Danny Cohen or John Shoch (ien 22, not online): a Name is what we seek,
an Address is where it is, and a route is how we get there.
"A Partial Specification of an International Transmission Protocol"
by Vint Cerf.
- From time to time, I have departed from the proposed
terminology, but I believe I have done this only where
it seems appropriate.
- ...it will be necessary to assume that a terminal cannot be
a HOST unless it is capable of performing according to the
specified protocol... (on the topic of what qualifies a
machine to be directly connected to a TCP network)
- The job of the TCP is merely to take a stream of messages
produced by one HOST and reproduce the stream at a foreign
receiving HOST without change.
- The international transmission protocol (ITP) ... is intended
to ... Be relatively easy to implement
- It is tempting to try to pick a message size which all
networks can transmit without alternation (e.g. 255 bytes),
but our guess is that such an attempt is unlikely to
- ...barring total disaster, all elements are eventually
acknowledged... (on the topic of flow control and
"A Protocol for Packet Network Intercommunication" by Vint Cerf and
- It would be extremely convenient if all the differences between
networks could be economically resolved by suitable
interfacing at the network boundaries.
- The responsibility for properly routing data resides in the
- The choice for network identification (8 bits) allows up to
256 distinct networks. This size seems sufficient for the
foreseeable future. Similarly, the TCP identifier field
permits up to 65536 distinct TCP's to be addressed, which
seems more than sufficient for any given network.
- We assume that 16 bits are sufficient to serve as internetwork
- No transmission can be 100 percent reliable.
- (thus far, no provision has been made for error
- Unfortunately, ... operating systems tend not to have an
infinite supply of internal port addresses. These internal
addresses are reassigned after the demise of each port.
- from "RFC 675"
by Vint Cerf, Yogen Dalal, and Carl Sunshine.
- all network communication is viewed as inter-process
- The TCP acts in many ways like a postal service since it
provides a way for processes to exchange letters with each
other. ... In addition to acting like a postal service,
the TCP insures end-to-end acknowledgment, error correction,
duplicate detection, sequencing, and flow control.
- Conceptually, the TCP is made up of several processes. Some
of these deal with USER/TCP commands, others with packets
arriving from the network.
- ...half-open connections are expected to be unusual...
- From the standpoint of controlling buffer congestion, it
appears better to TREAT INCOMING PACKETS WITH HIGHER PRIORITY
THAN OUTGOING PACKETS.
- ...some wasteful processing may occur...This is the general
problem of trying to implement a protocol that is fundamentally
asynchronous, but at least it is immune to harmful
- ...the cost of buffering to smooth out bursty traffic is borne
partly by the TCP's, and partly by the user at the send side.
None of it is borne by the communication subnet.
- We view the [connection] measurement process as something
which occurs internal to the TCP but which is controllable
from the outside...We would restrict access to TCP
measurement control to a few designated sites...
- from "BBN report #2918"
- We have also come to believe that it is essential to have a
reset mechanism to prevent "impossible" deadlocks and other
conditions that may result from hardware or software
- We also recognize that there are no absolutes at this stage in
the development of computer networks...(on differing
opinions on design strategies)
- from IEN 2 by Jon Postel (1977)
- ...internetwork communication should be view[ed] as having
two components: the hop by hop relaying of a message, and
the end to end control of the conversation. This leads to
a proposal for a hop by hop oriented internet protocol, an
end to end oriented host level protocol, and the interface
- We are screwing up in our design of internet protocols by
violating the principle of layering.
- Ah the data, at last a space for the reason we are going
through all this nonsense.
- Only hop to hop error control should be attempted in the
internet protocol. Specific host level protocols such
as TCP can provide for end to end error control.
- from Internet System Handbook by many authors (1993)
- At the heart of the Internet are two protocols, IP and TCP...
The IP, or Internet Protocol, smoothes over the differences
between different kinds of network technologogies, so that
they offer a common set of delivery services, regardless of
their physical characteristics. The TCP, or Transmission
Control Protocol, provides reliability over the IP-mesh.
-Vint Cerf (ch 4 preface)
- The problem of providing reliable, end-to-end communications
between computers using multiple packet networks was the
central focus of the internetting research effort.
-Vint Cert (ch 4 intro)
- This 8-bit [TTL] field was originally intended to be used to
specify how long, in seconds, an internetwork packet could
persist in the internet before being discarded by a router
or a host. ... Effectively this field has become a kind of
maximum hop count.
-Vint Cerf (section 4.2.6)
The Design Philosophy of the DARPA Internet Protocols by
David D. Clark (1988)
- The ideas developed in [the DARPA packet switched protocol
development project] have also influenced other protocol
suites, most importantly the connectionless configuration of
the ISO protocols.
- ...an understanding of the history of the design provides a
necessary context for current design extensions
- ...the applications being supported, such as remote login,
were naturally served by the packet switching paradigm, and
the networks which were to be integrated together in this
project were packet switching networks. So packet switching
was accepted as a fundamental component of the Internet
- The initial concept of TCP was that it could be general enough
to support any needed type of service. However, as the full
range of needed services became clear, it seemed too difficult
to build support for all of them into one protocol.
- It was thus decided, fairly early in the development of the
Internet architecture, that more than one transport service
would be required, and the architecture must be prepared to
tolerate simultaneously transports which wish to constrain
reliability, delay, or bandwidth, at a minimum. This goal
caused TCP and IP, which originally had been a single
protocol in the architecture, to be separated into two
- The Internet architecture achieves ... flexibility by making
a minimum set of assumptions about the function which the
net will provide. The basic assumption is that the network
can transport a packet or datagram ... must be of a reasonable
size ... The network must have some suitable form of
addressing if it is more than a point to point link.
- By engineering these services at the transport, for example
reliable delivery via TCP, the engineering must be done only
once, and the implementation must be done only once for each
- The [Internet] architecture tried very hard not to constrain
the range of service which the Internet could be engineered
- The relationship between architecture and performance is an
extremely challenging one. The designers of the Internet
architecture felt very strongly that it was a serious mistake
to attend only to logical correctness and ignore the issue
- the role of a datagram ... is as a building block, and not as
a service in itself.
- The original ARPANET host-to-host procotol provided flow
control based on both bytes and packets. This seemed overly
complex and the designers of TCP felt that only one form of
regulation would be sufficient. The choice was to regulate
the delivery of bytes, rather than packets.
last modified at 21:15 21 mar 2001 by
feedback and questions are welcome!