RFC 4923 - Quality of Service (QoS) Signaling in a Nested Virtua
Network Working Group F. Baker Request for Comments: 4923 Cisco Systems Category: Informational P. Bose Lockheed Martin August 2007 Quality of Service (QoS) Signaling in a Nested Virtual Private Network Status of This Memo This memo provides information for the Internet community. It does not specify an Internet standard of any kind. Distribution of this memo is unlimited. Copyright Notice Copyright (C) The IETF Trust (2007). Abstract Some networks require communication between an interior and exterior portion of a Virtual Private Network (VPN) or through a concatenation of such networks resulting in a nested VPN, but have sensitivities about what information is communicated across the boundary, especially while providing quality of service to communications with different precedence. This note seeks to outline the issues and the nature of the proposed solutions based on the framework for Integrated Services operation over Diffserv networks as described in RFC 2998. Table of Contents 1. Introduction ....................................................3 1.1. Problem Statement ..........................................3 1.2. Background Information and Terminology .....................4 1.3. Nested VPNs ................................................5 1.4. Signaled QoS Technology ....................................7 1.5. The Resource Reservation Protocol (RSVP) ...................9 1.6. Logical Structure of a VPN Router .........................10 2. Reservation and Preemption in a Nested VPN .....................13 2.1. Reservation in a Nested VPN ...............................14 2.2. Preemption in a Nested VPN ................................16 2.3. Working through an Example ................................17 2.3.1. Initial Routine Reservations - Generating Network State ......................................18 2.3.2. Initial Routine Reservations - Request Reservation ........................................19 2.3.3. Installation of a Reservation Using Precedence .....20 2.3.4. Installation of a Reservation Using Preemption .....21 3. Data Flows within a VPN Router .................................24 3.1. VPN Routers That Carry Data across the Cryptographic Boundary ....................................24 3.1.1. Plaintext to Ciphertext Data Flows .................24 3.1.2. Ciphertext to Plaintext Data Flows .................27 3.2. VPN Routers That Use the Network Guard for Signaling across the Cryptographic Boundary ...............28 3.2.1. Signaling Flow .....................................29 3.2.2. Use Case with Network Guard ........................30 4. Security Considerations ........................................33 5. Acknowledgements ...............................................34 6. References .....................................................34 6.1. Normative References ......................................34 6.2. Informative References ....................................35 1. Introduction 1.1. Problem Statement More and more networks wish to guarantee secure transmission of IP traffic across public LANs or WANs and therefore use Virtual Private Networks. Some networks require communication between an interior and exterior portion of a VPN or through a concatenation of such networks resulting in a nested VPN, but have sensitivities about what information is communicated across the boundary, especially while providing quality of service to communications with different precedence. This note seeks to outline the issues and the nature of the proposed solutions. The outline of the QoS solution for real- time traffic has been described at a high level in [RFC4542]. The key characteristics of this proposal are that o it uses standardized protocols, o it includes reservation setup and teardown for guaranteed and controlled load services using the standardized protocols, o it is independent of link delay, and therefore consistent with high delay*bandwidth networks as well as the more common variety, o it has no single point of failure, such as a central reservation manager, o it provides for the preemption of established data flows, o in that preemption, it not only permits a policy-admitted data flow in, but selects a specific data flow to exclude based upon control input rather than simply accepting a loss of service dictated at the discretion of the network control function, and o it interoperates directly with SIP Proxies, H.323 Gatekeepers, or other call management subsystems to present the other services required in a preemptive or preferential telephone network. The thrust of the memo surrounds VPNs that use encryption in some form, such as IPsec and their subsequent nesting across multiple network domains. This specific type of VPNs is further clarified in Section 1.3, which describes the nested VPN as an example of an IPsec or IPsec like VPN under the context of a 'customer provisioned' VPN. As a result, we will discuss the VPN router supporting "plaintext" and "ciphertext" interfaces. However, the concept extends readily to any form of aggregation, including the concept proposed in [RFC3175] of the IP traffic entering and leaving a network at identified points, and the use of other kinds of tunnels including Generic Routing Encapsulation (GRE), IP/IP, MPLS, and so on. 1.2. Background Information and Terminology A note on the use of the words "priority" and "precedence" in this document is in order. The term "priority" has been used in this context with a variety of meanings, resulting in a great deal of confusion. The term "priority" is used in this document to identify one of several possible datagram scheduling algorithms. A scheduler is used when deciding which datagram will be sent next on a computer interface; a priority scheduler always chooses a datagram from the highest priority class (queue) that is occupied, shielding one class of traffic from most of the jitter by passing jitter it would otherwise have experienced to another class. [RFC3181] applies the term to a reservation, in a sense that this document will refer to as "precedence". The term "precedence" is used in the sense implied in the phrase "Multi-Level Precedence and Preemption" [ITU.MLPP.1990]; some classes of sessions take precedence over others, which may result in bandwidth being admitted that might not otherwise have been or may result in the prejudicial termination of a lower-precedence session under a stated set of circumstances. For the purposes of the present discussion, "priority" is a set of algorithms applied to datagrams, where "precedence" is a policy attribute of sessions. The techniques of priority comparisons are used in a router or a policy decision point to implement precedence, but they are not the same thing. Along the same lines, it is important for the reader to understand the difference between QoS policies and policies based on the "precedence" or "importance" of data to the person or function using it. Voice, regardless of the precedence level of the call, is impeded by high levels of variation in network-induced delay. As a result, voice is often serviced using a priority queue, transferring jitter from that application's traffic to other applications. This is as true of voice for routine calls as it is for flash traffic. There are classes of application traffic that require bounded delay. That is a different concept than "no jitter"; they can accept jitter within stated bounds. Routing protocols such as OSPF or BGP are critical to the correct functioning of network infrastructure. While they are designed to work well with moderate loss levels, they are not helped by them, and even a short period of high loss can result in dramatic network events. Variation in delay, however, is not at all an issue if it is within reasonable bounds. As a result, it is common for routers to treat routing protocol datagrams in a way that limits the probability of loss, accepting relatively high delay in some cases, even though the traffic is absolutely critical to the network. Telephone call setup exchanges have this characteristic as well: faced with a choice between loss and delay, protocols like SIP and H.323 far prefer the latter, as the call setup time is far less than it would be if datagrams had to be retransmitted, and this is true regardless of whether the call is routine or of high precedence. As such, QoS markings tell us how to provide good service to an application independent of how "important" it may be at the current time, while "importance" can be conveyed separately in many cases. 1.3. Nested VPNs One could describe a nested VPN network in terms of three network diagrams. Figure 1 shows a simple network stretched across a VPN connection. The VPN router (where, following [RFC2460], a "router" is "a node that forwards packets not explicitly addressed to itself"), performs the following steps: o receives an IP datagram from a plaintext interface, o determines what remote enclave and therefore other VPN router to forward it to, o ensures that it has a tunnel mode security association (as generally defined in [RFC4301], Section 4) with that router, o encloses the encrypted datagram within another VPN (e.g., IPsec) and IP header, and o forwards the encapsulated datagram toward the remote VPN router. The receiving VPN router reverses the steps: o determines what security association the datagram was received from, o decrypts the interior datagram, o forwards the now-decapsulated datagram on a plaintext interface. The use of IPsec in this manner is described as the tunnel mode of [RFC4301] and [RFC4303]. Host Host Host Host Host Host /------------------/ /------------------/ Router -------Router | VPN-Router || || IPsec Tunnel through routed network || VPN-Router | Router -------Router /------------------/ /------------------/ Host Host Host Host Host Host Figure 1: VPN-Connected Enclave An important point to understand is that the VPN tunnel, like other features of the routed network, are invisible to the host. The host can infer that "something out there" is affecting the Path MTU, introducing delay, or otherwise affecting its data stream, but if properly implemented, it should be able to adapt to these. The words "if properly implemented" are the bane of every network manager, however; substandard implementations do demonstrably exist. Outside of the enclave, the hosts are essentially invisible. The communicating enclaves look like a simple data exchange between peer hosts across a routed network, as shown in Figure 2. Hosts Not Visible /==================/ Router | VPN-Router /---------------------/ Inner Domain /---------------------/ VPN-Router | Router /==================/ Hosts Not Visible Figure 2: VPN-Connected Enclave from Exterior Perspective Such networks can be nested and re-used in a complex manner. As shown in Figure 3, a pair of enclaves might communicate across a ciphertext network that, for various reasons, is itself re-encrypted and transmitted across a larger ciphertext network. The reasons for doing this vary, but they relate to information-hiding in the wider network, different levels of security required for different enclosed enclaves, and so on. Host Host Host Host Host Host /------------------/ /------------------/ Router -------Router | VPN-Router VPN-Router VPN-Router /---------------------/ /----------/ Router -------------Router | VPN-Router VPN-Router /-----------/ /----------/ Router -------Router | | Router -------Router /-----------/ /----------/ VPN-Router VPN-Router | Router ------------Router /---------------------/ /----------/ VPN-Router VPN-Router VPN-Router | Router -------Router /------------------/ /------------------/ Host Host Host Host Host Host Figure 3: Nested VPN The key question this document explores is "how do reservations, and preemption of reservations, work in such an environment?" 1.4. Signaled QoS Technology The Integrated Services model for networking was originally proposed in [RFC1633]. In short, it divides all applications into two broad classes: those that will adapt themselves to any available bandwidth, and those that will not or cannot. In the words of [RFC1633]: One class of applications needs the data in each packet by a certain time and, if the data has not arrived by then, the data is essentially worthless; we call these "real-time" applications. Another class of applications will always wait for data to arrive; we call these "elastic" applications. The Integrated Services model defines data flows supporting applications as either "real-time" or "elastic". It should be noted that "real-time" traffic is also referred to as "inelastic" traffic, and that elastic traffic is occasionally referred to as "non-real- time". In this view, the key issue is the so-called "playback point": a real-time application is considered to have a certain point in time at which data describing the next sound, picture, or whatever to be delivered to a display device or forfeit its utility, while an elastic application has no such boundary. Another way to look at the difference is that real-time applications have an irreducible lower bound on their bandwidth requirements. For example, the typical G.711 payload is delivered in 160-byte samples (plus 40 bytes of IP/ UDP/RTP headers) at 20 millisecond intervals. This will yield 80 kbps of bandwidth, without silence suppression, and not accounting for the layer 2 overhead. To operate in real-time, a G.711 codec requires the network over which its data will be delivered to support communications at 80 kbps at the IP layer with roughly constant end- to-end delay and nominal or no loss. If this is not possible (if there is significant loss or wide variations in delay), voice quality will suffer. On the other hand, if many megabits of capacity are available, the G.711 codec will not increase its bandwidth requirements either. Although adaptive codecs exist (e.g., G.722.2 or G.726), the adaptive mechanism can either require greater or lesser bandwidth and can adapt only within a certain range of bandwidth requirements beyond which the quality of the data flow required is not met. Elastic applications, however, will generally adapt themselves to any network: if the bottleneck provides 9600 bits per second, a Web transfer or electronic mail exchange will happen at 9600 bits per second, and if hundreds of megabits are available, the TCP (or SCTP) transport will increase their transfer rate in an attempt to reduce the time required to accomplish the transfer. For real-time applications, those that require data to be delivered end to end with at least a certain rate and with delays varying between stated bounds, the Integrated Services architecture proposes the use of a signaling protocol that allows the communicating applications and the network to communicate about the application requirements and the network's capability to deliver them. Several such protocols have been developed or are under development, notably including the Resource Reservation Protocol (RSVP) and Next Steps in Signaling (NSIS). The present discussion is limited to RSVP, although any protocol that delivers a similar set of capabilities could be considered. 1.5. The Resource Reservation Protocol (RSVP) RSVP is initially defined in [RFC2205] with a set of datagram processing rules defined in [RFC2209] and datagram details for Integrated Services [RFC2210]. Conceptually, this protocol specifies a way to identify data flows from a source application to a destination application and request specific resources for them. The source may be a single machine or a set of machines listed explicitly or implied, whereas the destination may be a single machine or a multicast group (and therefore all of the machines in it). Each application is specified by a transport protocol number in the IP protocol field, or may additionally include destination and perhaps source port numbers. The protocol is defined for both IPv4 [RFC0791] and IPv6 [RFC2460]. It was recognized immediately that it was also necessary to provide a means to perform the same function for various kinds of tunnels, which implies a relationship between what is inside and what is outside the tunnel. Definitions were therefore developed for IPsec [RFC2207] and [RFC4860] and for more generic forms of tunnels [RFC2746]. With the later development of the Differentiated Services Architecture [RFC2475], definitions were added to specify the Differentiated Services Code Point (DSCP) [RFC2474] to be used by a standard RSVP data flow in [RFC2996] and to use a pair of IP addresses and a DSCP as the identifying information for a data flow [RFC3175]. In addition, the initial definition of the protocol included a placeholder for policy information, and for preemption of reservations. This placeholder was later specified in detail in [RFC2750] with a view to associating a policy [RFC2872] with an identity [RFC3182] and thereby enabling the network to provide a contracted service to an authenticated and authorized user. This was integrated with the Session Initiation Protocol [RFC3261] in [RFC3312]. Preemption of a reservation is specified as in [RFC3181] -- a reservation that is installed in the network using an Preemption Priority and retained using a separate Defending Priority may be removed by the network via an RESV Error signal that removes the entire reservation. This has issues, however, in that the matter is often not quite so black and white. If the issue is that an existing reservation for 80 kbps can no longer be sustained but a 60 kbps reservation could, it is possible that a VoIP sender could change from a G.711 codec to a G.729 codec and achieve that. Or, if there are multiple sessions in a tunnel or other aggregate, one of the calls could be eliminated leaving capacity for the others. [RFC4495] seeks to address this issue. In a similar way, a capability was added to limit the possibility of control signals being spoofed or otherwise attacked [RFC2747] [RFC3097]. [RFC3175] describes several features that are unusual in RSVP, being specifically set up to handle aggregates in a service provider network. It describes three key components: o The RFC 3175 session object, which identifies not the IP addresses of the packets that are identified, but the IP addresses of the ingress and egress devices in the network, and the DSCP that the traffic will use. o The function of a reservation "aggregator", which operates in the ingress router and accepts individual reservations from the "customer" network. It aggregates the reservations into the ISP core in a tunnel or an MPLS LSP, or as a traffic stream that is known to leave at the deaggregator. o The function of a reservation "deaggregator", which operates in the egress router and breaks the aggregate reservation and data streams back out into individual data streams that may be passed to other networks. In retrospect, the Session Object specified by RFC 3175 is useful but not intrinsically necessary. If the ISP network uses tunnels, such as MPLS LSPs, IP/IP or GRE tunnels or enclosing IPsec Security Associations, the concepts of an aggregator and a deaggregator work in the same manner, although the reservation mechanism would be that of [RFC3473] and [RFC3474], [RFC2207], [RFC4860], or [RFC2746]. 1.6. Logical Structure of a VPN Router The conceptual structure of a VPN router is similar to that of any other router. In its simplest form, it is physically a two or more port device (similar to that shown in Figure 4), which has one or more interfaces to the protected enclave(s) and one or more interfaces to the outside world. On the latter, it structures some number of tunnels (in the case of an IPsec tunnel, having security associations) that it can treat as point-to-point interfaces from a routing perspective. +---------+ +-------+ +----+----+ +---------+ | RSVP | |Routing| |Net Guard| |IPsec Mgr| +----+----+ +---+---+ +----+----+ +----+----+ | | | | +----+-----------+------------+-----------------+----+ | IP | +-----------+--------------------+------------+------+ | | | | +-----+-----+ +----+------+ | | Encrypt/ | | Encrypt/ | | |Decrypt for| |Decrypt for| | | Security | | Security | | |Association| |Association| | +-----+-----+ +----+------+ | | | +-----------+------------+ +-----+------------+------+ | Plaintext | | Ciphertext | | Interface | | Interface | +------------------------+ +-------------------------+ Figure 4: Logical Structure of a VPN Router The encrypt/decrypt unit may be implemented as a function of the plaintext router, as a function on its interface card, or as a function of an external device with a private interface to the IP routing functionality of the plaintext router. These are conceptually equivalent, although there are practical differences in implementation. The key issue is that when IP routing presents a message to the encrypt/decrypt unit for transmission, it must also be presented with the IP address of the plaintext routing peer, whether host or router, to which the security association must be established. This IP Address is used to select (and perhaps create) the security association, and in turn select the appropriate set of security parameters. This could also be implemented by presenting the local Security Parameter Index (SPI) for the data, if it has been created out of band by the Network Management Process. In addition, it is necessary for aggregated signaling to be generated for the ciphertext domain. This may be accomplished in several ways: o by having the RSVP process on the plaintext router generate the messages and having the encrypt/decrypt unit bypass them into the ciphertext network o by having the plaintext RSVP process advise a process in the encrypt/decrypt implementation of what needs to be generated using some local exchange, and having it generate such messages, or o by having a separate parallel network management system intermediate between the plaintext and ciphertext routers, in which case, the encrypt/decrypt unit and the parallel network system must use the same address, and the ciphertext router must distinguish between traffic for them based on SPI or the presence of encryption. Control plane signaling using this additional path is described in Section 3.2. The information flow between the plaintext and ciphertext domains includes o IP datagrams via the encrypt/decrypt unit, o RSVP signaling via the bypass path, o Control information coordinating security associations, and o precious little else. 2. Reservation and Preemption in a Nested VPN / \ ( +--+ +--+ enclave ) ,---------. .----------. \ |H2+---+R2| / ,-' ` +--+ +--+`--.\ +--+ ++-+ / / +--+ +--+ |H1+---+R1| \`. | ,' / |R3+---+H3| +--+ +-++ ) '--. +----++ _.-' ( ++-+ +--+ | / _.`---|VPN2||''-. \ | enclave +----+-i.--'' +----++ `----.\ +----+ enclave --------|VPN1|' | ``|VPN3| , ,+----+ | +----+------' ,' --+-------+----------+------------------+---`. ,' ++-+ `. ,' |R7+--------+ `. / interface +--+ | \ domain 1 +-+--+ \ _.--------|VPN7|--------. ,-----'' +--+-+ `------. . `-. ,-' | `-. .-' `-: inner domain +-++ `.' ( |R9| ) .'. ++-+ ;-. .' `-. | ,-' `-. ' `------. +-+--+ _.-----' ` interface `---------|VPN8|-------'' domain 2 +-+--+ / \ | +--+ / `. +----------+R8| ,' `. ++-+ ,' `. --+------------------+-----------+------+-- ,' ,-----+----+ | +----+------. ,' |VPN6|. | _.|VPN4| ` +----+.`----. +----+ _.--'' ,+----+ | \ `--=.-|VPN5|---:' / | +--+ ++-+ : ,-'' +----+ `--. ; ++-+ +--+ |H6+---+R6| | ,' | `.| |R4+---+H4| +--+ +--+ ;/ +--+ ++-+ : +--+ +--+ // |H5+---+R5| \ enclave ,'( +--+ +--+ `. enclave `. ,' \ enclave / '-. , `-------' \ / `-------' Figure 5: Reservations in a Nested VPN Let us discuss how a resource reservation protocol, and specifically RSVP, might be used in a nested virtual private network. 2.1. Reservation in a Nested VPN A reservation in a nested VPN is very much like a reservation in any other network, with one exception: it is composed of multiple reservations that must be coordinated. These include a reservation within the originating and receiving enclaves and a reservation at each layer of the VPN, as shown in Figure 5. Thus, when a host in one enclave opens a reservation to a host in another enclave, a reservation of the appropriate type and size is created end to end. As it traverses the VPN router leaving its enclave, the reservation information and the data are placed within the appropriate tunnel (e.g., the IPsec Security Association for its precedence level to the appropriate remote VPN router). At the remote VPN router, it is extracted from the tunnel and passed on its way to the target system. The data in the enclave will be marked with a DSCP appropriate to its application and (if there is a difference) precedence level, and the signaling datagrams (PATH and RESV) are marked with a DCLASS object indicating that value. RSVP signaling datagrams (PATH and RESV) are marked with a DCLASS object indicating the value used for the corresponding data. The DSCP on the signaling datagrams, however, is a DSCP for signaling, and has the one provision that if routing varies by DSCP, then it must be a DSCP that is routed the same way as the relevant data. The [RFC2872] policy object specifies the applicable policy (e.g., "routine service for voice traffic") and asserts a [RFC3182] credential indicating its user (which may be a person or a class of persons). As specified in [RFC3181], it also specifies its Preemption Priority and its Defending Priority; these enable the Preemption Priority of a new session to be compared with the Defending Priority of previously admitted sessions. On the ciphertext side of the VPN router, no guarantees result unless the VPN router likewise sets up a reservation to the peer VPN Router across the ciphertext domain. Thus, the VPN router sets up an [RFC2207], [RFC4860], or [RFC3175] reservation to its peer. The Session Object defined by [RFC2207] or [RFC4860] contains a field called a "virtual destination port", which allows the multiplexing of many reservations over a common security association and, in the latter case, a common DSCP value. Thus, the voice traffic at every precedence level might use the Expedited Forwarding (EF) DSCP and service as described in [RFC3246], but the reservations would be for "the aggregate of voice sessions at precedence Pn between these VPN routers". This would allow the network administration to describe policies with multiple thresholds, such as "a new session at precedence Pn may be accepted if the total reserved bandwidth does not exceed threshold Qn; if it is necessary and sufficient to accept the reservation, existing reservations at lower precedences may be preemptively reduced to make acceptance of the new session possible". In the [RFC3175] case, since the DSCP must be used to identify both the reservation and the corresponding data stream, the aggregate reservations for different precedence levels require different DSCP values. In either case, the fundamental necessity is for one VPN router to act as what [RFC3175] calls the "aggregator" and another to act as the "deaggregator", and extend a VPN tunnel between them. If the VPN Tunnel is an IPsec Security Association between the VPN routers and the IP packet is entirely contained within (such as is used to cross a firewall), then the behavior of [RFC2746] is required of the tunnel. That bearer will have the following characteristics: o it will have a DSCP corollary to the DSCP for the data or the same DSCP as the data it carries, o the reservations and data will be carried in security associations between the VPN routers, and o the specification for the reservation for the tunnel itself will not be less than the sum of the requirements of the aggregated reservations. The following requirements relationships apply between the set of enclosed reservations and the tunnel reservation: o The sum of the average rates of the contained reservations, having been adjusted for the additional IP headers, will be less than or equal to the average rate of the tunnel reservation. o The sum of the peak rates of the contained reservations, having been adjusted for the additional IP headers, will be less than or equal to the peak rate of the tunnel reservation. o The sum of the burst sizes of the contained reservations, having been adjusted for the additional IP headers, will be less than or equal to the burst size of the tunnel reservation. o The Preemption Priority of a tunnel reservation is identical to that of the individual reservations it aggregates. o The Defending Priority of a tunnel reservation is identical to that of the individual reservations it aggregates. This would differ only in the case that measurement-based admission is in use in the tunnel but not in the end system. In that case, the tunnel's average bandwidth specification would be greater than or equal to the actual average offered traffic. Such systems are beyond the scope of this specification. As a policy matter, it may be useful to note a quirk in the way Internet QoS works. If the policies for various precedence levels specify different thresholds (e.g., "to accept a new routine call, the total reserved bandwidth after admission may not exceed X; to accept a call with a higher precedence level, the total reserved bandwidth after admission may not exceed X+Y, and this may be achieved by preempting a call with a lower precedence level"), the bandwidth Y effectively comes from the bandwidth in use by elastic traffic rather than forcing a preemption event. 2.2. Preemption in a Nested VPN As discussed in Section 1.5, preemption is specified in [RFC3181] and further addressed in [RFC4495]. The issue is that in many cases the need is to reduce the bandwidth of a reservation due to a change in the network, not simply to remove the reservation. In the case of an end-system-originated reservation, the end system might be able to accommodate the need through a change of codec; in the case of an aggregate of some kind, it could reduce the bandwidth it is sending by dropping one or more reservations entirely. In a nested VPN or other kind of aggregated reservation, this means that the deaggregator (the VPN router initiating the RESV signal for the tunnel) must o receive the RESV Error signaling it to reduce its bandwidth, o re-issue its RESV accordingly, o identify one or more of its aggregated reservations, enough to do the job, and o signal them to reduce their bandwidth accordingly. It is possible, of course, that it is signaling them to reduce their bandwidth to zero, which is functionally equivalent to removing the reservation as described in [RFC3181]. In the routers in the core, an additional case arises. One could imagine that some enclave presents the VPN with a single session, and that session has a higher precedence level. If some interior link is congested (e.g., the reserved bandwidth will exceed policy if the call is admitted), a session between a different pair of VPN routers must be preempted. More generally, in selecting a reservation to preempt, the core router must always select a reservation at the lowest available Defending Priority. This is the reason that various precedence levels must be kept separate in the core. 2.3. Working through an Example The network in Figure 5 shows three security layers: six plaintext enclaves around the periphery, two ciphertext domains connecting them at one layer (referred to in the diagram as an "interface domain"), and a third ciphertext domain connecting the first two (referred to in the diagram as an "inner domain"). The following distribution of information exists: o Each enclave has access to general routing information concerning other enclaves it is authorized to communicate with: systems in it can translate a DNS name for a remote host or domain and obtain the corresponding address or prefix. o Each enclave router also has specific routing information regarding its own enclave. o A default route is distributed within the enclave, pointing to its VPN router. o VPN Routers 1-6 are able to translate remote enclave prefixes to the appropriate remote enclave's VPN router addresses. o Each interface domain has access to general routing information concerning the other interface domains, but not the enclaves. Systems in an interface domain can translate a DNS name for a remote interface domain and obtain the corresponding address or prefix. o Each interface domain router also has specific routing information regarding its own interface domain. o A default route is distributed within the interface domain, pointing to the "inner" VPN router. o VPN Routers 7 and 8 are able to translate remote interface domain prefixes to remote VPN router addresses. o Routers in the inner domain have routing information for that domain only. While the example shows three levels, there is nothing magic about the number three. The model can be extended to any number of concentric layers. Note that this example places unidirectional reservations in the forward direction. In voice and video applications, one generally has a reservation in each direction. The reverse direction is not discussed, for the sake of clarity, but operates in the same way in the reverse direction and uses the same security associations. 2.3.1. Initial Routine Reservations - Generating Network State Now let us install a set of reservations from H1 to H4, H2 to H5, and H3 to H6, and for the sake of argument, let us presume that these are at the "routine" precedence. H1, H2, and H3 each initiate a PATH signal describing their traffic. For the sake of argument, let us presume that H1's reservation is for an [RFC2205] session, H2's reservation is for a session encrypted using IPsec, and therefore depends on [RFC2207], and H3 (which is a Public Switched Telephone Network (PSTN) gateway) sends an [RFC3175] reservation comprising a number of distinct sessions. Since these are going to H4, H5, and H6, respectively, the default route leads them to VPN1, VPN2, and VPN3, respectively. The VPN routers each ensure that they have an appropriate security association or tunnel open to the indicated remote VPN router (VPN4, VPN5, or VPN6). This will be a security association or tunnel for the indicated application at the indicated precedence level. Having accomplished that, it will place the PATH signal into the security association and forward it. If such does not already exist, following [RFC3175]'s aggregation model, it will now open a reservation (send a PATH signal) for the tunnel/SA within the interface domain; if the reservation does exist, the VPN router will increase the bandwidth indicated in the ADSPEC appropriately. In this example, these tunnel/SA reservations will follow the default route to VPN7. VPN7 ensures that it has an appropriate security association or tunnel open to VPN8. This will be a security association or tunnel for the indicated application at the indicated precedence level. Having accomplished that, it will place the PATH signal into the security association and forward it. If such does not already exist, following [RFC3175]'s aggregation model, it will now open a reservation (send a PATH signal) for the tunnel/SA within the interface domain; if the reservation does exist, the VPN router will increase the bandwidth indicated in the ADSPEC appropriately. In this example, this tunnel/SA reservation is forwarded to VPN8. VPN8 acts as an [RFC3175] deaggregator for the inner domain. This means that it receives the PATH signal for the inner domain reservation and stores state, decrypts the data stream from VPN7, operates on the RSVP signals as an RSVP-configured router, and forwards the received IP datagrams (including the updated PATH signals) into its interface domain. The PATH signals originated by VPN1, VPN2, and VPN3 are therefore forwarded towards VPN4, VPN5, and VPN6 according to the routing of the interface domain. VPN4, VPN5, and VPN6 each act as an [RFC3175] deaggregator for the interface domain. This means that it receives the PATH signal for the interface domain reservation and stores state, decrypts the data stream from its peer, operates on the RSVP signals as an RSVP- configured router, and forwards the received IP datagrams (including the updated PATH signals) into its enclave. The PATH signals originated by H1, H2, and H3 are therefore forwarded towards H4, H5, and H6 according to the routing of the enclave. H4, H5, and H6 now receive the original PATH signals and deliver them to their application. 2.3.2. Initial Routine Reservations - Request Reservation The application in H4, H5, and H6 decides to install the indicated reservations, meaning that they now reply with RESV signals. These signals request the bandwidth reservation. Following the trail left by the PATH signals, the RESV signals traipse back to their respective sources. The state left by the PATH signals leads them to VPN4, VPN5, and VPN6, respectively. If the routers in the enclaves are configured for RSVP, this will be explicitly via R4, R5, or R6; if they are not, routing will lead them through those routers. The various RSVP-configured routers en route in the enclave (including the VPN router on the "enclave" side) will verify that there is sufficient bandwidth on their links and that any other stated policy is also met. Having accomplished that, each will update its reservation state and forward the RESV signal to the next. The VPN routers will also each generate an RESV for the reservation within the interface domain, attempting to set or increase the bandwidth of the reservation appropriately. The various RSVP-configured routers en route in the interface domain (including VPN8) will verify that there is sufficient bandwidth on their links and that any other stated policy is also met. Having accomplished that, each will update its reservation state and forward the RESV signal to the next. VPN8 will also generate an RESV for the reservation within the inner domain, attempting to set or increase the bandwidth of the reservation appropriately. This gets the reservation to the inner deaggregator, VPN8. The various RSVP-configured routers en route in the inner domain (including VPN7) will verify that there is sufficient bandwidth on their links and that any other stated policy is also met. Having accomplished that, each will update its reservation state and forward the RESV signal to the next. This gets the signal to VPN7. VPN7 acts as an [RFC3175] aggregator for the inner domain. This means that it receives the RESV signal for the inner domain reservation and stores state, decrypts the data stream from VPN8, operates on the RSVP signals as an RSVP-configured router, and forwards the received IP datagrams (including the updated RESV signals) into its interface domain. The RESV signals originated by VPN4, VPN5, and VPN6 are therefore forwarded towards VPN1, VPN2, and VPN3 through the interface domain. VPN1, VPN2, and VPN3 each act as an [RFC3175] aggregator for the interface domain. This means that it receives the RESV signal for the interface domain reservation and stores state, decrypts the data stream from its peer, operates on the RSVP signals as an RSVP- configured router, and forwards the received IP datagrams (including the updated RESV signals) into its enclave. The RESV signals originated by H4, H5, and H6 are therefore forwarded towards H1, H2, and H3 according to the routing of the enclave. H1, H2, and H3 now receive the original RESV signals and deliver them to their application. 2.3.3. Installation of a Reservation Using Precedence Without going through the details called out in Sections 2.3.1 and 2.3.2, if sufficient bandwidth exists to support them, reservations of other precedence levels or other applications may also be installed across this network. If the "routine" reservations already described are for voice, for example, and sufficient bandwidth is available under the relevant policy, a reservation for voice at the "priority" precedence level might be installed. Due to the mechanics of preemption, however, this would not expand the existing "routine" reservations in the interface and inner domains, as doing this causes loss of information - how much of the reservation is now "routine" and how much is "priority"? Rather, this new reservation will open up a separate set of tunnels or security associations for traffic of its application class at its precedence between that aggregator and deaggregator. As a side note, there is an opportunity here that does not exist in the PSTN. In the PSTN, all circuits are potentially usable by any PSTN application under a certain set of rules (H channels, such as those used by video streams, must be contiguous and ordered). As such, if a channel is not made available to routine traffic but is made available to priority traffic, the operator is potentially losing revenue on the reserved bandwidth and deserves remuneration. However, in the IP Internet, some bandwidth must be kept for basic functions such as routing, and, in general, policies will not permit 100% of the bandwidth on an interface to be allocated to one application at one precedence. As a result, it may be acceptable to permit a certain portion (e.g., 50%) to be used by routine voice and a larger amount (e.g., 60%) to be used by voice at a higher precedence level. Under such a policy, a higher precedence reservation for voice might not result in the preemption of a routine call, but rather impact elastic traffic, and might be accepted at a time that a new reservation of lower precedence might be denied. In microwave networks, such as satellite or mobile ad hoc, one could also imagine network management intervention that could change the characteristics of the radio signal to increase the bandwidth under some appropriate policy. 2.3.4. Installation of a Reservation Using Preemption So we now have a number of reservations across the network described in Figure 5 including several reservations at "routine" precedence and one at "priority" precedence. For sake of argument, let us presume that the link from VPN7 to R9 is now fully utilized - all of the bandwidth allocated by policy to voice at the routine or priority level has been reserved. Let us further imagine that a new "priority" reservation is now placed from H3 to H6. The process described in Section 2.3.1 is followed, resulting in PATH state across the network for the new reservation. This is installed even though it is not possible to install a new reservation on VPN7- R9, as it does not install any reservation and the network does not know whether H6 will ultimately require a reservation. The process described in Section 2.3.2 is also followed. The application in H6 decides to install the indicated reservation, meaning that it now replies with an RESV signal. Following the trail left by the PATH signal, the RESV signal traipses back towards H3. VPN6 and (if RSVP was configured) R6 verify that there is sufficient bandwidth on their links and that any other stated policy is also met. Having accomplished that, each will update its reservation state and forward the RESV signal to the next. VPN6 also generates an RESV for the reservation within the interface domain, attempting to set or increase the bandwidth of the reservation appropriately. VPN6, R8, and VPN8's "interface domain" sides now verify that there is sufficient bandwidth on their links and that any other stated policy is also met. Having accomplished that, each will update its reservation state and forward the RESV signal to the next. VPN8 will also generate an RESV for the reservation within the inner domain, attempting to set or increase the bandwidth of the reservation appropriately. This gets the reservation to the inner deaggregator, VPN8. VPN8's "inner domain" side and R9 now verify that there is sufficient bandwidth on their links and that any other stated policy is also met. At R9, a problem is detected - there is not sufficient bandwidth under the relevant policy. In the absence of precedence, R9 would now return an RESV Error indicating that the reservation could not be increased or installed. In such a case, if a preexisting reservation of lower bandwidth already existed, the previous reservation would remain in place but the new bandwidth would not be granted, and the originator (H6) would be informed. Let us clarify what it means to be at a stated precedence: it means that the POLICY_DATA object in the RESV contains a Preemption Priority and a Defending Priority with values specified in some memo. With precedence, [RFC4495]'s algorithm would have the Preemption Priority of the new reservation compared to the Defending Priority of extant reservations in the router, of which there are two: one VPN7->VPN8 at "routine" precedence and one VPN7->VPN8 at "priority" precedence. Since the Defending Priority of routine reservation is less than the Preemption Priority of a "priority" reservation, the "routine" reservation is selected. R9 determines that it will accept the increase in its "priority" reservation VPN7->VPN8 and reduce the corresponding "routine" reservation. Two processes now occur in parallel: o The routine reservation is reduced following the algorithms in [RFC4495] and o The priority reservation continues according to the usual rules. R9 reduces its "routine" reservation by sending an RESV Error updating its internal state to reflect the reduced reservation and sending an RESV Error to VPN8 requesting that it reduce its reservation to a number less than or equal to the relevant threshold less the sum of the competing reservations. VPN8, acting as a deaggregator, makes two changes. On the "inner domain" side, it marks its reservation down to the indicated rate (the most it is now permitted to reserve), so that if an RESV Refresh event happens, it will request the specified rate. On the "interface domain" side, it selects one or more of the relevant reservations by an algorithm of its choosing and requests that it likewise reduce its rate. For the sake of argument, let us imagine that the selected reservation is the one to VPN5. The RESV Error now makes its way through R8 to VPN5, which similarly reduces its bandwidth request to the stated amount and passes a RESV Error signal on the "enclave" side requesting that the reservation be appropriately reduced. H5 is now faced with a decision. If the request is to reduce its reservation to zero, that is equivalent to tearing down the reservation. In this simple case, it sends an RESV Tear to tear down the reservation entirely and advises its application to adjust its expectations of the session accordingly, which may mean shutting down the session. If the request is to reduce it below a certain value, however, it may be possible for the application to do so and remain viable. For example, if a VoIP application using a G.711 codec (80 kbps) is asked to reduce its bandwidth below 70 kbps, it may be possible to renegotiate the codec in use to G.729 or some other codec. In such a case, the originating application should re-reserve at the stated bandwidth (in this case, 70 kbps), initiate the application level change, and let the application change the reservation again (perhaps to 60 kbps) when it has completed that process. At the time the reservation is being processed at R9, for the "priority" reservation, R9 believes that it has sufficient bandwidth and that any other stated policy is also met, and it forwards the RESV to VPN7. Each will update its reservation state and forward the RESV signal to the next. VPN7 now acts as an [RFC3175] aggregator for the inner domain. This means that it receives the RESV signal for the inner domain reservation and stores state, decrypts the data stream from VPN8, operates on the RSVP signals as an RSVP-configured router, and forwards the received IP datagrams (including the updated RESV signals) into its interface domain. The RESV signals originated by VPN4, VPN5, and VPN6 are therefore forwarded towards VPN1, VPN2, and VPN3 through the interface domain. VPN3 now acts as an [RFC3175] aggregator for the interface domain. This means that it receives the RESV signal for the interface domain reservation and stores state, decrypts the data stream from its peer, operates on the RSVP signals as an RSVP-configured router, and forwards the received IP datagrams (including the updated RESV signals) into its enclave. The RESV signal originated by H6 is therefore forwarded towards H3 according to the routing of the enclave. H3 now receives the original RESV signals and delivers it to the relevant application. 3. Data Flows within a VPN Router This section details the data flows within a VPN router, in the context of sessions as described in Section 2. It specifically identifies the signaling flow at a given VPN boundary and additionally elaborates the signaling mechanism with the aid of a Network Guard. A use case describing the proposal in the context of an operational scenario is presented herein. 3.1. VPN Routers That Carry Data across the Cryptographic Boundary 3.1.1. Plaintext to Ciphertext Data Flows +-----------------------+ +----------------------+ | +--------------------+| |+--------------------+| | |RSVP || ||Aggregate RSVP || | | || || || | |Per session: || ID ||Agg. Session || | | Destination ||--->|| Agg. Destination || | | Source || || Agg. Source= self || | | potential SPI || || Agg. SPI generated|| | | DSCP ---------> DSCP || | | vPort or protocol---------> vPort || | | and port || || || | | Mean rate ---------> Sum of mean rates || | | Peak rate ---------> f(Peak rates) || | | Burst Size ---------> Sum of Burst sizes|| | | || || || | +--------------------+| |+--------------------+| | +--------------------+| |+--------------------+| | | IP || || IP || | +--------------------+| |+--------------------+| | +--------------------+| |+--------------------+| | | Plaintext Interface|| ||Ciphertext Interface|| | +--------------------+| |+--------------------+| +-----------------------+ +----------------------+ Figure 6: Data Flows in a VPN Router Outbound Parameters on a reservation include: Destination Address: On the plaintext side, the VPN router participates in the end-to-end reservations being installed for plaintext sessions. These may include individual flows as described in [RFC2205], IPsec data flows [RFC2207], aggregate reservations [RFC3175], or other types. It passes an identifier for the ciphertext side of the deaggregator to its ciphertext unit. DSCP: The DSCP of the plaintext data flow is provided to the cipher text side. Virtual Port: The virtual destination port is provided to the cipher text side. This may be derived from an [RFC2207] session object or from policy information. Mean Rate: The sum of the plaintext mean rates is provided to the ciphertext unit. Peak Rate: A function of the plaintext peak rates is provided to the ciphertext unit. This function is less than or equal to the sum of the peak rates. Burst Size: The sum of the burst sizes is provided to the cipher text unit. Messages include: Path: The plaintext PATH message is sent as encrypted data to the ciphertext unit. In parallel, a trigger needs to be sent to the ciphertext unit that results in it generating the corresponding aggregated PATH message for the ciphertext side. Path Error: This indicates that a PATH message sent to the remote enclave was in error. In the error case, the message itself is sent on as encrypted data, but a signal is sent to the ciphertext side in case the error affects the ciphertext reservation (such as removing or changing state). Path Tear: The PATH Tear message is sent as encrypted data to the ciphertext unit. In parallel, a signal is sent to the cipher text side; it will trigger a Path Tear on its reservation in the event that this is the last aggregated session, or change the SENDER_TSPEC of the aggregated session. RESV: The plaintext RESV message is sent as encrypted data to the ciphertext unit. In parallel, a trigger needs to be sent to the ciphertext unit that results in it generating the corresponding aggregated RESV message for the ciphertext side. RESV Error: This indicates that a RESV message that was received as data and forwarded into the enclave was in error or needed to be preempted as described in [RFC3181] or [RFC4495]. In the error case, the message itself is sent on as encrypted data, but a signal is sent to the ciphertext side in case the error affects the ciphertext reservation (such as removing or changing state). RESV Tear: The RESV Tear message is sent as encrypted data to the ciphertext unit. In parallel, a signal is sent to the cipher text side; it will trigger a RESV Tear on its reservation in the event that this is the last aggregated session, or reduce the bandwidth of an existing reservation. RESV Confirm: This indicates that a RESV message received as data and forwarded into the enclave, and is now being confirmed. This message is sent as encrypted data to the ciphertext side, and, in parallel, a signal is sent to potentially trigger an RESV Confirm on the aggregate reservation. 3.1.2. Ciphertext to Plaintext Data Flows +-----------------------+ +----------------------+ | +--------------------+| |+--------------------+| | |RSVP || ||Aggregate RSVP || | | || || terminated || | |Per session: |+ || || | | Destination || || || | | Source <---------Decrypted RSVP || | | potential SPI || || message sent to || | | DSCP || || Plaintext unit || | | vPort or protocol || || *as data* for || | | and port || || normal processing || | | Mean rate || || || | | Peak rate || || || | | Burst Size || || || | | || || || | +--------------------+| |+--------------------+| | +--------------------+| |+--------------------+| | | IP || || IP || | +--------------------+| |+--------------------+| | +--------------------+| |+--------------------+| | |Plaintext Interface || ||Ciphertext Interface|| | +--------------------+| |+--------------------+| +-----------------------+ +----------------------+ Figure 7: Data Flows in a VPN Router Inbound The aggregate reservation is terminated by the ciphertext side of the VPN router. The RSVP messages related to the subsidiary sessions are carried in the encrypted tunnel as data, and therefore arrive at the plaintext side with other data. As the plaintext side participates in these reservations, some information is returned to the ciphertext size to parameterize the aggregate reservation as described in Section 3.1.1 in the processing of the outbound messages. 3.2. VPN Routers That Use the Network Guard for Signaling across the Cryptographic Boundary As described in Section 1.6 the Network Guard provides an additional path for the reservation signaling between the plaintext and cipher text domains. _.------------. ,--'' Plaintext Domain--. ,-' +--------+ +--------+ `-. ,' | Host | | Host | `. ,' +--------+ +--------+ `. ; : | +----------------------+ | : | +--------+ | | `. | | Router | | ,' `. | +---+----+ | ,' `- | +----------+ | ,' ---| +-+--+ +-+--+--+ |' |----|E/D |--|Net Grd| | VPN Router ,-'| +-+--+ +-+--+--+ |\ , | +----------+ | \ ,' | +---+----+ | `. ,' | | Router | | | / | +--------+ | \ ; +----------------------+ : | | : Ciphertext Domain ; Figure 8: RSVP Passage via Network Guard In this context, the VPN router is composed of a plaintext router, a ciphertext router, an encrypt/decrypt implementation (such as a line card or interface device), and a network management process that manages the encrypt/decrypt implementation and potentially passes defined information flows between the plaintext and ciphertext domains. If the Network Guard is implemented as a software process that exchanges configuration instructions between the routers, this is simple to understand. If it is built as a separate systems exchanging datagrams, it is somewhat more complex, but conceptually equivalent. For example, the ciphertext router would consider an IP datagram received via the Network Guard (control plane) as having been received from and concerning the interface used in the data plane to the encrypt/decrypt unit. 3.2.1. Signaling Flow Encrypt/decrypt units may not be capable of terminating and originating flows as described in Section 3.1, and policy may prevent knowledge of the ciphertext network addresses in the plaintext router. In such a case, the plaintext and ciphertext routers may use the Network Guard as the path for the signaling flows. The Network Guard performs the following functions to enable the flow of reservation signaling across the cryptographic domain o transforms plaintext session identifiers into ciphertext session identifiers and vice-versa in IP datagrams and RSVP objects (e.g. IP addresses) o performs resource management of aggregated reservations (e.g., including ciphertext encapsulation overhead to resources requested) o reads and writes configuration on the encrypt/decrypt units as necessary (e.g., reads plaintext to ciphertext IP address mapping) In addition, the plaintext and ciphertext routers must support a routing function or local interface that ensures that aggregated RSVP messages flow via the Network Guard. However, the signaling flow across the entire VPN router at a cryptographic boundary remains identical to the description in Section 3.1. A reader may note that the VPN router described in Figure 8 can be collapsed into a single router with two halves, or the Network Guard and the encrypt/decrypt units can be part of the plaintext router. The details of alternate logical and physical architectures for the VPN router are beyond the scope of this document. 3.2.2. Use Case with Network Guard ........................................ : VPN Router A : : : :+-----------++----------++-----------+: +------+ RSVP :| || NetGrd-A || |: |Host A|<---->:|PT-Router-A|+----------+|CT-Router-A|:::::::: +------+ :| || E/D-A || |: :: :+-----------++----------++-----------+: :: : A-RSVP : :: : <:::::::::::::> : :: :......................................: :: A-RSVP :: ,---. ,' `. / \ ; Interface : | Domain | : ; \ / `. ,' '---' A-RSVP :: ........................................ :: : VPN Router B : :: : : :: :+-----------++----------++-----------+: :: +------+ RSVP :| || NetGrd-B || |: :: |Host B|<---->:|PT-Router-B|+----------+|CT-Router-B|:::::::: +------+ :| || E/D-B || |: :+-----------++----------++-----------+: : A-RSVP : : <:::::::::::::> : :......................................: Figure 9: Aggregated RSVP via Network Guard The above figure depicts a simple use case for aggregated signaling with the Network Guard. In this scenario, Host A initiates RSVP signaling to Host B for a reservation. The RSVP signaling between the hosts is encapsulated by the VPN routers into encrypted tunnels. Aggregated RSVP signaling is triggered by VPN routers, and flows into the CT-Routers, as well as the interface domains, to reserve resources at RSVP-capable routers on the path. The aggregation/ deaggregation point for RSVP reservations in this use case are the PT-Routers. The signaling aggregation of RSVP into A-RSVP at the PT-Router is similar to the data flow described in Section 3.1. The Network Guard performs the additional functions described in Section 3.2.1 to transform plaintext A-RSVP messages into suitable ciphertext A-RSVP messages. A typical reservation set up in this case would follow these steps. o Host A sends RSVP PATH message to Host B. o PT-Router-A encapsulates RSVP PATH message in encrypted tunnel to VPN Router B. o CT Routers and Interface domain carry encrypted RSVP PATH message (like any other encrypted data message). o PT-Router-B decrypts RSVP Path Message and sends an E2E PathErr message to PT-Router-A in the encrypted tunnel. o PT-Router-B forwards decrypted plaintext RSVP PATH message to Host B. o PT-Router-A receives E2E PathErr and sends an aggregated RSVP PATH message towards PT-Router-B via the Network Guard. o The NetGrd-A transforms the plaintext aggregate RSVP into the ciphertext aggregate RSVP message as described in Section 3.2.1 and sends it to the CT-Router-A. o The ciphertext aggregated RSVP message travels through ciphertext routers in the interface domain. o CT-Router-B receives the ciphertext aggregate RSVP message and sends it to the NetGrd-B. o The NetGrd-B transforms the ciphertext aggregate RSVP into the plaintext aggregate RSVP message as described in Section 3.2.1 and sends it to the PT-Router-B. The subsequent RSVP and Aggregate RSVP signaling follows a similar flow, as described in detail in [RFC3175] and [RFC4860]to aggregate each plaintext reservation into a corresponding ciphertext reservation. This ensures that RSVP-capable ciphertext routers reserve the required resources for a plaintext end-to-end reservation. Subsequent mechanisms, such as preemption or the increase and decrease of resources reserved, may be applied to these reservations as described before in this document. The RSVP data flow as described in Section 3.1 within the VPN router (from the plaintext router to the ciphertext router via the Guard) provides necessary and sufficient information to routers in the ciphertext domain to implement the QoS solution presented in the document. In this description, we have described the Network Guard as being separate from the encrypt/decrypt unit. This separation exists because in certain implementations, it is mandated by those who specify the devices. The separation does not come for free, however; the separation of the devices for system-engineering purposes is expensive, and it imposes architectural problems. For example, when the Guard is used to aggregate RSVP messages or Protocol Independent Multicast (PIM) routing, the traffic is destined to the remote VPN router. This means that the Guard must somehow receive and respond to, on behalf of the VPN Router, messages that are not directed to it. Several possible solutions exist; they should be selected carefully based on the security and implementation needs of the environment. They are as follows: o In the simplest case, the Network Guard and encrypt/decrypt unit can be two independent functions that utilize a common network and MAC layer. This can allow the two functions to share a common MAC and IP address, so that traffic destined for one function is also received by the other. In the case that these two functions are physically separated on two devices, they can still share a common MAC and IP address; however, additional modifications may be required on the Guard to filter and not process IP traffic not destined for itself. o The ciphertext interface of the Guard could be placed into promiscuous mode, allowing it to receive all messages and discard all but the few it is interested in. The security considerations on putting a device in promiscuous mode at the VPN boundary needs to be taken into account in this method. o The Guard could be engineered to receive all from the ciphertext router and pass the bulk of it on to the VPN router through another interface. In this case, the Guard and the VPN router would use the same IP address. This mechanism puts the load of all data and management traffic destined for the VPN router upon the Guard. o The VPN router could be engineered to receive all traffic from the ciphertext router and pass any unencrypted traffic it receives to the Guard through another interface. In this case, the Guard and the VPN router would use the same IP address. o All the VPN router functions, as shown in Figure 9, could be incorporated into a single chassis, with appropriate internal traffic management to send some traffic into the plaintext enclave and some to the Guard. In this case, the Guard and the VPN router would be -- at least, functionally -- the same system. Of these, clearly the last is the simplest architecturally and the one that most minimizes the attendant risk. 4. Security Considerations The typical security concerns of datagram integrity, node and user authentication are implicitly met by the security association that exists between the VPN routers. The secure data stream that flows between the VPN routers is also used for the reservation signaling datagrams flowing between VPN routers. Information that is contained in these signaling datagrams receives the same level of encryption that is received by the data streams. One of the reasons cited for the nesting of VPN routes in Section 1.3 is the different levels of security across the nested VPN routers. If the security level decreases from one VPN router to the next VPN Router in the nested path, the reservation signaling datagrams will, by default, receive the lower security-level treatment. For most cases, the lower security treatment is acceptable. In certain networks, however, the reservation signaling across the entire nested path must receive the highest security-level treatment (e.g., encryption, authentication of signaling nodes). For example, the highest precedence level may only be signaled to VPN routers that can provide the highest security levels. If any VPN router in the nested path is incapable of providing the highest security level, it cannot participate in the reservation mechanism. In the general case, the nested path may contain routers that are either incapable of participating in VPNs or providing required security levels. These routers can participate in the reservation only if the lower security level is acceptable (as configured by policy) for the signaling of reservation datagrams. VPN routers encapsulate encrypted IP packets and prepend an extra header on each packet. These packets, whether used for signaling or data, should be identifiable, at a minimum by the IP addresses and DSCP value. Therefore, the prepended header should contain, at a minimum, the DSCP value corresponding to the signaled reservation in each packet. This may literally be the same DSCP as is used for the data (forcing control plane traffic to receive the same QoS treatment as its data), or a different DSCP that is routed identically (separating control and data-plane traffic QoS but not routing). Additionally security considerations as described in [RFC4860] and [RFC3175] are also applicable in this environment; they include the integrity of RSVP messages can be ensured via mechanisms described in [RFC2747] and [RFC3097] and related key management (through manual configuration or a key management protocol) at nodes between any aggregator and deaggregator pair that processes the messages. In addition, confidentiality can be provided between hops by employing IPsec. Further work in the IETF MSEC Working Group may be applicable in these environments for key management and confidentiality. 5. Acknowledgements Doug Marquis, James Polk, Mike Tibodeau, Pete Babendreier, Roger Levesque, and Subha Dhesikan gave early review comments. Comments by Sean O'Keefe, Tony De Simone, Julie Tarr, Chris Christou, and their associates resulted in Section 3.2. Francois Le Faucheur, Bruce Davie, and Chris Christou (with Pratik Bose) added [RFC4860], which clarified the interaction of this approach with the DSCP. 6. References 6.1. Normative References [RFC2205] Braden, B., Zhang, L., Berson, S., Herzog, S., and S. Jamin, "Resource ReSerVation Protocol (RSVP) -- Version 1 Functional Specification", RFC 2205, September 1997. [RFC2207] Berger, L. and T. O'Malley, "RSVP Extensions for IPSEC Data Flows", RFC 2207, September 1997. [RFC2746] Terzis, A., Krawczyk, J., Wroclawski, J., and L. Zhang, "RSVP Operation Over IP Tunnels", RFC 2746, January 2000. [RFC2750] Herzog, S., "RSVP Extensions for Policy Control", RFC 2750, January 2000. [RFC2996] Bernet, Y., "Format of the RSVP DCLASS Object", RFC 2996, November 2000. [RFC3175] Baker, F., Iturralde, C., Le Faucheur, F., and B. Davie, "Aggregation of RSVP for IPv4 and IPv6 Reservations", RFC 3175, September 2001. [RFC4495] Polk, J. and S. Dhesikan, "A Resource Reservation Protocol (RSVP) Extension for the Reduction of Bandwidth of a Reservation Flow", RFC 4495, May 2006. [RFC4542] Baker, F. and J. Polk, "Implementing an Emergency Telecommunications Service (ETS) for Real-Time Services in the Internet Protocol Suite", RFC 4542, May 2006. [RFC4860] Le Faucheur, F., Davie, B., Bose, P., Christou, C., and M. Davenport, "Generic Aggregate Resource ReSerVation Protocol (RSVP) Reservations", RFC 4860, May 2007. 6.2. Informative References [ITU.MLPP.1990] International Telecommunications Union, "Multilevel Precedence and Preemption Service", ITU-T Recommendation I.255.3, 1990. [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, September 1981. [RFC1633] Braden, B., Clark, D., and S. Shenker, "Integrated Services in the Internet Architecture: an Overview", RFC 1633, June 1994. [RFC2209] Braden, B. and L. Zhang, "Resource ReSerVation Protocol (RSVP) -- Version 1 Message Processing Rules", RFC 2209, September 1997. [RFC2210] Wroclawski, J., "The Use of RSVP with IETF Integrated Services", RFC 2210, September 1997. [RFC2460] Deering, S. and R. Hinden, "Internet Protocol, Version 6 (IPv6) Specification", RFC 2460, December 1998. [RFC2474] Nichols, K., Blake, S., Baker, F., and D. Black, "Definition of the Differentiated Services Field (DS Field) in the IPv4 and IPv6 Headers", RFC 2474, December 1998. [RFC2475] Blake, S., Black, D., Carlson, M., Davies, E., Wang, Z., and W. Weiss, "An Architecture for Differentiated Services", RFC 2475, December 1998. [RFC2747] Baker, F., Lindell, B., and M. Talwar, "RSVP Cryptographic Authentication", RFC 2747, January 2000. [RFC2872] Bernet, Y. and R. Pabbati, "Application and Sub Application Identity Policy Element for Use with RSVP", RFC 2872, June 2000. [RFC3097] Braden, R. and L. Zhang, "RSVP Cryptographic Authentication -- Updated Message Type Value", RFC 3097, April 2001. [RFC3181] Herzog, S., "Signaled Preemption Priority Policy Element", RFC 3181, October 2001. [RFC3182] Yadav, S., Yavatkar, R., Pabbati, R., Ford, P., Moore, T., Herzog, S., and R. Hess, "Identity Representation for RSVP", RFC 3182, October 2001. [RFC3246] Davie, B., Charny, A., Bennet, J., Benson, K., Le Boudec, J., Courtney, W., Davari, S., Firoiu, V., and D. Stiliadis, "An Expedited Forwarding PHB (Per-Hop Behavior)", RFC 3246, March 2002. [RFC3261] Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, A., Peterson, J., Sparks, R., Handley, M., and E. Schooler, "SIP: Session Initiation Protocol", RFC 3261, June 2002. [RFC3312] Camarillo, G., Marshall, W., and J. Rosenberg, "Integration of Resource Management and Session Initiation Protocol (SIP)", RFC 3312, October 2002. [RFC3473] Berger, L., "Generalized Multi-Protocol Label Switching (GMPLS) Signaling Resource ReserVation Protocol-Traffic Engineering (RSVP-TE) Extensions", RFC 3473, January 2003. [RFC3474] Lin, Z. and D. Pendarakis, "Documentation of IANA assignments for Generalized MultiProtocol Label Switching (GMPLS) Resource Reservation Protocol - Traffic Engineering (RSVP-TE) Usage and Extensions for Automatically Switched Optical Network (ASON)", RFC 3474, March 2003. [RFC4301] Kent, S. and K. Seo, "Security Architecture for the Internet Protocol", RFC 4301, December 2005. [RFC4303] Kent, S., "IP Encapsulating Security Payload (ESP)", RFC 4303, December 2005. Authors' Addresses Fred Baker Cisco Systems 1121 Via Del Rey Santa Barbara, California 93117 USA Phone: +1-408-526-4257 Fax: +1-413-473-2403 EMail: email@example.com Pratik Bose Lockheed Martin 700 North Frederick Ave Gaithersburg, Maryland 20871 USA Phone: +1-301-240-7041 Fax: +1-301-240-5748 EMail: firstname.lastname@example.org Full Copyright Statement Copyright (C) The IETF Trust (2007). This document is subject to the rights, licenses and restrictions contained in BCP 78, and except as set forth therein, the authors retain all their rights. This document and the information contained herein are provided on an "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Intellectual Property The IETF takes no position regarding the validity or scope of any Intellectual Property Rights 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; nor does it represent that it has made any independent effort to identify any such rights. Information on the procedures with respect to rights in RFC documents can be found in BCP 78 and BCP 79. Copies of IPR disclosures made to the IETF Secretariat 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 implementers or users of this specification can be obtained from the IETF on-line IPR repository at http://www.ietf.org/ipr. The IETF invites any interested party to bring to its attention any copyrights, patents or patent applications, or other proprietary rights that may cover technology that may be required to implement this standard. Please address the information to the IETF at email@example.com. Acknowledgement Funding for the RFC Editor function is currently provided by the Internet Society.