RFC 1006
ISO Transport Service on top of the TCP Version: 3
ISO Transport Service on top of the TCP ✨
Okay bestie, this one is RFC 1006 🦋, written by Marshall T. Rose and Dwight E. Cass from the Northrop Research and Technology Center. It's from the Network Working Group, dated May 1987, and it's on Version: 3 💖. It obsoletes RFC 983, the older memo this one replaces 🌸.
Status of this Memo
So this memo is a standard for the Internet community 🎀. Hosts on the Internet that choose to implement ISO transport services on top of the TCP are expected to adopt and implement this standard. TCP port 102 is reserved for hosts which implement this standard. Distribution of this memo is unlimited 🌈.
This memo specifies version 3 of the protocol and supersedes [RFC983]. The changes between the protocol in Request for Comments 983 and this memo are minor, but they're unfortunately incompatible 😭.
1. Introduction and Philosophy ✨
The Internet community already has a super mature set of transport and internetwork protocols (TCP/IP), and they're honestly so good at giving end-users network and transport services 💅. Meanwhile, the CCITT and the ISO have defined a bunch of session, presentation, and application recommendations that the international community and loads of vendors have adopted 📚. The goal is to offer these higher layers directly in the ARPA Internet, as much as possible, without disrupting the facilities that already exist. That lets users build expertise with ISO and CCITT applications that weren't previously available in the ARPA Internet 🧚. It also gives us a smoother convergence and transition from TCP/IP-based networks to ISO-based networks in the medium and long term 🫶.
There are two basic ways to "port" an ISO or CCITT application to a TCP/IP environment 🍓. The first option is to port each application separately and develop local protocols on top of TCP. That's useful in the short term, since special-purpose interfaces to TCP can be built quickly, but it lacks generality 🙈.
The second option is based on the fact that both the ARPA Internet protocol suite and the ISO protocol suite are layered systems (though the former uses layering from a more pragmatic perspective) 🧩. A key part of the layering principle is layer-independence. This section is redundant for most readers, but we need a little background to introduce the concept 💜.
Externally, a layer is defined by two definitions:
- a service-offered definition, which describes the services provided by the layer and the interfaces it provides to access those services; and,
- a service-required definition, which describes the services used by the layer and the interfaces it uses to access those services.
Collectively, all of the entities in the network which co-operate to provide the service are known as the service-provider. Individually, each of these entities is known as a service-peer 👑.
Internally, a layer is defined by one definition:
- a protocol definition, which describes the rules which each service-peer uses when communicating with other service-peers.
Putting all this together, the service-provider uses the protocol and services from the layer below to offer its service to the layer above. Protocol verification, for instance, deals with proving that this actually happens (and it's also a fertile field for many Ph.D. dissertations in computer science) 🎓.
The concept of layer-independence is quite simply this:
IF one preserves the services offered by the service-provider
THEN the service-user is completely naive with respect to the
protocol which the service-peers useFor the purposes of this memo, we'll use layer-independence to define a Transport Service Access Point (TSAP) that appears to be identical to the services and interfaces offered by the ISO/CCITT TSAP (as defined in [ISO8072]). But we'll actually implement the ISO TP0 protocol on top of TCP/IP (as defined in [RFC793,RFC791]), not on top of the ISO/CCITT network protocol 🦋. Since the transport class 0 protocol is used over the TCP/IP connection, it achieves identical functionality to transport class 4. So ISO/CCITT higher level layers (all session, presentation, and application entities) can operate fully without knowing they're running on a TCP/IP internetwork 💖.
2. Motivation 💌
When migrating from TCP/IP to the ISO protocols, there are several strategies one might undertake. This memo was written with one particular strategy in mind 🎯.
The migration strategy this memo uses is based on gatewaying between the TCP/IP and ISO protocol suites at the transport layer. There are two strong arguments for this approach:
- Experience teaches us that it takes just as long to get good implementations of the lower level protocols as it takes to get implementations of the higher level ones. In particular, it has been observed that there is still a lot of work being done at the ISO network and transport layers. As a result, implementations of protocols above these layers are not being aggressively pursued. Thus, something must be done "now" to provide a medium in which the higher level protocols can be developed. Since TCP/IP is mature and essentially provides identical functionality, it's an ideal medium to support this development 🌸.
- Implementing gateways at the IP and ISO IP layers is probably not of general use in the long term. In effect, this would require each Internet host to support both TP4 and TCP. So a better strategy is to implement a graceful migration path from TCP/IP to ISO protocols for the ARPA Internet, once the ISO protocols have matured sufficiently 🙈.
Both of these arguments indicate that gatewaying should occur at or above the transport layer service access point. Further, the first argument suggests that the best approach is to perform the gatewaying exactly AT the transport service access point, to maximize the number of ISO layers which can be developed 🔥.
NOTE: This memo does not intend to act as a migration or intercept document. It is intended ONLY to meet the needs discussed above. However, it would not be unexpected that the protocol described in this memo might form part of an overall transition plan. The description of such a plan, however, is COMPLETELY beyond the scope of this memo. 💀
Finally, in general, building gateways between other layers in the TCP/IP and ISO protocol suites is problematic, at best.
To summarize: the primary motivation for the standard described in this memo is to make it easier to gain experience with higher-level ISO protocols (session, presentation, and application). The stability and maturity of TCP/IP are ideal for providing solid transport services, independent of the actual implementation 🎀.
3. The Model 🧠
The [ISO8072] standard describes the ISO transport service definition, which from here on is called TP.
ASIDE: This memo references the ISO specifications rather than the CCITT recommendations. The differences between these parallel standards are quite small, and can be ignored, with respect to this memo, without loss of generality. For the relationships, here's the breakdown 🍓:
Transport service [ISO8072] [X.214]
Transport protocol [ISO8073] [X.224]
Session protocol [ISO8327] [X.225]The ISO transport service definition describes the services offered by the TS-provider (transport service) and the interfaces used to access those services. This memo focuses on how the ARPA Transmission Control Protocol (TCP) [RFC793] can be used to offer those services and provide those interfaces ✨.
+-----------+ +-----------+
| TS-user | | TS-user |
+-----------+ +-----------+
| |
| TSAP interface TSAP interface |
| [ISO8072] |
| |
+----------+ ISO Transport Services on the TCP +----------+
| client |-----------------------------------------| server |
+----------+ (this memo) +----------+
| |
| TCP interface TCP interface |
| [RFC793] |
| |For expository purposes, the following abbreviations are used:
- TS-peer: a process which implements the protocol described by this memo
- TS-user: a process talking using the services of a TS-peer
- TS-provider: the black-box entity implementing the protocol described by this memo
For the purposes of this memo, which describes version 2 of the TSAP protocol, all aspects of [ISO8072] are supported with one exception:
- Quality of Service parameters
In the spirit of CCITT, this is left "for further study" 🔮. A future version of the protocol will most likely support the QOS parameters for TP by mapping them onto various TCP parameters.
The ISO standards don't specify the format of a session port (termed a TSAP ID). This memo mandates the use of the GOSIP specification [GOSIP86] for the interpretation of this field. (Please refer to Section 5.2, entitled "UPPER LAYERS ADDRESSING".) 🧚
Finally, the ISO TSAP is fundamentally symmetric in behavior. There's no underlying client/server model. Instead of a server listening on a well-known port, when a connection is established the TS-provider generates an INDICATION event, which the TS-user presumably catches and acts upon 🦋. Although this might be implemented by having a server "listen" by hanging on the INDICATION event, from the perspective of the ISO TSAP, all TS-users just sit around in the IDLE state until they either generate a REQUEST or accept an INDICATION 💅.
4. The Primitives ✨
okay so this protocol just assumes TCP[RFC793] already gives it these service primitives 💖:
Events
connected - open succeeded (either ACTIVE or PASSIVE)
connect fails - ACTIVE open failed
data ready - data can be read from the connection
errored - the connection has errored and is now closed
closed - an orderly disconnection has started
Actions
listen on port - PASSIVE open on the given port
open port - ACTIVE open to the given port
read data - data is read from the connection
send data - data is sent on the connection
close - the connection is closed (pending data is
sent)the memo then shows how to use those TCP services to fake the primitives that [ISO8073] wants 🧚. bestie, this part is giving pure illusion energy 🦋:
Events
N-CONNECT.INDICATION
- An NS-user (responder) is notified that
connection establishment is in progress
N-CONNECT.CONFIRMATION
- An NS-user (responder) is notified that
the connection has been established
N-DATA.INDICATION
- An NS-user is notified that data can be
read from the connection
N-DISCONNECT.INDICATION
- An NS-user is notified that the connection
is closed
Actions
N-CONNECT.REQUEST
- An NS-user (initiator) indicates that it
wants to establish a connection
N-CONNECT.RESPONSE
- An NS-user (responder) indicates that it
will honor the request
N-DATA.REQUEST - An NS-user sends data
N-DISCONNECT.REQUEST
- An NS-user indicates that the connection
is to be closedand then there's the TS-user list, as defined in [ISO8072]. this is the set the protocol actually hands to the TS-user 💅:
Events
T-CONNECT.INDICATION
- a TS-user (responder) is notified that
connection establishment is in progress
T-CONNECT.CONFIRMATION
- a TS-user (initiator) is notified that the
connection has been established
T-DATA.INDICATION
- a TS-user is notified that data can be read
from the connection
T-EXPEDITED DATA.INDICATION
- a TS-user is notified that "expedited" data
can be read from the connection
T-DISCONNECT.INDICATION
- a TS-user is notified that the connection
is closed
Actions
T-CONNECT.REQUEST
- a TS-user (initiator) indicates that it
wants to establish a connection
T-CONNECT.RESPONSE
- a TS-user (responder) indicates that it
will honor the request
T-DATA.REQUEST - a TS-user sends data
T-EXPEDITED DATA.REQUEST
- a TS-user sends "expedited" data
T-DISCONNECT.REQUEST
- a TS-user indicates that the connection
is to be closed5. The Protocol 💅
the protocol in this memo is basically the same as the protocol for ISO transport class 0 🌈. the differences are these:
- for testing purposes, initial data may be exchanged during connection establishment
- for testing purposes, an expedited data service is supported
- for performance reasons, a much larger TSDU size is supported
- the network service used by the protocol is provided by the TCP
ISO transport swaps info between peers in chunks called transport protocol data units (TPDUs) 🎀. this protocol wraps those TPDUs up in chunks called TPKTs. the structure of TPKTs and how they relate to TPDUs get covered in the next section 🫶.
PRIMITIVES 🎀
the TCP service primitives map onto the ones transport class 0 expects, and this part is honestly so straightforward 😭:
network service TCP
--------------- ---
CONNECTION ESTABLISHMENT
N-CONNECT.REQUEST open completes
N-CONNECT.INDICATION listen (PASSIVE open)
finishes
N-CONNECT.RESPONSE listen completes
N-CONNECT.CONFIRMATION open (ACTIVE open)
finishes
DATA TRANSFER
N-DATA.REQUEST send data
N-DATA.INDICATION data ready followed by
read data
CONNECTION RELEASE
N-DISCONNECT.REQUEST close
N-DISCONNECT.INDICATION connection closes or
errorsmapping the parameters is also straightforward 🍓:
network service TCP
--------------- ---
CONNECTION RELEASE
Called address server's IP address
(4 octets)
Calling address client's IP address
(4 octets)
all others ignored
DATA TRANSFER
NS-user data (NSDU) data
CONNECTION RELEASE
all parameters ignoredCONNECTION ESTABLISHMENT 🦋
the connection establishment procedures are the same as the ones in [ISO8073], with three exceptions ✨.
first, for testing, the connection request and connection confirmation TPDUs can carry initial user data in their user data fields 🧪.
second, to experiment with expedited data, the connection request and connection confirmation TPDUs can negotiate expedited data transfer. that uses the negotiation mechanism from [ISO8073], like setting the "use of transport expedited data transfer service" bit in the "Additional Option Selection" variable part. the default is not to use the transport expedited data transfer service 🙈.
third, for good performance, the default TPDU size is 65531 octets instead of 128 octets 🚀. if you want a smaller, standard TPDU size, you use the negotiation mechanism from [ISO8073] again, like setting the desired bit in the "TPDU Size" variable part 💜.
to perform an N-CONNECT.REQUEST action, the TS-peer does an active open to the desired IP address on TCP port 102. when TCP signals success or failure, that results in an N-CONNECT.INDICATION action 🌸.
to wait for an N-CONNECT.INDICATION event, a server listens on TCP port 102. when a client successfully connects to that port, the event happens and an implicit N-CONNECT.RESPONSE action is performed 🧚.
NOTE: In most implementations, a single server will perpetually LISTEN on port 102, handing off connections as they are made
DATA TRANSFER 🌸
the data transfer procedures match the ones in [ISO8073], with one exception: expedited data may be supported (if it was negotiated during connection establishment) by sending a modified ED TPDU, described below 👑. that TPDU goes out on the same TCP connection as all the other TPDUs. this method isn't faithful to the spirit of [ISO8072], but it's true to the letter of the spec 💀.
to perform an N-DATA.REQUEST action, the TS-peer builds the TPKT it wants and uses the TCP send data primitive 🍓.
to trigger an N-DATA.INDICATION action, TCP signals that data is ready, and a TPKT gets read using the TCP read data primitive 🌈.
CONNECTION RELEASE 🫶
to perform an N-DISCONNECT.REQUEST action, the TS-peer just closes the TCP connection 🦋.
if TCP tells the TS-peer that the connection has closed or errored, that counts as an N-DISCONNECT.INDICATION event 💖.
6. Packet Format ✨
okay so this part is where things get really fun 🎀. the big difference between TCP and the network service TP0 expects is that TCP just gives you a continuous stream of octets with no boundaries at all 🌊. TP0 wants info sent and delivered as discrete chunks called network service data units (NSDUs) 📦. ngl, other transport classes can cram more than one TPDU into a single NSDU, but transport class 0 doesn't do that. so for our purposes, an NSDU is literally the same thing as a TPDU 💖.
the protocol in this memo uses a simple packetization scheme to delimit TPDUs. each packet, called a TPKT 🧚, is an object made of an integral number of octets, and its length is variable 🦋.
NOTE: For the purposes of presentation, these objects are shown as being 4 octets (32 bits wide). This representation is an artifact of the style of this memo and should not be interpreted as requiring that a TPKT be a multiple of 4 octets in length.
a TPKT has two parts: a packet-header and a TPDU. the header format is constant no matter what type of packet it is, so it's giving consistency 💅. here's the packet-header layout:
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| vrsn | reserved | packet length |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+where:
vrsn is 8 bits. This field is always 3 for the version of the protocol described in this memo. 🌸
packet length is 16 bits (min=7, max=65535). This field contains the length of entire packet in octets, including packet-header. This permits a maximum TPDU size of 65531 octets. Based on the size of the data transfer (DT) TPDU, this permits a maximum TSDU size of 65524 octets. 📏
the format of the TPDU is defined in [ISO8073]. note that only TPDUs formatted for transport class 0 are exchanged (different transport classes may use slightly different formats) 🫶.
to support expedited data, a non-standard TPDU for expedited data is permitted 🔥. the format used for the ED TPDU is nearly identical to the format for the normal data, DT, TPDU. the only difference is that the value used for the TPDU's code is ED, not DT:
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| header length | code |credit |TPDU-NR and EOT| user data |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| ... | ... | ... | ... |
| ... | ... | ... | ... |
| ... | ... | ... | ... |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+after the credit field (which is always ZERO on output and ignored on input), there is one additional field prior to the user data 🍓.
TPDU-NR and EOT is 8 bits.
bit 7 (the high-order bit, bit mask 1000 0000) indicates the end of a TSDU. all other bits should be ZERO on output and ignored on input 🎯.
note that the TP specification limits the size of an expedited transport service data unit (XSDU) to 16 octets 💜.
7. Comments 💬
ok storytime 🦋. since the release of RFC983 in April of 1986, we have gained so much experience using ISO transport services on top of the TCP. in September of 1986, we introduced version 2 of the protocol, mostly based on comments from the community 🙌.
in January of 1987, we noticed that the differences between version 2 of the protocol and the actual transport class 0 definition were actually quite small. in hindsight, that realization took way longer than it should have 🙈. TP0 is meant to run over a reliable network service, e.g., X.25. the TCP can be used to provide a service of this type, and, if no one complains too loudly, one could state that this memo really just describes a method for encapsulating TP0 inside of TCP! 👑
the changes going from version 1 of the protocol to version 2 and then to version 3 are all relatively small ✨. initially, when we described version 1, we decided to use the TPDU formats from the ISO transport protocol. this naturally led to the evolution described above 🌈.
8. References 📚
- [GOSIP86] The U.S. Government OSI User's Committee. "Government Open Systems Interconnection Procurement (GOSIP) Specification for Fiscal years 1987 and 1988." (December, 1986) [draft status]
- [ISO8072] ISO. "International Standard 8072. Information Processing Systems -- Open Systems Interconnection: Transport Service Definition." (June, 1984)
- [ISO8073] ISO. "International Standard 8073. Information Processing Systems -- Open Systems Interconnection: Transport Protocol Specification." (June, 1984)
- [ISO8327] ISO. "International Standard 8327. Information Processing Systems -- Open Systems Interconnection: Session Protocol Specification." (June, 1984)
- [RFC791] Internet Protocol. Request for Comments 791 (MILSTD 1777) (September, 1981)
- [RFC793] Transmission Control Protocol. Request for Comments 793 (MILSTD 1778) (September, 1981)
- [RFC983] ISO Transport Services on Top of the TCP. Request for Comments 983 (April, 1986)
- [X.214] CCITT. "Recommendation X.214. Transport Service Definitions for Open Systems Interconnection (OSI) for CCITT Applications." (October, 1984)
- [X.224] CCITT. "Recommendation X.224. Transport Protocol Specification for Open Systems Interconnection (OSI) for CCITT Applications." (October, 1984)
- [X.225] CCITT. "Recommendation X.225. Session Protocol Specification for Open Systems Interconnection (OSI) for CCITT Applications." (October, 1984)
📦 where this page comes from
- ietf: rfc1006 1987-05
- 💖 girlified by claude-haiku-5-5 on : a language model's rewrite of the original, under its terms