Internet Engineering Task Force T. Przygienda/P. Droz INTERNET DRAFT Fore/IBM 30 October 1997 OSPF over ATM and Proxy PAR Status of This Memo This document is an Internet Draft, and can be found as draft-ietf-ospf-atm-00.txt in any standard internet drafts repository. Internet Drafts are working documents of the Internet Engineering Task Force (IETF), its Areas, and its Working Groups. Note that other groups may also distribute working documents as Internet Drafts. Internet Drafts are draft documents valid for a maximum of six months. Internet Drafts may be updated, replaced, or obsoleted by other documents at any time. It is not appropriate to use Internet Drafts as reference material, or to cite them other than as a ``working draft'' or ``work in progress.'' Please check the I-D abstract listing contained in each Internet Draft directory to learn the current status of this or any other Internet Draft. Abstract This draft specifes for OSPF implementors and users mechanisms describing how the protocol operates in ATM networks over PVC and SVC meshes with the presence of Proxy PAR. These recommendations do not require any protocol changes and allow for simpler, more efficient and cost- effective network designs. It is recommended that OSPF implementations should be able to support logical interfaces, each consisting of one or more virtual circuits and used as numbered logical point-to-point links (one VC) or logical NBMA networks (more than one VC) where a solution simulating broadcast interfaces is not appropriate. Proxy PAR can help to distribute configuration changes of such interfaces when OSPF capable routers are reconfigured on the ATM cloud. Przygienda, Droz Expires 5 May 1998 [Page 1] Internet Draft OSPF over ATM and Proxy PAR 30 October 1997 1. Introduction 1.1. Introduction to Proxy PAR Proxy PAR [CPS96, PD97] is an extension allowing for different ATM attached devices to interact with PAR capable switches and obtain information about non-ATM services without executing PAR [Ca96] which is an extension of PNNI [AF96b] themselves. The client side is much simpler in terms of implementation complexity and memory requirements than a complete PAR stack and should allow for easy implementation in e.g. existing IP routers. Additionally, clients can use Proxy PAR to register different non-ATM services and protocols they support. Proxy PAR has consciously not been included as part of ILMI due to the complexity of PAR information passed in the protocol and the fact that it is intended for integration of non-ATM protocols and services only. A device executing Proxy PAR does not necessarily need to execute ILMI or UNI signaling although this normally will be the case. The context or reference model is therefore aligned with the one included in [AF96a]. The protocol in itself does not specify how the distributed service registration and data delivered to the client is supposed to be driving other protocols so OSPF routers finding themselves through proxy PAR could use this information in e.g. RFC1577 [Lau94] fashion, forming a full mesh of point- to-point connections to interact with each other to simulate broadcast interfaces. For the same purpose LANE [AF95] or MARS [Arm96] could be used. Contrary to such solutions, this RFC elaborates on how to handle virtual OSPF interfaces over ATM such as NBMA, point-to-multipoint or point-to-point and allow for their auto-configuration in presence of Proxy PAR. One advantage is the circumvention of server solutions that often present single points of failure or hold large amounts of configuration information. The other main benefit is the possibility to execute OSPF on top of partially meshed VC topologies. As a by-product, Proxy PAR could provide the ATM address resolution for IP attached devices but such resolution can be achieved by other protocols under specification in IETF as well, e.g. [CH97a, CH97b]. Last but not least, it should be mentioned here that the protocol coexists with and complements the ongoing work in IETF on server detection via ILMI extensions [Dav97] and opaque LSAs [CH97a, CH97b]. Przygienda, Droz Expires 5 May 1998 [Page 2] Internet Draft OSPF over ATM and Proxy PAR 30 October 1997 1.1.1. Proxy PAR scopes Any Proxy PAR registration is carried only within a defined scope that is set during registration and is equivalent to the PNNI routing level. Since no assumptions except scope values can be made about the information distributed (e.g. IP addresses bound to NSAPs are not assumed to be aligned with them in any respect such as encapsulation or functional mapping), registration information cannot be summarized. This makes a careful handling of scopes necessary to preserve the scalability. 1.2. Introduction to OSPF OSPF (Open Shortest Path First) is an Interior Gateway Protocol (IGP) and described in [Moy94, Moy97] from which most of the following paragraphs has been taken almost literally. OSPF distributes routing information between routers belonging to a single Autonomous System. The OSPF protocol is based on link-state or SPF technology. It was developed by the OSPF working group of the Internet Engineering Task Force. It has been designed expressly for the TCP/IP internet environment, including explicit support for IP subnetting, and the tagging of externally-derived routing information. OSPF also utilizes IP multicast when sending/receiving the updates. In addition, much work has been done to produce a protocol that responds quickly to topology changes, yet involves small amounts of routing protocol traffic. To cope with the needs of NBMA and demand circuits capable networks such as Frame Relay or X.25, [Moy95] has been made available that standardizes extensions to the protocol allowing for efficient operation over on-demand circuits. OSPF supports three types of networks today: - Point-to-point networks: A network that joins a single pair of routers. Point- to-point networks can either be numbered or unnumbered in the latter case the interfaces do not have IP addresses nor masks. Even when numbered, both sides of the link do not have to agree on the IP subnet. - Broadcast networks: Networks supporting many (more than two) attached routers, together with the capability to address a single physical message to all of the attached routers Przygienda, Droz Expires 5 May 1998 [Page 3] Internet Draft OSPF over ATM and Proxy PAR 30 October 1997 (broadcast). Neighboring routers are discovered dynamically on these nets using OSPF's Hello Protocol. The Hello Protocol itself takes advantage of the broadcast capability. The protocol makes further use of multicast capabilities, if they exist. An Ethernet is an example of a broadcast network. - Non-broadcast networks: Networks supporting many (more than two) attached routers, but having no broadcast capability. Neighboring routers are maintained on these nets using OSPF's Hello Protocol. However, due to the lack of broadcast capability, some configuration information is necessary for the correct operation of the Hello Protocol. On these networks, OSPF protocol packets that are normally multicast need to be sent to each neighboring router, in turn. An X.25 Public Data Network (PDN) is an example of a non-broadcast network. OSPF runs in one of two modes over non-broadcast networks. The first mode, called non-broadcast multi-access (NBMA), simulates the operation of OSPF on a broadcast network. The second mode, called Point-to-MultiPoint, treats the non-broadcast network as a collection of point-to-point links. Non-broadcast networks are referred to as NBMA networks or Point-to-MultiPoint networks, depending on OSPF's mode of operation over the network. 2. OSPF over ATM 2.1. Model Parallel to [dR94] that describes the recommended operation of OSPF over Frame Relay networks, a similar model is assumed where the underlying ATM network can be used to model single VCs as point-to-point interfaces or collections of VCs can be accessed as an non-broadcast interface in NBMA or point-to-multipoint mode. Such a VC or collection of VCs is called a logical interface and specified through its type (either point-to-point, NBMA or point-to-point), IP instance (presenting an incarnation of IP with its own address spaces), address and mask. Layer 2 specific configuration such as address resolution method, class and quality of service of used circuits and other must be also included. As logical consequence thereof, a single, physical interface could encompass multiple IP subnets or even multiple, independent IP instances. In contrary to layer 2 and IP addressing information, when running Proxy PAR, most of the OPSF information needed to operate such a logical interface Przygienda, Droz Expires 5 May 1998 [Page 4] Internet Draft OSPF over ATM and Proxy PAR 30 October 1997 does not have to be configured into routers statically but can be provided through Proxy PAR queries. This allows for much more dynamic configuration of VC meshes in OSPF environments than e.g. in Frame Relay solutions. 2.2. OSPF Configuration Interaction with Proxy PAR To achieve the goal of simplification of VC mesh reconfiguration, Proxy PAR allows the router to learn automatically most of the configuration that has to be provided to OSPF. Non-broadcast and point-to-point interface information can be learned across an ATM cloud as described in the ongoing sections. It is up to the implementation to possibly allow for a mixture of Proxy PAR autoconfiguration and manual configuration of neighbor information. Moreover, manual configuration could e.g. override or complement information derived from a proxy PAR client. Additionally, OSPF extensions to handle on-demand circuits [Moy95] can be used to allow for graceful tearing down of VCs not carrying any OSPF traffic over prolonged periods of time. Even after autoconfiguration of interfaces has been provided, the problem of VC setups in an ATM network is unsolved since none of the normally used mechanisms such as RFC1577 [Lau94] or LANE [AF95] are assumed to be present. Section 2.3 describes the behavior of OSPF routers to allow for router connectivity necessary. 2.2.1. Autoconfiguration of non-broadcast interfaces Proxy PAR allows to autoconfigure the list of all routers residing on the same IP network in the same IP instance by simply querying the Proxy PAR server. Each router can easily obtain the list of all OSPF routers on the same subnet with their router priorities and ATM address bindings. This is the precondition for OSPF to work properly across such logical NBMA interfaces. Note that the memberlist, when learned through Proxy PAR queries, can dynamically change with PNNI (in)stability and general ATM network behavior. It maybe preferable for an implementation to withdraw list membership e.g. much slower than detect new members. Relying on OSPF mechanism to discover lack of reachability in the overlaying logical IP network could alleviate the risk of thrashing DR elections and excessive information flooding. Once the DR registration is completed and the router has not been elected DR or BDR, an implementation of [Moy95] Przygienda, Droz Expires 5 May 1998 [Page 5] Internet Draft OSPF over ATM and Proxy PAR 30 October 1997 can ignore the fact that all routers on the specific NBMA subnet are available in its configuration since it only needs to maintain VCs to the DR and BDR. 2.2.2. Autoconfiguration of point-to-multipoint interfaces Point-to-Multipoint interfaces in ATM networks only make sense if no VCs can be dynamically set up since an SVC-capable ATM network normally presents a NBMA cloud. This is e.g. the case if the intended use of the network is only to execute OSPF in presence of a partial PVC or SPVC mesh. Such a collection could be modeled using the point-to-multipoint OSPF interface and the neighbor detection could be provided by Proxy PAR or other means. In Proxy PAR case the router queries for all OSPF routers on the same network in the same IP instance but it installs in the interface configuration only routers that are already reachable through preset PVCs. The underlying assumption is that a router understands the remote NSAP of a PVC and can compare it with appropriate Proxy PAR registrations. If the remote NSAP of the PVC is unknown, alternative autodiscovery mechanisms have to be used e.g. inverse ARP [BB92]. 2.2.3. Autoconfiguration of numbered point-to-point interfaces OSPF point-to-point links do not necessarily have an IP address assigned and even when having one, the mask is undefined. As a precondition to successfully register a service with Proxy PAR, IP address and mask is required. Therefore, if a router desires to use Proxy PAR to advertise the local end of a point-to- point link to the router it intends to form an adjacency with, an IP address has to be provided and a netmask set or a default of 255.255.255.254 (this gives as the default case a subnet with 2 routers on it) assumed. To allow the discovery of the remote end of the interface, IP address of the remote side has to be provided and a netmask set or a default of 255.255.255.254 assumed. Obviously the discovery can only be successfull when both sides of the interface are configured with the same network mask and are within the same IP network. The situation where more than two possible neighbors are discovered through queries and the interface type is set to point-to-point presents a configuration error. Przygienda, Droz Expires 5 May 1998 [Page 6] Internet Draft OSPF over ATM and Proxy PAR 30 October 1997 2.2.4. Autoconfiguration of unnumbered point-to-point interfaces For reasons given already in [dR94] using unnumbered point-to-point interfaces with Proxy PAR is not a very attractive alternative since the lack of an IP address prevents efficient registration and retrieval of configuration information. Relying on the numbering method based on MIB entries generates conflicts with the dynamic nature of creation of such entries and is beyond the scope of this work. 2.3. Proxy PAR Interaction with OSPF Configuration To allow other routers to discover an OSPF interface automatically, the IP address, mask, Area ID, interface type and router priority information given must be registered with the Proxy PAR server at an appropriate scope. A change in any of these parameters has to force a reregistration with Proxy PAR. 2.3.1. Registration of non-broadcast interfaces For an NBMA interface the appropriate parameters are available and can be registered through Proxy PAR without further complications. 2.3.2. Registration of point-to-multipoint interfaces In case of a point-to-multipoint interface the router registers its information in the same fashion as in the NBMA case except that the interface type is modified accordingly. 2.3.3. Registration of point-to-point interfaces In case of point-to-point numbered interfaces the address mask is not specified in the OSPF configuration. If the router has to use Proxy PAR to advertise its capability, a mask must be defined or a default value of 255.255.255.254 used. Przygienda, Droz Expires 5 May 1998 [Page 7] Internet Draft OSPF over ATM and Proxy PAR 30 October 1997 2.3.4. Registration of unnumbered point-to-point interfaces Due to the lack of a configured IP address and difficulties generated by this fact as described earlier, registration of unnumbered point-to-point interfaces is not covered in this document. 2.4. Connection setup mechanisms This sections describes OSPF behavior in an ATM network under different assumptions in terms of signaling capabilities and preset connectivity. 2.4.1. OSPF in PVC environments In environments where only partial PVCs (or SPVCs) meshes are available and modeled as point-to-multipoint interfaces, the routers see reachable routers through autodiscovery provided by Proxy PAR. This leads to expected OSPF behavior. In cases where a full mesh of PVCs is present, such an interface should preferably be modeled as broadcast and Proxy PAR discovery should be superfluous. 2.4.2. OSPF in SVC environments In SVC-capable environments the routers can initiate VCs after having discovered the appropriate neighbors, preferably driven by the need to send data such as Hello-packets. Since this can lead to race conditions where both sides can open a VC and it is desirable to minimize this valuable resource, if the router with lower Router ID detects that the VC initiated by the other side is bidirectional, it is free to close its own VC and use the detected one. The existence of VCs used for OSPF exchanges is orthogonal to the number and type of VCs the router chooses to use within the logical interface to forward data to other routers. OSPF implementations are free to use any of these VCs to send packets if their endpoints are adequate and must accept hello packets arriving on any of the VCs belonging to the logical interface. Przygienda, Droz Expires 5 May 1998 [Page 8] Internet Draft OSPF over ATM and Proxy PAR 30 October 1997 3. Acknowledgments Comments and contributions from several sources, especially Rob Coltun and John Moy are included in this work. 4. Security Consideration Security issues are not discussed in this memo. References [AF95] ATM-Forum. LAN Emulation over ATM 1.0. ATM Forum af-lane-0021.000, January 1995. [AF96a] ATM-Forum. Interim Local Management Interface (ILMI) Specification 4.0. ATM Forum 95-0417R8, June 1996. [AF96b] ATM-Forum. Private Network-Network Interface Specification Version 1.0. ATM Forum af-pnni-0055.000, March 1996. [Arm96] G. Armitage. Support for Multicast over UNI 3.0/3.1 based ATM Networks, RFC 2022. Internet Engineering Task Force, November 1996. [BB92] T. Bradley and C. Brown. Inverse Address Resolution Protocol, RFC 1293. Internet Engineering Task Force, January 1992. [Ca96] R. Callon and al. An Overview of Pnni Augmented Routing. ATM Forum 96-0354, April 1996. [CH97a] R. Coltun and J. Heinanen. Opaque LSA in OSPF. Internet Draft, 1997. [CH97b] R. Coltun and J. Heinanen. The OSPF Address Resolution Advertisement Option. Internet Draft, 1997. [CPS96] R. Coltun, T. Przygienda, and S. Shew. MIPAR: Minimal PNNI Augmented Routing. ATM Forum 96-0838, June 1996. Przygienda, Droz Expires 5 May 1998 [Page 9] Internet Draft OSPF over ATM and Proxy PAR 30 October 1997 [Dav97] M. Davison. Simple ILMI-Based Server Discovery. Internet Draft, 1997. [dR94] O. deSouza and M. Rodrigues. Guidelines for Running OSPF Over Frame Relay Networks, RFC 1586. Internet Engineering Task Force, March 1994. [Lau94] M. Laubach. Classical IP and ARP over ATM, RFC 1577. Internet Engineering Task Force, January 1994. [Moy94] J. Moy. OSPFv2, RFC 1583. Internet Engineering Task Force, March 1994. [Moy95] J. Moy. Extending OSPF to Support Demand Circuits, RFC 1793. Internet Engineering Task Force, April 1995. [Moy97] J. Moy. OSPFv2, RFC 2178. Internet Engineering Task Force, July 1997. [PD97] T. Przygienda and P. Droz. Proxy PAR. ATM Forum 97-0495, 97-0705, 97-0882, July 1997. Authors' Addresses Tony Przygienda FORE Systems 6905 Rockledge Drive Suite 800 Bethesda, MD 20817 prz@fore.com Patrick Droz IBM Research Division Saumerstrasse 4 8803 Ruschlikon Switzerland dro@zurich.ibm.com Przygienda, Droz Expires 5 May 1998 [Page 10]