A Protocol for Packet Network Intercommunication
1974 · paper · Computer science
Presents a protocol for sharing resources across different packet-switching networks. It handles varying packet sizes, transmission failures, sequencing, flow control, error checking and process-to-process connections, and raises open issues such as internetwork routing, accounting and timeouts.
Checked: PDF from cs.princeton.edu, fetched 4 Oct 2026. Page numbers are those of that PDF.
4 findings
Substance Receiver rule 3 (p.8) says in-window, out-of-order packets "are optionally kept or discarded, but are now acknowledged"; the rest of the paper treats these packets as not acknowledged, so 'now' should evidently be 'not'.
- p. 8 Receiver rule 3)
discarded, but are now acknowledged
- p. 9 TCP Input/Output Handling, paragraph 'If there are no more input buffers'
packets in it can be safely discarded since none of them have yet been acknowledged
- p. 8
optionally kept or discarded, but are now acknowledged.
- p. 9 TCP Input/Output Handling, last paragraph
them have yet been acknowledged
Judge's note: Checked on the page 8 and page 9 crops. Rule 1 acknowledges only packets that coincide with the left window edge. The 'Furthermore' observation on p.8 says that for in-window packets not at the left edge, an acknowledgment of the current left edge would prompt 'retransmission of the unacknowledged bytes'. P.9 says packets queued for later processing 'can be safely discarded since none of them have yet been acknowledged'. A packet that may be discarded cannot already have been acknowledged as received, and the contrastive 'but' only makes sense with 'not'. As printed, the rule states the opposite of the protocol behaviour described elsewhere in the paper.
Second check: another model agreed
Report a problem with this finding
Reference or label Page 3 says the flag field "is discussed in the section on retransmission and duplicate detection later", but that section never mentions the flag field; it is explained in 'Reassembly and Sequencing' ("The flag field of the internetwork header is used for this purpose (see Fig. 8)").
- p. 3 The Gateway Notion, paragraph 'Since the GATEWAY must understand'
and is discussed in the section on retransmission and
- p. 6 Reassembly and Sequencing
The flag field of the
- p. 3 Gateway notion, header description
is discussed in the section on retransmission and duplicate detection later
- p. 6 Reassembly and Sequencing
The flag field of the internetwork header is used for this purpose (see Fig. 8).
Judge's note: Read the whole 'Retransmission and Duplicate Detection' section (pages 7-8) in the text: it covers timeouts, acknowledgements and the window, with no occurrence of 'flag'. The ES and EM flags and Fig. 8 are in 'Reassembly and Sequencing' (page 6 image), and page 6 itself says the flag field's use is 'explained in the following section', i.e. that one. SYN and REL are covered later under 'Connection-Free Protocols with Associations'. The pointer on page 3 names the wrong section.
Second check: another model agreed
Report a problem with this finding
Reference or label "The w bytes serve as the window (see Fig. 11)." Fig. 11 is "Conceptual TCB format"; the window is drawn in Fig. 10, "The window concept". Should read Fig. 10.
- p. 7 Retransmission and Duplicate Detection, paragraph 'Suppose that the sequence number field'
the window (see Fig. 11)
- p. 7 Figure captions
Fig. 11. Conceptual TCB format.
- p. 7 Retransmission and duplicate detection, paragraph after 'Suppose that the sequence number field...'
The w bytes serve as the window (see Fig. 11).
- p. 7 Fig. 11 caption
Conceptual TCB format.
Judge's note: Page 7 image: Fig. 10 shows the left window edge, the window span and the packet sequence number space 0..n-1; Fig. 11 is the ten-row TCB layout. The preceding paragraph cites Fig. 10 for the window strategy, and page 9 introduces Fig. 11 as the TCB format. Fig. 11 does have a Window field in row 10, but the sentence defines the w-byte window in sequence space before the TCB has been introduced. Wrong figure pointer.
Second check: another model agreed
Report a problem with this finding
Wording Observations run "First ... Second, w must be less than n/2; otherwise a retransmission may appear to the receiver to be a new transmission in the case that the receiver can either save or discard arriving packets whose sequence numbers do not coincide with the receiver's left window. ... Fourth, multiple packets can be acknowledged simultaneously. Fifth ... Sixth": there is no "Third", and the "Second" sentence appears to have lost text.
- p. 8 Retransmission and Duplicate Detection, paragraph 'We make some observations on this strategy'
otherwise a retransmission may appear to the receiver to be a new transmission in the case that the receiver can either save or discard arriving packets
- p. 8 same paragraph
Fourth, multiple packets can be
- p. 8 Retransmission and Duplicate Detection, paragraph 'We make some observations on this strategy'
in the case that the receiver can either save or discard arriving packets
- p. 8 Observations on window strategy
Fourth, multiple packets can be acknowledged simultaneously.
Only the authors can say what is right here.
Judge's note: Page 8 image, left column: ordinals printed are First, Second, Fourth, Fifth, Sixth, Finally; no Third anywhere in the paragraph (also searched the text layer). The clause "in the case that the receiver can either save or discard arriving packets" is not a failure condition for the n/2 bound, and the next sentence "Thus, in the simplest implementation, the receiver need not buffer more than one packet" follows from a save-or-discard option, not from the bound. It reads as if the end of the second observation and the opening of a third were dropped (this copy is a re-typeset reprint). The exact missing words cannot be recovered from the file, hence author query.
Second check: another model agreed
Report a problem with this finding
26 minor (typos and the like; most checked by the judge only)
Reference or label Reference "[15] N. G. Anslow and J. Hanscoff, 'Implementation of international data exchange networks'" is listed but never cited in the text; the introduction cites "[1]-[7],[14],[17]" and "[8]-[12],[16]".
- p. 13 Reference [15]
[15] N. G. Anslow and J. Hanscoff
- p. 1 Introduction, first paragraph
packet switching networks [1]-[7],[14],[17]
- p. 13 Reference list
[15] N. G. Anslow and J. Hanscoff, “Implementation of international data exchange networks,”
Judge's note: I searched the full text for every bracketed citation. The body cites [1]-[7], [14], [17]; [8]-[12], [16]; [13]; [8]; [18]; [9], [10]; [12]. No citation or range covers [15], while every other entry from [1] to [18] is cited. The entry is on the page 13 image and the citation groups are on the page 1 crop. An uncited reference is harmless to the content; it was probably meant for the network group on page 1, but only the authors can say.
Second check: another model agreed · checked by hand
Report a problem with this finding
Reference or label In 'The Gateway Notion', addressing is said to be "discussed in greater detail in the next section", but the next section is 'Process Level Communication'; addressing is treated two sections on, in 'Address Formats', 'TCP Addressing' and 'Port Addressing'.
- p. 3 The Gateway Notion, paragraph describing Fig. 3
discussed in greater detail in the next section
- p. 4 Section heading following The Gateway Notion
PROCESS LEVEL COMMUNICATION
- p. 5 Section heading
ADDRESS FORMATS
Only the authors can say what is right here.
Judge's note: Page 3 image has the sentence; page 4 image shows the next heading is 'Process Level Communication', which deals with segments, multiplexing and the process header and does not discuss the source/destination address format. Page 5 image shows 'Address Formats', 'TCP Addressing' and 'Port Addressing'. Elsewhere 'next section' and 'following section' are used literally (page 6). Minor: a reader finds the material one section later; 'next section' may have been meant loosely or left over from an earlier arrangement.
Second check: another model agreed
Report a problem with this finding
Wording Text says "with the write position always to the left (module the buffer size) of the read position", while Fig. 12 (cited as the buffer layout) draws "Next Write" at the far right, after "Next Read" and "End Read"; the same cluster also lists misspellings not covered by other clusters: "A principle reason" (p. 1), "maximum transmissin size" and "Mutliplexing" (p. 4), "implicity" and "acknowlegded" (p. 8).
- p. 10 top of left column (sentence starts on page 9)
always to the left (module the buffer size) of the read position
- p. 9 Fig. 12, Transmit buffer layout
Next Write
- p. 10
always to the left (module the buffer size) of the
- p. 10 Top of page
always to the left (module the buffer size)
Judge's note: This cluster mixes two things. (1) Text versus figure: the sentence runs from the foot of PDF page 9 to the top of page 10. Fig. 12 is on PDF page 11 (not page 9 as the cluster location says); its pointer labels run left to right as Current Ack, Next Seq. No./Next Read, End Read, Next Write. In the drawing the write position is to the right of the read position. In a ring buffer the phrase 'to the left, modulo the buffer size' can be read as 'the writer may not wrap round past the reader', so the two can be reconciled; it is an ambiguity in wording, not a wrong fact, hence S3 observation. (2) Typos: I confirmed on the page images 'A prin-ciple reason' (p. 1, should be 'principal'), 'transmissin' and 'Mutliplexing' (p. 4), 'implicity' and 'acknowlegded' (p. 8). All are S4 corrections. 'module', 'paket' and 'APRA' are also in this cluster but are handled as C032, C033 and C034. The reviewer's side remark that page 1 gives 'ARPA/IPTO' is not a defect: page 1 spells out 'Information Processing Technology Office'.
Second check: another model agreed
Report a problem with this finding
Wording "We recall from Fig. 3 that an internetwork header contains both a sequence number and a byte count, as well as a flag field and a check sum." In Fig. 3 the Internetwork Header bracket ends at FLAG FIELD; CHECKSUM follows TEXT, and page 3 calls it "a trailing check sum".
- p. 6 Segment and Packet Formats, second paragraph
as well as a flag field and a check sum
- p. 3 Gateway Notion, description of Fig. 3
and a trailing check sum used for end-to-end
- p. 3 Fig. 3
CHECKSUM
Judge's note: Enlarged Fig. 3: the bracket labelled Internetwork Header spans SOURCE, DESTINATION, SEQUENCE NO., BYTE COUNT, FLAG FIELD; TEXT and CHECKSUM lie outside it. Page 3 text says the remainder of the packet is text "and a trailing check sum". Fig. 9 also shows CK after TEXT, outside the bracketed internetwork header. The page 6 sentence loosely attributes the check sum to the header; a wording slip (header for packet), no effect on the design described.
Second check: another model agreed
Report a problem with this finding
Wording Overlap example (p.8): "broken into three 200-byte packets going through a different HOST". The surrounding text is about alternate routing "through different GATEWAYS" and the first transmission passes "through one GATEWAY", so 'GATEWAY' appears to be meant.
- p. 8 paragraph 'A 600-byte packet might pass through one GATEWAY'
into three 200-byte packets going through a
- p. 8 preceding sentence
alternate routing of packets through different
- p. 8 Paragraph beginning 'A 600-byte packet'
different HOST. Since each byte has a sequence
- p. 8 Overlap problem paragraph
three 200-byte packets going through a different HOST
Only the authors can say what is right here.
Judge's note: The page 8 crop confirms the wording. The preceding sentence says the overlap problem comes from 'alternate routing of packets through different GATEWAYS', and the first half of the example uses 'one GATEWAY'. A gateway appears to each network as a host, so the wording could be loose usage rather than an error. It is therefore raised as an author query about wording, not as an outright error. The byte counts are consistent.
Second check: another model agreed
Report a problem with this finding
Typo Author footnote on p.1 reads "Engineering, Standford University, Stanford, Calif."; 'Standford' should be 'Stanford'.
- p. 1 author footnote
Standford University, Stanford, Calif.
- p. 1 Footnote author affiliation
Standford University
- p. 1
Engineering, Standford University, Stanford, Calif.
Judge's note: The page 1 crop shows 'Standford University', followed on the same line by the correctly spelt city 'Stanford, Calif.' The misspelling is in the printed text, not an extraction artefact.
Second check: another model agreed
Report a problem with this finding
Typo "A prin-ciple reason for developing such networks" uses 'principle' where the adjective 'principal' is meant.
- p. 1 Introduction, first paragraph
A prin-
- p. 1 Introduction, first paragraph
A prin-ciple reason for developing such networks
Judge's note: Page 1 image, Introduction first paragraph: the word breaks as 'prin-' / 'ciple'. The ending is 'ciple', so this is the wrong word, not a hyphenation artefact.
Second check: none (judge only)
Report a problem with this finding
Typo "To make the data meaningful, computer and terminals share a common protocol (i.e, a set of agreed upon conventions)": singular 'computer' should be 'computers', and 'i.e,' lacks its second full stop.
- p. 1 Introduction, first paragraph
make the data meaningful, computer and terminals
- p. 1 Introduction, first paragraph
(i.e, a set of agreed upon
Judge's note: Page 1 image, Introduction first paragraph. Both slips are visible: 'computer and terminals share' follows 'between computers and terminals' in the previous sentence, and the abbreviation is printed '(i.e, a set'.
Second check: none (judge only)
Report a problem with this finding
Typo "The transmit time for this data is usually dependent upon internal network parameters": 'transit time' is probably meant, since the listed factors (routeing, propagation delays) describe time to cross the subnet.
- p. 1 right column, paragraph 'In a typical packet switching subnet'
The transmit time for this data is
Judge's note: Page 1 image, right column, confirms the wording 'transmit time'. The paper uses 'transit' elsewhere ('during transit', page 3). 'Transmit time' is intelligible, so either form could be intended; only the authors can say.
Second check: none (judge only)
Report a problem with this finding
Typo "determination of status of the networks components" is missing the possessive apostrophe ("network's components").
- p. 1 right column, paragraph 'In a typical packet switching subnet'
determination of status of the networks components
Judge's note: Page 1 image, right column, end of the paragraph 'In a typical packet switching subnet': printed 'networks components' with no apostrophe. The paper uses the apostrophe elsewhere ("any network's packet size", page 3).
Second check: none (judge only)
Report a problem with this finding
Typo Spelling varies between "routeing" ("and signalling strategies, routeing, propagation", page 1; "routeing issues", page 5) and "routing" (abstract, page 2, "essential to routing and delivery" on page 5).
- p. 1 right column
and signalling strategies, routeing, propagation
- p. 5 TCP Addressing
routeing issues
- p. 5 Address Formats
essential to routing and delivery
Judge's note: Page 1 and page 5 images show both spellings; on page 5 'routing' (Address Formats, and 'routing responsibility') and 'routeing' (TCP Addressing) sit within a column of each other. Both spellings are legitimate; the defect is only the inconsistency.
Second check: none (judge only)
Report a problem with this finding
Typo Sentence begins in lower case after a full stop: "are typically different in each network. thus, to obtain verification of certain conditions".
- p. 2 item 5) under Fig. 1
thus, to obtain verification of certain conditions
Judge's note: Page 2 image, item 5) below Fig. 1: a full stop after 'network' at the line end, and the next line starts with lower-case 'thus,'. Either 'Thus' or a comma/semicolon was intended.
Second check: none (judge only)
Report a problem with this finding
Typo "a packet may be fragmented by one or more GATEWAY" should be the plural 'GATEWAYS'.
- p. 3 right column, last paragraph
be fragmented by one or more GATEWAY
Judge's note: Page 3 image, right column last paragraph: 'be fragmented by one or more GATEWAY. We'. The plural 'GATEWAYS' is used elsewhere (e.g. Fig. 2 caption and page 3 left column). Similar missing plural or article at 'We take the position that GATEWAY should not perform this function' (page 3) and 'smaller pieces by GATEWAY creates' (page 6), both visible in the images.
Second check: none (judge only)
Report a problem with this finding
Typo "Within a HOST we assume that existence of a transmission control program (TCP)": 'that' should be 'the'.
- p. 4 Process Level Communication, second paragraph
we assume that existence of a
Judge's note: Page 4 image, Process Level Communication, second paragraph. The sentence is ungrammatical as printed.
Second check: none (judge only)
Report a problem with this finding
Typo P.4, Process Level Communication: "the local network may limit the maximum transmissin size"; 'transmissin' should be 'transmission'.
- p. 4 Process Level Communication, second paragraph
maximum transmissin size
- p. 4
network may limit the maximum transmissin size,
- p. 4 Process Level Communication, third paragraph
Mutliplexing and demultiplexing
- p. 4
the local network may limit the maximum transmissin size
Judge's note: The page 4 crop shows 'transmissin' in the printed text. The 'Mutliplexing' slip one reviewer also mentioned is judged separately under C015.
Second check: another model agreed
Report a problem with this finding
Typo P.4: "Mutliplexing and demultiplexing of segments among processes"; 'Mutliplexing' should be 'Multiplexing'.
- p. 4 Process Level Communication, third paragraph
Mutliplexing and demultiplexing of segments
- p. 4
Mutliplexing and demultiplexing of segments among processes
- p. 4 Process level communication
Mutliplexing and demultiplexing
Judge's note: The page 4 crop shows 'Mutliplexing' at the start of the paragraph. The next sentence spells 'multiplex' correctly.
Second check: another model agreed
Report a problem with this finding
Typo "of special importance when packets must be split apart from propagation through the next local network": "from" should be "for".
- p. 7 left column, paragraph 'The ES and EM flags'
importance when packets must be split apart from
- p. 7 Reassembly and Sequencing, paragraph on ES and EM flags
split apart from
- p. 7 paragraph before Fig. 9 discussion
must be split apart from propagation through the next local network
Judge's note: Read on the page 7 image, left column. Packets are split so they can propagate through the next network; "from propagation" does not parse. Wrong word, evident fix.
Second check: none (judge only)
Report a problem with this finding
Typo "(voie virtuelle transmission mode [8]) and the ARPANET very distant HOST connection [18]), is proposed here": unmatched closing parenthesis after "[18]".
- p. 7 right column, paragraph 'A window strategy'
connection [18]), is proposed here
- p. 7
ARPANET very distant HOST connection [18]), is proposed here
Judge's note: Page 7 image, right column: one opening parenthesis before "voie virtuelle", closed after "[8]", then a second closing parenthesis after "[18]" with no opening partner.
Second check: another model agreed
Report a problem with this finding
Typo Sender rule 3 (p.8): "(advancing the right window edge implicity)"; 'implicity' should be 'implicitly'.
- p. 8 top of left column
edge implicity)
- p. 8
advancing the right window edge implicity).
- p. 8 Flow Control, first sentence
TCP is ultimately acknowlegded
- p. 8 Sender rule 3
(advancing the right window edge implicity)
Judge's note: The page 8 crop shows 'implicity' at the top of the left column. The 'acknowlegded' slip one reviewer also mentioned is judged separately under C027.
Second check: another model agreed
Report a problem with this finding
Typo Flow Control (p.8): "TCP is ultimately acknowlegded by returning the"; 'acknowlegded' should be 'acknowledged'.
- p. 8 Flow Control, first paragraph
TCP is ultimately acknowlegded by returning the
- p. 8
destination TCP is ultimately acknowlegded by returning the
- p. 8 Flow control
ultimately acknowlegded by returning
- p. 8
is ultimately acknowlegded by returning the
Judge's note: The page 8 crop shows 'acknowlegded' in the first sentence of the Flow Control section.
Second check: another model agreed
Report a problem with this finding
Typo TCB field 2: "This is the full net/HOST/TCP/port of the receiver." Field 1 reads "the full net/HOST/TCP/port address of the transmitter"; the word "address" is missing in field 2.
- p. 9 TCB field list, item 2)
TCP/port of the receiver
Judge's note: Page 9 image, right column, items 1 and 2. The parallel item has "address"; item 2 omits it. Meaning is unaffected and the shorter form is readable as it stands.
Second check: none (judge only)
Report a problem with this finding
Typo "always to the left (module the buffer size) of the read position" should read "modulo the buffer size".
- p. 10 top of left column
(module the buffer size)
Judge's note: Seen on the page 10 image, first line of the left column. Page 8 uses the correct term: 'must be made modulo n'. Also quoted inside C031; kept here as the single entry for this typo.
Second check: none (judge only)
Report a problem with this finding
Typo Connections and Associations (p.10): "communication in paket switched networks has been influenced"; 'paket' should be 'packet'.
- p. 10 Connections and Associations, first paragraph
communication in paket switched networks
- p. 10
communication in paket switched networks has been influenced
- p. 10
communication in paket switched networks has
Judge's note: The page 10 crop shows 'paket' in the first paragraph of the section.
Second check: another model agreed
Report a problem with this finding
Typo Footnote 6 reads "S. Crocker of APRA/IPT."; the agency is ARPA.
- p. 11 footnote 6
S. Crocker of APRA/IPT.
- p. 11
6 S. Crocker of APRA/IPT.
Judge's note: Seen on the page 11 crop and full page. The paper itself spells the agency 'Advanced Research Projects Agency' (affiliation on page 1, Kahn biography on page 13) and uses ARPA and ARPANET throughout, so the fix is internal to the paper.
Second check: another model agreed
Report a problem with this finding
Typo Reference [7] gives the title as "Flow control in a resource-shaping computer network"; probably "resource-sharing".
- p. 12 Reference [7]
resource-shaping computer network
- p. 13 References, [17]
Resource-sharing computer
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: the true title of Kahn and Crowther's 1972 IEEE Trans. Commun. paper, which cannot be checked from this PDF. The page 12 image does print 'resource-shaping'. Inside the paper, references [1] ('resource sharing'), [12] ('resource sharing computer network') and [17] ('Resource-sharing computer communication networks', same author) all use 'sharing', and 'resource-shaping' appears nowhere else, so a one-letter slip is likely.
Second check: none (judge only)
Report a problem with this finding
Typo Kahn biography: "National Lecturer for the Assocation for Computing Machinery in 1972" should read "Association".
- p. 13 Robert E. Kahn biography, last sentence
National Lecturer for the Assocation
- p. 13
National Lecturer for the Assocation for Computing Machinery in 1972.
- p. 13 Biography of Robert E. Kahn, last sentence
Assocation
Judge's note: Seen on the page 13 image, last sentence of the Kahn biography. The line above spells 'Mathematical Association of America' correctly.
Second check: another model agreed