DCCP WG G.Fairhurst Internet-Draft University of Aberdeen Intended status: Proposed Standard November 18, 2007 Expires: May 18, 2008 The DCCP Service Code draft-ietf-dccp-serv-codes-03.txt Status of this Memo By submitting this Internet-Draft, each author represents that any applicable patent or other IPR claims of which he or she is aware have been or will be disclosed, and any of which he or she becomes aware will be disclosed, in accordance with Section 6 of BCP 79. 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 This Internet-Draft will expire on May 18, 2008. Abstract This document describes the usage of Service Codes by the Datagram Congestion Control Protocol, RFC 4340. This document motivates the setting of Service Codes by applications. Service Codes provide a method to identify the intended service/application to process a DCCP connection request. This provides improved flexibility in the use and assignment of port numbers for connection multiplexing. The use of a DCCP Service Code can also enable more explicit coordination of services with middleboxes (e.g. network address translators and firewalls). It updates the description provided in RFC 4340. Fairhurst Expires May 18, 2008 [Page 1] Internet-Draft DCCP Service Codes November 2007 Table of Contents 1. Introduction...................................................3 1.1. History...................................................3 1.2. Conventions used in this document.........................6 2. An Architecture for Service Codes..............................6 2.1. IANA Port Numbers.........................................7 2.2. DCCP Service Code Values..................................8 2.3. Service Code Registry.....................................8 2.4. Zero Service Code.........................................9 2.5. Invalid Service Code......................................9 2.6. SDP for describing Service Codes..........................9 3. Use of the DCCP Service Code..................................10 3.1. Setting Service Codes at the Sender......................10 3.2. Using Service Codes in the Network.......................10 3.3. Using Service Codes at the Receiver......................11 3.3.1. Reception of a DCCP-Request.........................12 3.3.2. Multiple Associations of Service Codes..............13 3.3.3. Automatically launching a Server....................13 4. Benchmarking Services Described in this document..............14 4.1. Echo.....................................................14 4.2. Daytime..................................................14 4.3. Character generator......................................14 4.4. Time service.............................................15 4.5. Generic PerfTest service.................................15 4.6. PERF service.............................................15 5. Security Considerations.......................................16 5.1. Interactions of Service Codes and port numbers...........16 5.2. Interactions with IPsec..................................16 6. IANA Considerations...........................................17 6.1. Port number values allocated by this document............17 6.2. Service Code values allocated by this document...........18 7. Acknowledgments...............................................19 8. References....................................................19 8.1. Normative References.....................................19 8.2. Informative References...................................19 9. Author's Addresses............................................21 9.1. Intellectual Property Statement..........................22 9.2. Disclaimer of Validity...................................22 9.3. Copyright Statement......................................22 APPENDIX A: API support for Service Codes........................23 Fairhurst Expires May 18, 2008 [Page 2] Internet-Draft DCCP Service Codes November 2007 1. Introduction DCCP specifies a Service Code as a 4-byte value (32 bits) that describes the application-level service to which a client application wishes to connect ([RFC4340], Section 8.1.2). A Service Code identifies the protocol (or a standard profile, e.g. [ID.DCCP.RTP]) to be used at the application layer. It is not intended to be used to specify a variant of an application, or a specific variant of a protocol (section 2.2). Service Codes allow a flexible correspondence between application- layer services and port numbers, which affects how applications interact with DCCP. This decouples the use of ports for connection demultiplexing and state management from their use to indicate a desired service. An application identifies the requested service by the Service Code value in a DCCP-REQUEST. Each application may listen on one or more ports associated with one or more Service Codes ([RFC4340], 8.1.2). The use of Service Codes can assist in identifying the intended service when the server by a Middleboxes (a network address translator (NAT) [RFC2663], NAT-PT [RFC2766], Firewalls, etc). Middleboxes that desire to identify the type of data being transported by a flow, should utilize the Service Code for this purpose. When consistently used, the Service Code can provide a more specific indication of the actual service (e.g. indicating the type of multimedia flow, or intended application behaviour). The more flexible use of server ports can offer benefit to applications where servers need to handle very large numbers of simultaneous open ports to the same service. RFC 4340 omits to describe the motivation behind Service Codes, nor does it describe properly how well-known (server) ports relate to Service Codes. The intent of this document is to clarify these issues. 1.1. History It is simplest to understand the motivation for defining Service Codes by describing the history of the DCCP protocol. Most current Internet transport protocols (TCP [RFC793], UDP [RFC768], SCTP [RFC2960], UDP-Lite [RFC3828]) used "well-known" port numbers [RFC814]. These 16-bit values indicate the application service associated with a connection or message. The server port must Fairhurst Expires May 18, 2008 [Page 3] Internet-Draft DCCP Service Codes November 2007 be known to the client to allow a connection to be established. This may be achieved using out-of-band signaling (e.g. described using SDP [RFC4566]), but more commonly a well-known port is allocated to a particular protocol or application; for example HTTP commonly uses port 80 and SMTP commonly uses port 25. Making a port number well- known [RFC1122] involves registration with the Internet Assigned Numbers Authority (IANA), which includes defining a service by a unique keyword and reserving a port number from among a fixed pool [IANA]. This fixed space of port numbers is globally reserved [ID.Portnames]. In the earliest draft of DCCP the authors wanted to address the issue of well-known ports in a future-proof manner, since this method suffers from several problems: o The port space is not sufficiently large for ports to be easily allocated (e.g. in an unregulated manner). Thus, many applications operate using unregistered ports, possibly colliding with use by other applications. o The use of port-based firewalls encourages application-writers to disguise one application as another in an attempt to bypass firewall filter rules. This motivates firewall writers to use deep packet inspection in an attempt to identify the service associated with a port number. o ISPs often deploy transparent proxies, primarily to improve performance and reduce costs. For example, TCP requests destined to TCP port 80 are often redirected to a web proxy. These issues are coupled. When applications collide on the same "well-known", but unregistered port, there is no simple way for network security equipment to tell them apart, with the likelihood of introducing problems with interaction of features. There is little that a transport protocol designer can do about applications that attempt to masquerade as other applications. For ones that are not attempting to hide, the problem may be simply that they cannot trivially obtain a well-known port. Ideally, it should be sufficiently easy that every application-writer can request a well-known port and get one instantly with no questions asked. The 16-bit port space traditionally used is not large enough to support such a trivial allocation of well-known ports. Thus, the design of DCCP sought an alternative solution. The idea was simple. A 32-bit server port space should be sufficiently large that it enables use of very simple allocation policies. However, Fairhurst Expires May 18, 2008 [Page 4] Internet-Draft DCCP Service Codes November 2007 overhead considerations made a 32-bit port value undesirable (DCCP needed to be useful for low rate applications). The solution in DCCP to this problem was the use of a 32-bit Service Code [RFC4340] that is included only in the DCCP-Request packet. This was intended to perform the primary role of a well-known server port, in that it would be trivially simply to obtain a unique value for each application. Placing the value in a request packet, requires no additional overhead for the actual data flow. It is however sufficient for both the end systems, and provides any stateful middleboxe(s) along the path with additional information to understand what applications are being used. The original draft of the DCCP specification did not use traditional ports; instead the client allocated a 32-bit identifier to uniquely identify the connection. The server listened on a socket bound only to a Service Code. This solution was unambiguous; the Service Code was the only identifier for a listening socket at the server side. The DCCP client included a Service Code in the request, allowing it to reach the corresponding listening application. This design suffered from the downside of being sufficiently different from existing protocols that there were concerns that it would hinder the use of DCCP through NATs and other middleboxes. RFC 4340 abandoned the use of a 32-bit connection identifier in favor of two traditional 16-bit ports, one chosen by the server and one by the client. This allows middleboxes to utilize similar techniques for DCCP, UDP, TCP, etc. (e.g. NAT). This also has the advantage that two servers associated with the same Service Code could co-exist on the same server host. However, it introduced a new problem: "How does the server port relate to the Service Code?" The intent was that the Service Code identified the application or protocol using DCCP, providing middleboxes with information about the intended use of a connection, and that the pair of ports effectively formed a 32- bit connection identifier, which was unique between a pair of end- systems. The large number of available unique Service Code values allows all applications to be assigned a unique Service Code. However, there remains a current problem: The server port is chosen by the server, but the client needs to know this to establish a connection. It was undesirable to mandate out-of-band communication to discover the server port. A solution is to register well-known DCCP ports. The limited availability of well-known DCCP ports appears to contradict the benefits of DCCP Service Codes, because although it may be trivial to Fairhurst Expires May 18, 2008 [Page 5] Internet-Draft DCCP Service Codes November 2007 obtain a service code, it has not traditionally been trivial to obtain a well-known port from IANA and in the long-run it may not be possible to uniquely allocate a unique well-known DCCP port to new applications. As port numbers become scarce, this motivates the need to associate more than one Service Code with a listening port (e.g. two different applications could be assigned the same well-known port, and need to run on the same host at the same time). No protocols issues arise from a port being associated with two Service Codes, each bound to different applications does not raise any protocol issues. An incoming DCCP-Request is directed to the correct application. Service Codes provide flexibility in the way clients identify the server application to which they wish to communicate. The Service Code mechanism allows a server to associate a set of server ports with a service. The set may be common with other services available at the same server host, allowing a larger number of concurrent connections for a particular service than possible when the service is identified by a single well-known port number. There has been confusion concerning how well-known ports relate to well-known Service Codes. The goal of this document is to clarify the issues concerning the use and allocation of Service Codes. RFC4340 states that Service Codes are not intended to be DCCP- specific. Service Codes, or similar concepts may therefore also be useful to other IETF transport protocols. 1.2. 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 [RFC2119]. All protocol code points and values are transmitted in network byte order (most significant byte first), with the most significant bit of each byte is placed in the left-most position of an 8-bit field. 2. An Architecture for Service Codes DCCP defines the use of a combination of ports and Service Codes to identify the server application ([RFC4340], 8.1.2). These are described in the following sections. Fairhurst Expires May 18, 2008 [Page 6] Internet-Draft DCCP Service Codes November 2007 2.1. IANA Port Numbers In DCCP the packets belonging to a connection are de-multiplexed based on a combination of four values {source IP address, source port, dest IP address, dest port}, as in TCP. An endpoint address is associated with a port number, (e.g. forming a socket); and a pair of associations uniquely identifies each connection. Ports provide the fundamental per-packet de-multiplexing function. The Internet Assigned Numbers Authority currently manages the set of globally reserved port numbers [IANA]. The source port associated with a connection request, often known as the "ephemeral port", traditionally includes the range 49152-65535, and should also include the 1024-49151 range. The value used for the ephemeral port is usually chosen by the client operating system. It has been suggested that a randomized choice of port number value can help defend against "blind" attacks [ID.TSVWG.RAND] in TCP. This method may be applicable to other IETF-defined transport protocols, including DCCP. Traditionally, the destination (server) port value that is associated with a service is determined either by an operating system index to a copy of the IANA table (e.g., getportbyname() in Unix, which indexes the /etc/services file), or directly mapped by the application. The UDP and TCP port number space: 0..65535, is split into three ranges [RFC2780]: o 0..1023 "well-known", also called "system" ports o 1024..49151 "registered", also called "user" ports o 49152..65535 "dynamic", also called "private" ports DCCP supports well-known and reserved ports. These are allocated in the DCCP IANA port numbers registry ([RFC4340], 19.9). Each registered DCCP port MUST be associated with at least one pre-defined Service Code. Applications that do not need to use a server port in the well-known range SHOULD use a dynamic server port (i.e. do not require to be registered in the DCCP port registry). Clients can identify the server port value for the services to which they wish to connect using a range of methods. One common method is by reception of a SDP record (section 2.6) exchanged out-of-band (e.g. using SIP [RFC3261] or RTSP [RFC2326]). DNS SRV resource records also provide a way to identify a server port for a particular service based on the services string name [RFC2782]. Applications that do not use out-of-band signalling can still communicate, providing both client and server agree the port value to be used (e.g. by hashing the 32-bit Service Code to a value in the dynamic port range). Note that more Fairhurst Expires May 18, 2008 [Page 7] Internet-Draft DCCP Service Codes November 2007 than one DCCP server may share the same server port, since in DCCP the Service Code mechanism is the method for unique identification of a service. 2.2. DCCP Service Code Values DCCP specifies a 4 byte Service Code ([RFC4340],8.1.2) represented in one of three forms as: a decimal number (the canonical method), a four character ASCII string, or an eight digit hexadecimal number. The Service Code identifies the application-level service to which a client application wishes to connect. Examples of services are RTP, TIME, ECHO. In a different example, DTLS provides a transport-service (not an application-layer service), therefore applications using DTLS are individually identified by a set of corresponding service codes. A single passive listening port may be associated with more than one Service Code value, which may be associated with one or different server applications. Endpoints MUST associate a Service Code with every DCCP socket [RFC4340], both actively and passively opened. The application will generally supply this Service Code. It permits a more flexible correspondence between services and port numbers than possible using the corresponding socket pair (4-tuple of layer-3 addresses and layer-4 ports). This decouples the use of ports for connection demultiplexing and state management, from their use to indicate a desired endpoint service. The Service Code value is present only in DCCP-Request ([RFC4340],5.2)and DCCP-Response packets ([RFC4340],5.3). Applications/protocols that provide version negotiation or indication in the protocol operating over DCCP do not require a new server port for each new protocol version. New versions of such applications/protocols SHOULD continue to use the same Service Code. If the application developers feel that the new version provides significant new capabilities (e.g. that will change the behavior of middleboxes), they MAY allocate a new Service Code associated with the same or a different set of well-known ports. If the new Service Code is associated with well-known ports, the DCCP Well-Known Ports registry MUST also be updated to include the new Service Code value. Fairhurst Expires May 18, 2008 [Page 8] Internet-Draft DCCP Service Codes November 2007 2.3. Service Code Registry The set of registered Service Codes specified for use within the general Internet are defined in an IANA-controlled name space. IANA manages new allocations of Service Codes in this space ([RFC4340], 19.8, updated by this document). Private service codes are not centrally allocated and are denoted by the range 1056964608- 1073741823 (i.e. whose first hexadecimal digit has the ASCII value for '?'). Associations of Service Code with Well-Known Ports are defined in the IANA DCCP Port Registry (section 2.1). 2.4. Zero Service Code A Service Code of zero is "permanently reserved (it represents the absence of a meaningful Service Code)" [RFC4340]. This indicates that no application information was provided. RFC 4340 stated that applications MAY be associated with this Service Code in the same way as other Service Code values. This use is permitted for any server port. This document clarifies section 19.8 of RFC 4340: "Applications SHOULD NOT use a Service Code of zero. Application writers that need a temporary Service Code value SHOULD choose a value from the private range (Section 2.3). Applications intended for deployment in the Internet are encouraged to use an IANA-defined Service Code. If no specific Service Code exists, they SHOULD request a new assignment for their protocols from the IANA. 2.5. Invalid Service Code RFC4340 defines the Service Code value of 0xFFFFFFFF as Invalid. The Invalid Service Code is provided so implementations can use a special four-byte value to indicate "no valid service code". Implementations MUST NOT accept a DCCP-Connect with this value, and SHOULD NOT allow applications to bind to this Service Code value. 2.6. SDP for describing Service Codes Methods that currently signal destination port numbers, such as the Session Description Protocol (SDP) [RFC4566] require extension to support DCCP Service Codes [ID.DCCP.RTP]. Fairhurst Expires May 18, 2008 [Page 9] Internet-Draft DCCP Service Codes November 2007 3. Use of the DCCP Service Code The basic operation of Service Codes is as follows: o A sending host: . issues a DCCP-Request with a Service Code and chooses a destination port number that is expected to be associated with the specified Service Code at the destination. o A server that receives a DCCP-Request: . determines whether an available service matching the Service Code is supported for the specified destination server port. The session is associated with the Service Code and a corresponding server. A DCCP-Response is returned. . if the service is not available, the session is rejected and a DCCP-Reset packet is returned. This section explicitly updates RFC 4340 as follows: "A DCCP implementation SHOULD allow multiple applications using different DCCP Service Codes to listen on the same server port. A DCCP implementation SHOULD provide a method that informs a server of the Service Code value that was selected by an active connection." The remainder of this section describes processing of DCCP Service Codes at the sending and receiving hosts and within the network by middleboxes. 3.1. Setting Service Codes at the Sender A client application MUST associate every DCCP connection (and hence every DCCP active socket) with a single Service Code value [RFC4340]). This value is used in the corresponding DCCP-Request packet. 3.2. Using Service Codes in the Network Port numbers and IP addresses are the traditional methods to identify a flow within an IP network. When the DCCP header has not been encrypted, Middleboxes [RFC3234] SHOULD use the Service Code to identify the application-service (even when running on a non-standard port). When consistently used, the Service Code can provide a more specific indication of the actual service (e.g. indicate the type of Fairhurst Expires May 18, 2008 [Page 10] Internet-Draft DCCP Service Codes November 2007 multimedia flow, or intended application behaviour). Middlebox devices are therefore expected to check Service Code values as well as, or even instead of port numbers for DCCP. DCCP connections identified by the Service Code continue to use IP addresses and ports, although neither port number may be well- known/reserved. Network address and port translators, known collectively as NATs [RFC2663][RFC2766], not only interpret DCCP ports, but may also translate/modify them [RFC2993]. Interpreting DCCP Service Codes can reduce the need to correctly interpret port numbers, leading to new opportunities for network address and port translators. The DCCP Service Code may allow services to be identified behind NATs, if NATs are not further extended to translate Service Codes. Although Service Codes label a connection and can (and is encouraged to) associate specific delivery properties (e.g. use Service Codes to identify the real-time nature of a flow that claims to be using RTP), there is no guarantee that the actual connection data corresponds to the associated Service Code. A middlebox implementor may still therefore desire to use deep packet inspection, and other means, in an attempt to verify the content of a connection. The use of the DCCP Service Code can potentially lead to interactions with other protocols that interpret or modify DCCP port numbers [RFC3234]. The following recommendations are provided: o A middlebox SHOULD use the Service Code value to assist in determining the behaviour to be applied to a packet flow (e.g. default keep-alive interval, NAT translation, etc). o A middlebox SHOULD NOT modify the Service Code, unless they also change the service that a connection is accessing. o A middlebox MAY send a DCCP-Reset in response to a packet with a Service Code that is considered unsuitable. 3.3. Using Service Codes at the Receiver A Service Code is used by a host that receives a DCCP-Request to associate a DCCP connection with the corresponding application service. At the server, this association must be explicit, i.e. if the connection is accepted, the requested Service Code must have been previously associated with the listening port at the server. Fairhurst Expires May 18, 2008 [Page 11] Internet-Draft DCCP Service Codes November 2007 A number of options are presented for servers using passively listening sockets. As an example, consider the four cases that could arise when two DCCP server applications listen on the same host: o The simplest case is when the two servers are associated with different Service Codes and are bound to different server ports (section 3.3.1). o The two servers may be associated with the same DCCP Service Code value, but be bound to different server ports (section 3.3.1). o The two servers could use different DCCP Service Code values, and be bound to the same server port (section 3.3.2). o The two servers could attempt to use the same DCCP Service Code and bind to the same server port. A DCCP implementation MUST disallow this, since there is no way for the DCCP host to direct a new connection to the correct server application. RFC 4340 ([RFC4340, 8.1.2) states that an implementation: o MUST associate each active socket with exactly one Service Code on a specified server port. o MAY, at the discretion of an implementation, associate more than one Service Code with a passive socket. This document updates RFC4340 in the following way: o SHOULD allow more than one Service Code to be associated with a passive server port, enabling multiple applications, or multiple versions of an application, to listen on the same port, differentiated by Service Code. o MUST also allow a server to use a single Service Code for more than one server port. o SHOULD provide a method that informs a server of the Service Code value that was selected by an active connection. 3.3.1. Reception of a DCCP-Request When a DCCP-Request is received, and the specified destination port is not bound to a server, the host MUST reject the connection by issuing a DCCP-Reset with Reset Code "Connection Refused". A host MAY also use the Reset Code "Too Busy" ([RFC4340], 8.1.3). Fairhurst Expires May 18, 2008 [Page 12] Internet-Draft DCCP Service Codes November 2007 When the destination port is bound to a server, the host MUST also verify that the server port has been associated with the specified Service Code. Two cases can occur: o If the receiving host is listening on the specified server port and the DCCP-Request uses one of the Service Codes associated with the server port, the host accepts the connection. Once connected, the server returns a copy of the Service Code in the DCCP-Response packet completing the initial handshake [RFC4340]. o If the server port is not associated with the requested Service Code, the server MUST reject the request by sending a DCCP-Reset packet with Reset Code 8, "Bad Service Code" ([RFC4340], 8.1.2). After a connection has been accepted, the protocol control block is associated with the pair of ports and the pair of IP addresses and a single Service Code value. 3.3.2. Multiple Associations of Service Codes and Ports at the Server RFC4340 states that a single passively opened (listening) port MAY be associated with multiple Service Codes, although an active (open) connection can only be associated with a single Service Code. This document updates RFC4340 to add: "A Service Code MAY be associated with more than one destination port (corresponding to a specified set of server port values)." A single application may wish to accept connections for more than one Service Code using the same server port. This approach can simplify middlebox processing, e.g. it should not be necessary to create more than one hole in a firewall for this to be the case; for example DTLS connections and unencrypted connections for the same application will normally use different Service Codes to distinguish them, but because this is the same application, it makes sense to use the same port. 3.3.3. Automatically launching a Server A host implementation may permit a service to be associated with a server port (or range of ports) that is not permanently running at the Server. In this case, the arrival of a DCCP-Request may require a method to associate a DCCP-Request with a server that handles the corresponding Service Code. This operation could resemble that of "inetd" [inetd]. This may allow a server to offer more than the limit of 65,536 services determined by the size of the Port field (fewer if system/user/dynamic boundaries are preserved). The upper limit is Fairhurst Expires May 18, 2008 [Page 13] Internet-Draft DCCP Service Codes November 2007 based solely on the number of unique connections between two hosts (i.e., 4,294,967,296). As in the previous section, when the specified Service Code is not associated with the specified server port, the connection MUST be aborted and a DCCP Reset message sent [RFC4340]. 4. Benchmarking Services Described in this document A number of simple services are commonly supported by systems using DCCP and UDP, this section defines corresponding services for DCCP. These services are useful to debug and benchmark bidirectional DCCP connections. The IANA section of this document allocates a corresponding set of code points for these services. 4.1. Echo The operation of the DCCP echo service follows that specified for UDP [RFC862]: a server listens for DCCP connections; once a client has set up a connection, each data packet sent to the server will be copied (echoed) back to the client. 4.2. Daytime The DCCP daytime service is operationally equivalent to the connection-based TCP daytime service [RFC867]: any data received is discarded by the server; and generates a response sent in a DCCP data packet containing the current time and data as an ASCII string; after which the connection is closed. 4.3. Character generator The operation of the DCCP chargen service corresponds to the connection-based TCP chargen protocol [RFC864]: A server listens for incoming requests and, once a client has established a connection, continuously sends datagrams containing a random number (between 0 and 512, up to the current Path MTU) of characters. The service terminates when the user either closes or aborts the connection. Congestion control is enforced using the mechanisms [RFC4340] and related documents. If necessary the receiver can enforce flow control on this service by using either or both of the Slow Receiver ([RFC4340], 11.6) and Data Dropped ([RFC4340], 11.7) DCCP options to signal the server to slow- down. Fairhurst Expires May 18, 2008 [Page 14] Internet-Draft DCCP Service Codes November 2007 The chargen protocol provides a useful service that may be used for testing and measurement of bidirectional DCCP connectivity, as well as congestion control responsiveness. The datagram-based variant of chargen can be emulated with the DCCP ECHO service by changing the format of the datagrams sent by the client, hence these services complement each other. 4.4. Time service The format of timestamps and the operation of the DCCP time service is equivalent with the TCP time protocol variant [RFC868]: a server listens for incoming connections; after a client has established a new connection, the server sends a 4-byte timestamp; whereupon the client closes the connection. 4.5. Generic PerfTest service The PerfTest service specified by this document provides a generic service that may be used to benchmark and measure both unidirectional and bidirectional DCCP connections, as well as server and host DCCP stacks. These services are identified by the Service Code "XPER". This document does not specify a specific port number for this service. The payload of DCCP packets associated with this service does not have a specified format. They are silently discarded by the receiver, and used only for gathering numerical performance data. Tools that have specific payload formats should register their own Service Code value with IANA (e.g. section 4.6). This Service Code is for benchmarking applications that transmit data in one-direction only. A benchmarking application expects responses to the messages it sends requires a different Service Code. (This could result in different Middlebox treatment.) 4.6. PERF service The PERF service specified by this document describes the service supported by the open-source iperf benchmarking program [iperf]. This may be used to benchmark and measure both unidirectional and bidirectional DCCP connections, as well as server and host DCCP stacks. This service is identified by a Service Code "PERF" and is associated with a well-known port number that currently coincides with that used by the iperf benchmarking program [iperf]. Fairhurst Expires May 18, 2008 [Page 15] Internet-Draft DCCP Service Codes November 2007 5. Security Considerations This document does not describe new protocol functions. The document discusses the usage of Service Codes. There are four areas of security that are important: 1. Interaction with NATs and firewalls (section 3.2 describes middlebox behaviour). 2. Interpretation of DCCP Service Codes over-riding traditional use of reserved/well-known port numbers (section 5.1) 3. Interaction with IPsec and DTLS security (section 5.2). 4. Services used for benchmarking and testing may also be used to generate traffic for other purposes, and also pose an opportunity for a Denial of Service attack. Care needs to be exercised when enabling these services in an operational network, or appropriate rate-limits should be provided to mitigate these effects. 5.1. Interactions of Service Codes and port numbers The same service (application) may be accessed using more than one Service Code. Examples include the use of separate Service Codes for an application layered directly upon DCCP and one using DTLS transport over DCCP. Other possibilities include the use of a private Service Code that maps to the same application as assigned to an IANA-defined Service Code value, or a single application that provides more than one service. Different versions of a service (application) may also be mapped to a corresponding set of Service Code values. Care needs to be exercised when interpreting the mapping of a Service Code value to the corresponding service. Processing of Service Codes may imply more processing than currently associated with incoming port numbers. Implementers need to guard against increasing opportunities for Denial of Service attack. 5.2. Interactions with IPsec IPsec uses port numbers to perform access control in transport mode [RFC4301]. Security policies can define port-specific access control (PROTECT, BYPASS, DISCARD), as well as port-specific algorithms and keys. Similarly, firewall policies allow or block traffic based on port numbers. Fairhurst Expires May 18, 2008 [Page 16] Internet-Draft DCCP Service Codes November 2007 Use of port numbers in IPsec selectors and firewalls may assume that the numbers correspond to well-known services. It is useful to note that there is no such requirement; any service may run on any port, subject to mutual agreement between the endpoint hosts. Use of the Service Code may interfere with this assumption both within IPsec and in other firewall systems, but it does not add a new vulnerability. New implementations of IPsec and firewall systems may interpret the Service Code when implementing policy rules, but should not rely on either port numbers or Service Codes to indicate a specific service. This is not an issue for IPsec because the entire DCCP header and payload are protected by all IPsec modes. None of the DCCP header is protected by application-layer security, e.g., DTLS [ID.DTLS.DCCP], so again this is not an issue [RFC4347]. 6. IANA Considerations A set of new services are defined in section 6 and are summarized in this section. >>> Author Note: This section requires consideration by the IANA and the DCCP WG - - issues need to be identified. [XX To encourage application writers to register their applications, and to avoid restricting DCCP service codes to a 16-bit space, we revise RFC 4340 as follows: "IANA should allocate well-known DCCP ports on demand to anyone to applies, without requiring a specification or additional justification. Each well-known port request MUST be for a specific registered DCCP Service Code. The procedure may allow both to be assigned in the same request. IANA MUST use an allocation policy that attempts to minimize server port collisions, but it is expected that the same well-known port will sometimes be allocated to more than one Service Code." XX] 6.1. Port number values allocated by this document IANA action is required to assign ports for use by DCCP. This document requests allocation of the following code points from the IANA DCCP Port numbers registry: >>>>>> IANA ACTION Please replace IANA THIS RFC, with the allocated RFC number. <<< Fairhurst Expires May 18, 2008 [Page 17] Internet-Draft DCCP Service Codes November 2007 echo 7/dccp Echo SC:ECHO # IETF dccp WG, [IANA - THIS RFC] daytime 13/dccp DayTime SC:DTIM # IETF dccp WG, [IANA - THIS RFC] chatgen 19/dccp Chargen SC:CHAR # IETF dccp WG, [IANA - THIS RFC] time 37/dccp Timeserver SC:TIME # IETF dccp WG, [IANA - THIS RFC] perf 5001/dccp iPerf SC:PERF # IETF dccp WG, [IANA - THIS RFC] 6.2. Service Code values allocated by this document This document solicits IANA action to allocate the following code points from the Service Code registry [IANA-SC]. The requested assignments are listed below and summarized in table 1. This set of Service Codes may be utilized for testing DCCP implementations and transmission paths. >>> IANA Please replace tbd by the assigned a port number in section 6.1. +----------+------+----+-------------------------------+----------+ | Service | ASCII|Port| Description | Ref | | Code (SC)| Code | | | | +----------+------+----+-------------------------------+----------+ |1162037327| ECHO | 7| Echo service | [RFC862] | |0x4543484f| | | | | |1146374477| DTIM | 13| Daytime server | [RFC867] | |0x4454494d| | | | | |1128808786| CHAR | 19| Character generator (chargen) | [RFC864] | |0x43484152| | | | | |1414090053| TIME | 37| Timeserver | [RFC868] | |0x54494d45| | | | | |1346720326| PERF |5001| iPerf | [*] | |0x50455246| | | | | |1481655634| XPER | - | Generic Performance Service | [*] | |0x58504552| | | | | +----------+------+----+-------------------------------+----------+ Table 1: Allocation of Service Codes by this document. Notes: 1) Port is the default port associated with this service. 2) * Reference is this document. Fairhurst Expires May 18, 2008 [Page 18] Internet-Draft DCCP Service Codes November 2007 This document notes that it is NOT required to supply an approved document (e.g. a published RFC) to support an application for a DCCP Service Code or port number value, although RFCs may be used to request Service Code values via the IANA Considerations section (e.g. [ID.DTLS.DCCP], [ID.DCCP.RTP]). 7. Acknowledgments This work has been supported by the EC IST SatSix Project. Significant contributions to this document resulted from discussion with Joe Touch, and this is gratefully acknowledged. The author also thanks Ian McDonald, Fernando Gont, Eddie Kohler and the DCCP WG for helpful comments on this topic, and Gerrit Renker for his help in determining DCCP behaviour, review of the document, and compilation of useful test applications defined in the IANA section of this document. Mark Handley provided significant input to the text on definition of Service Codes and their usage. He also contributed much of the material that has formed the historical background section. 8. References 8.1. Normative References [RFC1122] Braden, R. (ed.), "Requirements for Internet Hosts: Communication Layers, " STD 3, RFC 1122, Oct. 1989 (STANDARD). [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997 (BEST CURRENT PRACTICE). [RFC4340] Kohler, E., M. Handley, S. Floyd, "Datagram Congestion Control Protocol (DCCP)", RFC 4340, Mar. 2006 (PROPOSED STANDARD). 8.2. Informative References [IANA] Internet Assigned Numbers Authority, www.iana.org [IANA-SC] IANA DCCP Service Code Registry http://www.iana.org/assignments/service-codes [ID.Portnames] J. Touch, "A TCP Option for Port Names", IETF Work in Progress, draft-touch-tcp-portnames-00.txt. Fairhurst Expires May 18, 2008 [Page 19] Internet-Draft DCCP Service Codes November 2007 [ID.DTLS.DCCP] T.Phelan, "Datagram Transport Layer Security (DTLS) over the Datagram Congestion Control Protocol (DCCP)", IETF Work in Progress, draft-phelan-dccp-dtls-xx.txt. [ID.DCCP.RTP] C. Perkins, "RTP and the Datagram Congestion Control Protocol (DCCP)", IETF Work in Progress, draft-ietf-dccp- rtp-xx.txt. [ID.TSVWG.RAND] M. Larsen, F. Gont, "Port Randomization", IETF Work in Progress, draft-larsen-tsvwg-port-randomization-00. [inetd] The extended intetd project, http://xinetd.org/ [iperf] http://dast.nlanr.net/Projects/Iperf/ [RFC768] Postel, J., "User Datagram Protocol", STD 6, RFC 768, August 1980. [RFC793] Postel, J., "Transmission Control Protocol", STD 7, RFC 793, Sept. 1981 (STANDARD). [RFC814] Clark, D., "NAME, ADDRESSES, PORTS, AND ROUTES", RFC 814, July 1982 (UNKNOWN). [RFC862] Postel, J., "Echo Protocol", STD 20, RFC 862, May 1983. [RFC864] Postel, J., "Character Generator Protocol", STD 22, RFC 864, May 1983. [RFC867] Postel, J., "Daytime Protocol", STD 25, RFC 867, May 1983. [RFC868] Postel, J. and K. Harrenstien, "Time Protocol", STD 26, RFC 868, May 1983. [RFC2326] Schulzrinne, H., Rao, A., and R. Lanphier, "Real Time Streaming Protocol (RTSP)", RFC 2326, April 1998. [RFC2663] Srisuresh, P. and M. Holdrege, "IP Network Address Translator (NAT) Terminology and Considerations", RFC 2663, August 1999. [RFC2766] Tsirtsis, G. and P. Srisuresh, "Network Address Translation - Protocol Translation (NAT-PT)", RFC 2766, February 2000. Fairhurst Expires May 18, 2008 [Page 20] Internet-Draft DCCP Service Codes November 2007 [RFC2780] Bradner, S. and V. Paxson, "IANA Allocation Guidelines For Values In the Internet Protocol and Related Headers", BCP 37, RFC 2780, March 2000. [RFC2782] Gulbrandsen, A., P. Vixie, L. Esibov, "A DNS RR for specifying the location of services (DNS SRV)," RFC 2782, February 2000. [RFC2960] Stewart, R., Xie, Q., Morneault, K., Sharp, C., Schwarzbauer, H., Taylor, T., Rytina, I., Kalla, M., Zhang, L., and V. Paxson, "Stream Control Transmission Protocol", RFC 2960, October 2000. [RFC2993] Hain, T., "Architectural Implications of NAT", RFC 2993, November 2000. [RFC3234] Carpenter, B. and S. Brim, "Middleboxes: Taxonomy and Issues", RFC 3234, February 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. [RFC3493] Gilligan, R., Thomson, S., Bound, J., McCann, J., and W. Stevens, "Basic Socket Interface Extensions for IPv6", RFC 3493, February 2003. [RFC3828] Larzon, L-A., Degermark, M., Pink, S., Jonsson, L-E., and G. Fairhurst, "The Lightweight User Datagram Protocol (UDP- Lite)", RFC 3828, July 2004. [RFC4301] Kent, S. and K. Seo, "Security Architecture for the Internet Protocol", RFC 4301, December 2005. [RFC4347] Dierks, T. and E. Rescorla, "The Transport Layer Security (TLS) Protocol Version 1.1", RFC 4346, April 2006. [RFC4566] Handley, M., Jacobson, V., and C. Perkins, "SDP: Session Description Protocol", RFC 4566, July 2006. 9. Author's Addresses Godred (Gorry) Fairhurst, School of Engineering, University of Aberdeen, Kings College, Aberdeen, AB24 3UE, UK Email: gorry@erg.abdn.ac.uk URL: http://www.erg.abdn.ac.uk/users/gorry Fairhurst Expires May 18, 2008 [Page 21] Internet-Draft DCCP Service Codes November 2007 9.1. Intellectual Property Statement 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 ietf-ipr@ietf.org. 9.2. Disclaimer of Validity 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. 9.3. 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. Fairhurst Expires May 18, 2008 [Page 22] Internet-Draft DCCP Service Codes November 2007 APPENDIX A: API support for Service Codes A potential issue in defining an API for DCCP arises when an application binds to a port it needs to specify the associated DCCP Service Code. This requires an API that allows a service to be associated with a Service Code in addition to a port number. One approach is to use separate commands as follows: o Extend the existing port number indicator command (e.g., Unix bind() or connect() calls) to also select a specific Service Code where desired. o Extend the existing socket parameterization command (e.g., Unix setsockopt()) to set a service-code option. This is implemented in the present Linux API for a DCCP socket (where the Service Code should be wrapped by htonl/ntohl to ensure network byte order). o An information base (table) may be used by servers to identify the set of Service Codes that are associated with each port and the corresponding set of server applications. The current socket API generally requires separate requests to bind the port and to set the Service Code for the socket. This is not a problem, providing that an implementation requires both to be specified before the socket is allowed to accept connections. The host API SHOULD provide a method that returns the Service code of an incoming connection request to the application. This may be used by an application to correctly process a connection that arrives at a port for which it has registered more than one Service Code. >>> Author note: May need to discuss: get_port_and_service_code_by_name(char *what_service_do_you_want) char *get_service_code_by_number(unsigned sc) and interactions with getadddrinfo() address/port lookup routine, which has been introduced to simplify the migration to IPv6 ([RFC3493], 6.1). Functions such as getnameinfo and getservent may also need to be updated. >>> End Author Note. Fairhurst Expires May 18, 2008 [Page 23] Internet-Draft DCCP Service Codes November 2007 >>> RFC Editor please remove this section prior to publication. Change Log. 01 introduced: - a replacement of the word *range* when referring to sets of dccp ports (they are not necessarily contiguous), noted by E. Kohler. - Addition of some Service Codes in IANA section. 02 introduced: - add the use of profiles with DCCP, identified by Service Code, but not the use of protocol variants. - further detail on implementation levels (more input would be good) - added security consideration for traffic generators - added ref to UDPL for completeness - Corrected NiTs found by Gerrit Renker +++++++++++++++++++++++++++ WG 00 (first WG version) This introduced revisions to make it a WG document. - Corrected language and responded to many helpful comments from Fernando Gont and Ian McDonald. - Added a test for which server behaviour is used. - Added some speculative text on how to implement the SC. - More input and discussion is requested from the WG. - Added an informative appendix on host configuration. - Merging of some sections to remove repetition and clarify wording. +++++++++++++++++++++++++++ Fairhurst Expires May 18, 2008 [Page 24] Internet-Draft DCCP Service Codes November 2007 WG 01 Historical material was added. Comments from the list have been included. The concept of adding weak semantics to a SC=0 was removed. This was added at the request of implementers, with the aim of offering easier implementation on at least one target platform. It has been removed in this document because it weakens interoperability and complicates the Spec. The proposal to allow several levels of support was introduced in previous drafts following suggestions from the WG, but was removed in this revision. The method was seen to introduce complexity, and resulted in complex interoperability scenarios. Removed "test" method, this was no longer required. Draft was reorganized to improve clarity and simplify concepts. ---- WG 02 Updated following comments from Eddie Kohler. ---- WG 03 Fixed NiTs and addressed issues marked in previous version. Added 2 para at end of port section saying how to use well-known ports and that you do not need to register them. To-Do: Check RFC4340 differences. Section 6 - - Update to IANA section needs to be determined. Text on the API needs to be added. Could specify a hash algorithm - - if useful for SC->sport ---- Fairhurst Expires May 18, 2008 [Page 25]