WLAN networks with 828 different SSIDs: 830 Connection managers, excepted for the RIM Blackberry, construct 831 the list of preferred SSID giving priority to the last SSID on 832 which they have managed to attach. Excepted for the RIM 833 blackberry, the user is not allowed to define its preferred 834 access. So, if the terminal discovers and manages to attach to 835 SSID1, SSID1 becomes the preferred access for future attachment. 836 If the terminal moves out of SSID1 coverage and attaches to a new 837 SSID, SSID2. SSID2 will then be the preferred access of the 838 connection manager. Then, if the terminal processes to WLAN 839 attachment within both SSID1 and SSID2 coverage, the connection 840 manager will select SSID2 for attachment. The RIM Blackberry 841 behaves differently: the user is allowed to define its list of 842 preferred accesses. The connection manager selects the first SSID 843 of the preferred list of SSIDs configured in the device and that 844 is available based on the WLAN scan the device has performed. 846 All connection managers behave in the same way when the terminals 847 fails to attach to the selected SSID: the connection manager 848 automatically selects the next SSID in the list of preferred SSID. 849 Fallback come into play at expiration of a timeout from few 850 seconds to about 3 minutes. 852 When the IP stack fails to obtain an IP address after successful 853 WLAN attachment, all the considered handsets, excepted the iPhone, 854 try to connect another SSID. LG Pathfinder and HTC just select 855 the next SSID in the list. The BlackBerry, performs a new WLAN 856 scan and, among the networks available, selects the next SSID in 857 the preferred list. 859 When the terminal receives signals from different point of attachment 860 with same SSID: 862 The connection manager selects the point of attachment with best 863 signal strength; no other criteria (e.g. MAC address) is taken 864 into account. If the handset fails to attach to the selected 865 point of attachment (e.g. due to L2 authentication failure), the 866 connection manager selects the point of attachment with lower 867 signal strength. However, this fallback is not supported on the 868 LG pathfinder. If no more points of attachment (corresponding to 869 the preferred SSID) are available, the connection manager selects 870 the second SSID in the list of preferred SSID. 872 Whatever is the handset, fallback on L3 attachment failure is not 873 supported if the terminal remains under coverage of the same WLAN 874 access point. Actually, the connection manager always selects the 875 most powerful signal strength without considering IP configuration 876 results. In other words, if the terminal is unable to set up the 877 IP connectivity on one WLAN access, the connection manager will 878 not try to attach to an alternative point of attachment (or SSID) 879 as long as the signal strength of the first radio link is the most 880 powerful. Situation is not the same for mobile terminal since the 881 signal strength of the alternative point of attachment could 882 become better while the terminal is moving. If so, the terminal 883 automatically restarts IP connectivity process (excepted the HTC 884 Magic which requires the user to manually restart the L3 885 attachment). 887 4. Acknowledgements 889 Authors of the document would like to thank following people for 890 their input and feedback: Dan Wing, Hui Deng, Jari Arkko, Julien 891 Laganier and Steinar H. Gunderson. 893 5. IANA Considerations 895 This memo includes no request to IANA. 897 6. Security Considerations 899 This document describes current operating system implementations and 900 how they handle the issues raised in the MIF problem statement. 901 While it is possible that the currently implemented mechanisms 902 described in this document may affect the security of the systems 903 described, this document merely reports on current practice. It does 904 not attempt to analyze the security properties (or any other 905 architectural properties) of the currently implemented mechanisms. 907 7. Contributors 909 The following people contributed most of the per-Operating System 910 information found in this document: 912 o Marc Blanchet, Viagenie 914 o Hua Chen, Leadcoretech, Ltd. 916 o Yan Zhang, Leadcoretech Ltd. 918 o Shunan Fan, Huawei Technology 920 o Jian Yang, Huawei Technology 922 o Gabriel Montenegro, Microsoft Corporation 924 o Shyam Seshadri, Microsoft Corporation 926 o Dave Thaler, Microsoft Corporation 928 o Kevin Chin, Microsoft Corporation 930 o Teemu Savolainen, Nokia 932 o Tao Sun, China Mobile 934 o George Tsirtsis, Qualcomm. 936 o David Freyermuth, France telecom. 938 o Aurelien Collet, Altran. 940 o Giyeong Son, RIM. 942 8. References 944 8.1. Normative References 946 [I-D.ietf-mif-problem-statement] 947 Blanchet, M. and P. Seite, "Multiple Interfaces and 948 Provisioning Domains Problem Statement", 949 draft-ietf-mif-problem-statement-11 (work in progress), 950 March 2011. 952 8.2. Informative References 954 [ANDROID] Google Inc., "Android developers: package android.net", 955 2009, . 958 [ANDROID-RFC3484] 959 Gunderson, S., "RFC 3484 support for Android", 2010, . 963 [BLACKBERRY] 964 Research In Motion Limited, "BlackBerry Java Development 965 Environment - Fundamentals Guide: Wireless gateways", 966 2009, . 969 [I-D.montenegro-mif-multihoming] 970 Montenegro, G., Thaler, D., and S. Seshadri, "Multiple 971 Interfaces on Windows", 972 draft-montenegro-mif-multihoming-00 (work in progress), 973 March 2009. 975 [I-D.yang-mif-connection-manager-impl-req] 976 Yang, J., Sun, T., and S. Fan, "Multi-interface Connection 977 Manager Implementation and Requirements", 978 draft-yang-mif-connection-manager-impl-req-00 (work in 979 progress), March 2009. 981 [I-D.zhang-mif-connection-manager-arena] 982 Zhang, Y., Sun, T., and H. Chen, "Multi-interface Network 983 Connection Manager in Arena Platform", 984 draft-zhang-mif-connection-manager-arena-00 (work in 985 progress), February 2009. 987 [ISCDHCP] Internet Software Consortium, "ISC DHCP", 2009, 988 . 990 [NRPT] Windows, "Name Resolution Policy Table", February 2010, < 991 http://technet.microsoft.com/en-us/magazine/ 992 ff394369.aspx>. 994 [OPENBSDDHCLIENT] 995 OpenBSD, "OpenBSD dhclient", 2009, 996 . 998 [PHYSTECHDHCPC] 999 Phystech, "dhcpcd", 2009, 1000 . 1002 [PUMP] RedHat, "PUMP", 2009, . 1004 [RFC1122] Braden, R., "Requirements for Internet Hosts - 1005 Communication Layers", STD 3, RFC 1122, October 1989. 1007 [RFC3484] Draves, R., "Default Address Selection for Internet 1008 Protocol version 6 (IPv6)", RFC 3484, February 2003. 1010 [RFC4311] Hinden, R. and D. Thaler, "IPv6 Host-to-Router Load 1011 Sharing", RFC 4311, November 2005. 1013 [RFC5113] Arkko, J., Aboba, B., Korhonen, J., and F. Bari, "Network 1014 Discovery and Selection Problem", RFC 5113, January 2008. 1016 [S60] Nokia Corporation, "S60 Platform: IP Bearer Management", 1017 2007, . 1021 [UDHCP] Busybox, "uDHCP", 2009, . 1024 [WINDOWSMOBILE] 1025 Microsoft Corporation, "SDK Documentation for Windows 1026 Mobile-Based Smartphones: Connection Manager", 2005, 1027 . 1029 [WNDS-RFC3484] 1030 Microsoft Corporation, "SDK Documentation for Windows 1031 Mobile-Based Smartphones: Default Address Selection for 1032 IPv6", 2005, 1033 . 1035 Authors' Addresses 1037 Margaret Wasserman 1038 Painless Security, LLC 1039 356 Abbott Street 1040 North Andover, MA 01845 1041 USA 1043 Phone: +1 781 405-7464 1044 Email: mrw@painless-security.com 1045 URI: http://www.painless-security.com 1047 Pierrick Seite 1048 France Telecom - Orange 1049 4, rue du clos courtel BP 91226 1050 Cesson-Sevigne 35512 1051 France 1053 Email: pierrick.seite@orange-ftgroup.com