Internet Draft L. Yang Expiration: December 2002 Intel Labs File: draft-yang-forces-model-00.txt J. Halpern Working Group: ForCES R. Gopal Nokia R. Dantu Netrake June 2002 ForCES Forwarding Element Functional Model draft-yang-forces-model-00.txt Status of this Memo This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC2026. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet-Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as ``work in progress.'' The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt. The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html. Conventions used in this document The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC-2119]. 1. Abstract This document defines a functional model for ForCES forwarding elements (FEs). This model is used to describe the state of ForCES forwarding elements within the context of the ForCES protocol, so that ForCES control elements (CEs) can control the FEs accordingly. The model is to specify what logical function instances are present in the FEs and in what order these functions are performed. The Internet Draft ForCES FE Functional Model June 2002 forwarding element model defined herein is intended to satisfy the requirements specified in the ForCES requirements draft [FORCES- REQ]. Using this model, predefined or vendor specific logical functions can be expressed and configured. However, the definition of these components are not described and defined in this document. Definitions A set of terminology associated with the ForCES requirements is defined in [FORCES-REQ] and is not copied here. The following list of terminology is relevant to the FE model defined in this document. Datapath -- A conceptual path taken by packets within the forwarding plane, inside an FE. There might exist more than one datapath within an FE. Forwarding Element (FE) Block -- An abstraction of the basic packet processing functions in the datapath. It is the building block of FE functionality. This concept abstracts away implementation details from the parameters of interest for configuration, control and management by CE. The general rule of thumb on the granularity of FE blocks is that FE blocks should be coarse grained, stateful and focus on a particular domain. For example, RFC1812 compliant IPv4 forwarder, classifiers, meters, etc.. Forwarding Element (FE) Stage -- Representation of a FE block instance in a FE's datapath. As a packet flows through an FE along a datapath, it flows through one or multiple distinct stages, with each stage implementing an instance of a certain logical function. There may be one or more combination of such instances in a FE's datapath. Using NAT as an example, one NAT function is typically performed before the forwarding stage (packets arriving externally have their public addresses replaced with private addresses) and one NAT function is performed after (for packets exiting the domain, their private addresses are replaced by public ones). So there are three stages (NAT, forwarding, and NAT again) in this example FE, with two NAT instances present in two different stages. 2. Motivation and Requirements of FE model The ForCES architecture allows Forwarding Elements (FEs) of varying functionality to participate in a ForCES network element (NE). The implication of this varying functionality is that CEs can make only minimal assumptions about the functionality provided by its FEs. Before CEs can configure and control the forwarding behavior of FEs, CEs need to query and discover the capabilities and state of their FEs. [FORCES-REQ] mandates that this capability and state information be expressed in the form of a FE model, and this model will be used as the basis for CEs to control FEs' capabilities and manipulate FEs' state via ForCES protocol. [FORCES-REQ] describes all the requirements placed on the FE model in detail. We provide a brief summary here to highlight some of the design issues we face. Yang, et. al. Expires December 2002 [Page 2] Internet Draft ForCES FE Functional Model June 2002 . The FE model MUST express what logical functions can be applied to packets as they pass through a FE. . The FE model MUST be capable of supporting/allowing variations in the way logical functions are implemented on a FE. . The model MUST be capable of describing the order in which these logical functions are applied in a FE. . The FE model SHOULD be extendable and should have provision to express new or vendor specific logical functions. . . Should be able to support minimal set of logical functions that are already identified such as port functions, forwarding functions, QoS functions, filtering functions, high- touch functions, security functions, and vendor-specific functions. Since the motivation of an FE model is to allow CEs later to control and configure FEs' behavior via ForCES protocol, it becomes essential to examine and understand what kind of control and configuration CEs might do to FEs. We believe that there are roughly three levels of control and configuration that CEs can do to FEs. The first level of control and configuration is the simplest of all. It assumes that each FE's capability is already given and remains static in its lifetime, and CEs can only control its behavior by manipulating its state. We call this "static FE" control and configuration. For example, Figure 2 and 3 each shows an FE configuration example by representing the processing steps in a directed graph interconnecting all the functional stages that packets can possibly traverse. If such a configuration remains static during FE's lifetime, then all CE can control is the parameters associated with each stage in the graph, for example, the routing table in the LPM forwarder in Figure 2, or the token bucket parameters associated with meter1 in Figure 3. But CE cannot reconfigure the graph topology dynamically, such as adding another meter or queue onto the FE in Figure 3 on the fly. For this kind of static FE control and configuration purpose, the useful FE model is consisted of the state information that FEs allow CEs to manipulate and the statistics and events that FEs can collect and report back to CEs. Such a state model doesn't need to describe a lot of other information, like the packet formats supported between meter1 and counter1 in Figure 3, for example. Because such information is only useful when the graph is re-configured dynamically on the fly. The second level of control and configuration builds on top of the first that is just described. Using Figure 3 as an example, instead of presenting the static FE graph to CE, FE can convey its capabilities to CE by telling CE that "This FE can support one classifier with up to N filters. This FE can also support M meters, X queues, etc." We call this dynamic FE control and configuration. For such control and configuration, a more powerful and flexible FE capability model is required in addition to the FE state model. For example, it becomes necessary to model the capability of the building blocks like classifiers, filters, meters etc. it also needs Yang, et. al. Expires December 2002 [Page 3] Internet Draft ForCES FE Functional Model June 2002 to describe the packet formats supported at each input and output of each block to allow correct reconfiguration of the graph by CE. The third level of control and configuration is even more powerful and future looking. In addition to dynamic configuration, CEs might even be allowed to download a given functionality onto FEs at run time. The FE model proposed in this document intends to fully support the static FE control and configuration. It is also our intention to allow extension later to support dynamic FE control and configuration. However, this FE model currently makes no attempt to address issues beyond the simple static FE control and configuration. 3. Capability and configuration Model versus State Model FE Model describes both the static capabilities and the configured capabilities. At any time after initial configuration the logical functions associated with an FE model may exhibit different characteristics and may lead to change in state of FE. The capability describes set of parameters or attributes of one or more logical functions of a FE, which may be useful for configuring and managing a FE. The state model describes the instantaneous values or operational behaviour of a FE. Typical capabilities model describes at the coarsest level such aspects as - this FE can handle IPv4 and IPv6 - this FE can perform 6-tuple classification (or can only do DA based processing, ...) - this FE can perform metering - this FE can handle multiple queues with multiple priorities - this FE can add and remove encapsulating headers of types: IPSec, GRE, L2TP Where it gets more complicated is coping with the detailed limits. Issues such has how many classifiers can the FE handle? How many headers can the FE add and remove? How many queues, and how many buffer pools can the FE support? How many meters can the FE provide? How flexibly can these various parts be interconnected? While one could try to build an object model for representing capabilities in full, other efforts have found this to be a significant undertaking. A middle of the road approach is to define coarse-grained capabilities and simple capacity measures. Then, if the CE attempts to instruct the FE to set up some specific behavior it is not capable of, the FE will return an error indicating the problem. A state model describes the current state of the FE. It lists - on a given port the packets are classified using a given classification filter - a given classifier results in packets being metered in a certain way, and then marked in a certain way Yang, et. al. Expires December 2002 [Page 4] Internet Draft ForCES FE Functional Model June 2002 - packets coming from specific markers are delivered into a shared queue for handling, while other packets are delivered to a different queue - a specific scheduler with specific behavior and parameters will service these collected queues. While the DiffServ and QDDIM models are not designed with the primary goal of direct machine implementation, we will start with that model as it is better than no starting point. Alternative suggestions for a state model which can be implemented more generally without significant transformation are sought from the community. 4. FE Model This section proposes a ForCES FE model to satisfy all the requirements in [FORCES-REQ] for FE control and configuration. The approach taken is to first define a set of well-known FE logical functions (FE blocks) as the basic building blocks, so that any FE can be modeled by describing the kind of FE blocks or variants of these blocks it contains, and how the instances of these blocks are interconnected together. This model also allows new logical functions be added to accommodate future innovation in forwarding plane. We believe this approach strikes a good balance between flexibility and extensibility of the model and ease of use by the CE. 4.1. FE Blocks FE blocks are the very basic building blocks from which FEs model its overall packet processing behavior. The concept of a FE block is akin to that of an abstract base class in object-oriented terminology. Different FEs can have different implementation of the same block, but the packet processing behavior looks the same to the CE. Sometimes it is difficult to decide the granularity of the FE blocks. The general rule of thumb is that FE blocks should be coarse grained, stateful and focus on a particular domain. Basically, if a logical function doesnĘt have anything to configure, doesnĘt keep any kind of inter-packet state information (tables, etc), then it probably is not a FE block. For example, CRC check and IP checksum are not considered to be FE blocks by themselves, instead should be part of an FE block (i.e. RFC 1812 IP forwarder). A well-defined block has a well-defined packet processing behavior, and a well-defined set of state and parameters that CE can potentially configure or control via ForCES. Obviously, a namespace is needed to specify different blocks. The namespace assigns a unique ID or label to each distinct block type. For each block, it is necessary to specify the relevant information and parameters such as: Yang, et. al. Expires December 2002 [Page 5] Internet Draft ForCES FE Functional Model June 2002 - how many inputs it takes and what kinds of packets and meta data it takes for each input; - how many outputs it produces and what kind of packets and meta data it emits for each output and; - the packet processing (such as modification) behavior; - what information is programmed into it (e.g., LPM list, next hop list, WRED parameters, etc.) and what parameters among them are configurable; - what statistics it keeps (e.g., drop count, CRC error count, etc.); - what events it can throw (e.g., table miss, port down, etc.). CEs later use the information to decide how to configure the parameters, how to modify the relevant data structures (like tables), what kind of statistics it can query from FE, and what actions to take once a certain event happens. We use the classifier defined in [DS-MODEL] as an example. "Classifiers are 1:N(fan-out) devices: they take a single traffic stream as input and generate N logically separate traffic streams as output. Classifiers are parameterized by filters and output streams. Packets from the input stream are sorted into various output streams by filters which match the contents of the packet or possibly match other attributes associated with the packet." To further define filters: "A filter consists of a set of conditions on the component values of a packet's classification key (the header values, contents, and attributes relevant for classification)." Figure 1 illustrates an example classifier. Unclassified classified traffic traffic +------------+ | |--> match Filter1 --> OutputA ------->| classifier |--> match Filter2 --> OutputB | |--> no match --> OutputC +------------+ Figure 1. An Example Classifier 4.2.FE Block Library We expect a small set of blocks to be defined initially, e.g., ingress port, egress port, classifier, forwarder, meter, marker, shaper, scheduler, queue, encapsulator, decapsulator, encrypter, decrypter, NAT, mux, demux, header compressor, header decompressor, etc. Such a set of blocks can be viewed as a FE block library. The minimum set of FE functions required in [FORCES_REQ] must be part of this library. It is expected that new FE blocks would be defined and added into this library over time. It is also expected that the FE model itself be decoupled from the ForCES protocol so that extension Yang, et. al. Expires December 2002 [Page 6] Internet Draft ForCES FE Functional Model June 2002 to basic blocks or changes to the existing blocks should not impact the protocol itself. This document only intends to describe the conceptual FE model and illustrate it with some examples. However, it is not the intention of this document to define any specific block or the library itself. Separate document(s) would be written to achieve that. 4.3. FE Stage As a packet flows through an FE along a datapath, it flows through one or multiple distinct stages, with each stage instantiating a certain FE logical function. So an FE stage is simply an instance of a FE block within an FE's datapath. Each FE allocates a FE-unique stage ID or label to each of its stages and passes the stage ID or label along with the corresponding block name as part of the FE model description. This allows multiple instances of the same block present in a FE's datapath. For each stage, the following information must be defined: - its FE-unique stage ID or label to identify, configure and manage the logical block of each by CE.; - the corresponding block name (from the namespace) that this stage instantiating; - all the information for this stage as described in Section 5.1; - the number of downstream stages to which this stage can send packets; - for each downstream stage: its stage ID or label. 4.4. Directed Graph of FE Blocks Once a library of well-understood FE blocks is defined, and a stage is represented as described in 5.3, any static FE can be modeled by a directed graph interconnecting all the stages present in the FE. The static FEs can be defined by identifying all the input stage(s) and specifying each stage as described in Section 5.3. Figure 2 shows one simple example of such an FE with three stages. To model a static FE, the following information needs to be represented: - number of stages in the FE; - for each stage: define the stage using the model in 5.3; - identify the input stages It is frequently necessary to share state between FE functional stages in the forwarding plane. Since this model uses blocks as the basic way of modeling forwarding plane packet processing stages, the state information is exposed as part of connecting stage. For example in Figure 2, stage #2 (IPv4 L3 LPM Forwarder) generates some meta data at its output to carry information on which port the packets should go to, and #3 (Enet-Egress-port-Manager) uses this meta data to direct the packets to the right egress port. Yang, et. al. Expires December 2002 [Page 7] Internet Draft ForCES FE Functional Model June 2002 +------------+ +------------+ +------------+ input | Ethernet | | | | Ethernet |output ------->| Ingress |-->| IPv4 L3 LPM|-->| Egress |-----> | Port Mgr | | Forwarder | | Port Mgr | +------------+ +------------+ +------------+ {stage ID=1, {stage ID=2, {stage ID=3, function= function= function= Enet-Ing-port, IPv4-L3-LPM-fwd, Enet-Eg-port-Mgr, #downstream=1, #downstream=1, #downstream=1, downstream={2} downstream={3} downstream=none } } } Figure 2. A very simple example of a static FE. Queue1 +---+ +--+ | A|------------------->| |--+ +->| | | | | | | B|--+ +--+ +--+ +--+ | | +---+ | | | | | | | Meter1 +->| |-->| | | | | | | | | | +--+ +--+ | | Counter1 Absolute Queue2| +--+ +---+ | Dropper1 +--+ +--->|A | | A|---+ | |------>|B | -------->| B|------------------------------>| | +--->|C |------> | C|---+ +--+ | +->|D | | X|-+ | | | +--+ +---+ | | +---+ +---+ Queue3| | Scheduler Classifier1 | | | A|------------>|A | +--+ | | | +->| | | |->| |--+ | | | B|--+ +--+ +->|B | | | | | +---+ | | | | +---+ +--+ | | Meter2 +->| |-+ Mux1 | | | | | | +--+ Queue4 | | Marker1 +--+ | +---------------------------->| |----+ | | +--+ Figure 3. An FE example with multiple datapath. There might be more than one datapath within an FE, typically due to blocks that are either 1:N or N:1. Figure 3 shows one such FE block example (based on an example in [MD-MODEL]). This FE implements QoS functions via combination of one or multiple instances of logical functions like classifier, meter, marker, queue, scheduler, etc. Yang, et. al. Expires December 2002 [Page 8] Internet Draft ForCES FE Functional Model June 2002 Some of the functions are 1:N (fan-out_ functions. For example, the stage labeled "Classifier1" is 1:4 function with four downstream stages, namely, Meter1, Queue2, Meter2, Queue4. A packet entering this FE can potentially take one of the six distinct datapath. Note that our stage representation can encode largely arbitrary topologies of the stages. The only restrictions on topology relate to the source and sink nature of ingress and egress port functions respectively. For example, egress port functions must not have nay downstream stages whereas no other stage may refer to an ingress port function as one of its downstream stages. An FE block may contain zero, one or more ingress port stages. Similarly, an FE block may contain zero, one or more egress port stages. In another word, not every FE block has to contain any ingress port or egress port stages. For example, Figure 4 shows two FE blocks. Block #1 contains one ingress port function but no egress port function, while block #2 contains one egress port function but no ingress port function. It is possible to connect these two FE blocks together to achieve the complete ingress-to-egress packet processing function. This provides the flexibility to spread the functions across multiple FEs and interconnect them together later for certain applications. Figure 4 shows such an example. ------------------------------------------------------- | +---------+ +------------+ +---------+ | input| | | | | | output | ---+->| Ingress |-->|Header |-->|IPv4 |---------+--->--+ | | port | |Decompressor| |Forwarder| FE | | | +---------+ +------------+ +---------+ Block #1| | ------------------------------------------------------| V | +-----------------------<-----------------------------+ | | |----------------------------------------- V | +------------+ +----------+ | | input | | | | output | +->--+->|Header |-->| Egress |---------+--> | |Compressor | | port | FE | | +------------+ +----------+ Block #2| -----------------------------------------| Figure 4. An example of two different FE blocks connected together. 5. Model Representation A formal data definition language is needed to represent the FE model described in this document. The following is a list of some potential candidates. A suitable candidate needs to be chosen as the representation of FE model. However, we intend to leave this as an open issue and much debate is needed in the ForCES WG before a Yang, et. al. Expires December 2002 [Page 9] Internet Draft ForCES FE Functional Model June 2002 decision can be made. Therefore, we only provide the candidate list and some initial discussion here without drawing a conclusion yet. - XML (Extensible Markup Language) Schema - ASN.1 (Abstract Syntax Notation One) - SMI (Structure of Management Information) [RFC1155] - SPPI (Structure of Policy Provisioning Information) [RFC3159] - SMIng (Next Generation Structure of Management Information) - UML (Universal Modeling Language) XML has the advantage of being human readable with relatively little effort. However, it is less efficient and requires XML parsing functions in the CE and FE. Currently XML is not widely deployed and used in network elements. ASN.1 format is human readable, and widely used in network protocols. SMI is based on a subset of ASN.1 and used to define Management Information Base (MIB) for SNMP. SPPI is the adapted subset of SMI used to define Policy Information Base (PIB) for COPS. SMIng is currently being developed by SMIng working group at IETF and represents a superset of SMIv2 and SPPI. The objective of SMIng is to replace both SMIv2 and SPPI with a single, updated language as the data definition language for the monitoring, configuration, and provisioning of network devices. 6. Security Considerations The FE model just describes the representation and organization of data sets and attributes in the forwarding plane. There is no communication protocol associated defined as part of the FE model therefore we do not see any security threats in this model. 7. Intellectual Property Right The authors are not aware of any intellectual property right issues pertaining to this document. 8. IANA consideration If we are going to identify and name a new logical components we may need to standardize those components and the attributes. 9. Normative References [RFC1812] F. Baker, "Requirements for IP Version 4 Routers", RFC1812, June 1995. [RFC1155] M. Rose, et. al., "Structure and Identification of Management Informationfor TCP/IP-based Internets", May 1990. Yang, et. al. Expires December 2002 [Page 10] Internet Draft ForCES FE Functional Model June 2002 [RFC3159] K. McCloghrie, et. al., "Structure of Policy Provisioning Information (SPPI)", August 2001. 10. Informative References [FORCES-REQ] T. Anderson, et. al., "Requirements for Separation of IP Control and Forwarding", work in progress, April 2002, . [GOPAL-MODEL] R. Gopal, "Forwarding Element Model", work in progress, February 2002, . [ANDERSON-MODEL] T. Anderson, "ForCES Architectural Framework and FE Functional Model", work in progress, November 2001, . [DS-MODEL] Y. Bernet, et. al., "An Informal Management Model for Diffserv Routers", work in progress, February 2001, . 11. Acknowledgments This document draws heavily from the concepts presented in [GOPAL- MODEL], [ANDERSON-MODEL] and [DS-MODEL]. In addition to the authors of these documents, the authors would also like to thank the following individuals for their invaluable technical input: David Putzolu, Hormuzd Khosravi, Eric Johnson, David Durham, Andrzej Matejko. 12. Authors' Addresses Lily L. Yang Intel Labs 2111 NE 25th Avenue Hillsboro, OR 97124 USA Phone: +1 503 264 8813 Email: lily.l.yang@intel.com Joel Halpern P.O.Box 6049 Leesburg, VA 20178 Phone: +1 703 371 3043 Email: jmh@joelhalpern.com Ram Gopal Nokia Research Center 5, Wayside Road, Yang, et. al. Expires December 2002 [Page 11] Internet Draft ForCES FE Functional Model June 2002 Burlington, MA 01803 Phone: +1 781 993 3685 Email: ram.gopal@nokia.com Ram Dantu Netrake Corporation 3000 Technology Drive Plano, Texas 75074 Phone: +1 214 291 1111 Email: ramd@netrake.com 1. Abstract........................................................1 2. Motivation and Requirements of FE model.........................2 3. Capability and configuration Model versus State Model...........4 4. FE Model........................................................5 4.1. FE Blocks..................................................5 4.2. FE Block Library...........................................6 4.3. FE Stage...................................................7 4.4. Directed Graph of FE Blocks................................7 5. Model Representation............................................9 6. Security Considerations........................................10 7. Intellectual Property Right....................................10 8. IANA consideration.............................................10 9. Normative References...........................................10 10. Informative References........................................11 11. Acknowledgments...............................................11 12. Authors' Addresses............................................11 Yang, et. al. Expires December 2002 [Page 12]