Organization: China Interactive Publishing Network (http://www.china-pub.com/)
RFC Document Chinese Translation Program (http://www.china-pub.com/compters/emook/aboutemook.htm)
E-mail: Ouyang@china-pub.com
Translator: McSewang Mcsewang@21cn.com)
Translation time: 2001-10-18
Copyright: This Chinese translation copyright belongs to China Interactive Publishing Network. Can be used for non-commercial use free reprint, but must
Keep the translation and copyright information of this document.
Network Working Group S. CROCKER
RFC-37 UCLA
20 march 1970
Postscript and other
(Network Meeting Epilogue, ETC.)
table of Contents
About the meeting 1
Mail list change 2
Process 2
A few days after the meeting 3
About the meeting
On March 17, 1970, I hosted a web conference in UCLA. About 25 people participated in this time
Meeting, including generations from Mit, LL, BBN, Harvard, Sri, Utah, UCSB, SDC, RAND, and UCLA
table. I propose amended to the protocol in the RFC # 33 document, which is outlined in the RFC # 36 document.
The main modifications are concentrated in the easy-to-permeability of dynamic reproduction.
The protocol based on socket and ordinary single connection is very different from the protocol in RFC # 11. Can promote
This progress is benefited from online conferences held in Utah on December 8, 1969, at the meeting (formerly
The limitations showing the requirements of the login and the incompetence of the forced connection are strong.
question. Therefore, the main purpose of the recent online conference is to seek everyone's opinions on the new agreement.
Remember that the situation may not be the same, but I think the agreement has been widely accepted, and the opinions and discussions of criticism
Concentrate in two cases:
1. Question the complexity and effectiveness of all protocols, especially the need for dynamic reproduction.
2. Other topics, especially translation conversions, more advanced languages and incompatible devices, and more.
The judgment of the basic concepts of sockets and connections is obviously lacking (even if there is also very beautiful). I will reach it.
Next consensus:
1. I am responsible for publishing a viable specification online before April 30.
2. Any interested group can immediately (at least) and I will contact me if you want to modify the agreement.
3. If some major corrections are included in the agenda, those interested groups can meet again. This will be in two to three
It is completed in the week.
4. Jim Forgie of the Lincoln Lab has already agreed to host a meeting on a higher level of network language, which is roughly
In the same spring in the spring.
Mail list change
The Paul Rovner of the Lincoln Laboratory is replaced:
James forgie
Mass. Institute of Technology
Lincoln Laboratory C158
P.O. Box 73
Lexington, mass. 02173
Televhone At (617) 862-5500 EXT. 7173
Added Professor George Mealy:
George mealy
RM. 220
AiTKEN Computation Lab.
Harvard University
Cambridge, Massachusetts 02138TelePhone At (617) 868-1020 EXT. 4355
process
In all our writings, a term "process" is used, which is used to describe a program that is assigned a
A specific counter pointer and a piece of address space. A program is just some binary stored in a file.
Bit. A new process can only be created by an existing process. The previous process must make a micro-operation to promote
Come this creation process. The process can be terminated by oneself or other (usually a higher level) process.
The above definition description is based on Vysotsky, etc. in the 1965 FJCC conference, "Structure of the Multics"
Supervisor's 206 and 207 pages presented for definitions given.
Since a process can create another process, it is difficult to distinguish from the outside for the general two processes.
There is no reasonable way to make the two processes directly request connections. The function of the socket is in two processes.
A standard interface is provided between.
A few days later
In some days after the meeting, I used Steve Wolfe (UCLA-CCN), Bill Crowther (BBN), John
Heafner and Erick Harslem (Rand) are conversed. Wolf's review will be published as RFC # 38 and have been classified
I will have a class below.
The following is a Crowther submitted:
Here is a brief description of the simplified opinions of the host protocol proposed at the March meeting. These comments
It has not been carefully considered, just as a reference view.
The first idea to connect
----------------------
An NCP that is just wanting to connect to tells it that each neighbor "I want to connect again". They have been waiting for transmission lines
There is no message in the road, I will issue "OK" response. Then it responded again: "Let's connect it."
Connection is established. In extreme cases, NCP will also receive a "I want to make a re-connection" response instead of "ok",
At this point, only one reactive process can continue, and the other must stop. So we will pick up from high-end host users
The "re-connection" information received is recognized as "OK", and the information received from the low-end user is noted "No, wait for me next time.
Pick up "to connect to high-end users.
Second idea for renewal
----------------------
There is no repetitive connection and link. Still establish a connection, but use any convenient link for information. Returning until a FRNM
Send next information from the same connection again. Add the number of sources and destination sockets in the packet.
I want to connect again, tell each neighbor "Please talk about the following ...". Keep a small paragraph in this connection
Time (in seconds) and packet groups and connection information to their destination address. I have not yet developed non-transmission information
The order of the order, but if you send a reconnection request when you have not received RFNM, perhaps everything will be smooth.
I don't think it is better than my better or big difference than me. I think so. I haven't been there.
There are too many ideas, but I am trying to find some ideas.
The second idea of Bill looks opposite to my view of the role of the link. A support for connection and links
The evidence of the restriction is that the number of connections between the two hosts may need to exceed 255, and even if this is not, this
The dependence on the elimination design is also very practical. On the other hand, recently "link equipment
Stop bit "(on page 22 on the release of the release of February 1970) in the upcoming BBN # 1822 report).
(Bill just add this feature to go in, no maintenance.) Another observation believes that from intuitive watching
The potential ability to use link bandwidth to transmit data is not a good idea. (Please pay attention to conflicts of various standards)
In a meeting with Rand's John Haefner and Eric Harslem, they pointed out that there is no need to reserve the current agreement.
Misuse and reporting, identity testing, reporting, expansion capabilities and test methods. Error detection and identity test will need one
Some discussions use to study what is useful, I hope this discussion can be carried out in the process of implementation. In the field
The expansion and trials of the protocol have been made now.
I recommend retaining two aspects of expansion. One is only part of 256 links, such as the first 32. another
On the one hand, from 255 or less, combined with the continuous code to assign the value from the number of links in the use, I think
In most of the time, we will not use more than 32 links, in addition, the network will never process from overload chains
The traffic hidden in the road is hidden. (These two points don't have to be realized at the same time: a host can define multiple links, but only
There is a small portion to be transmitted at any given time. )
Other ideas of Heafner and Harslen will appear in the NWG RFC document.
Subsequent activities
-------------------------------
In the next few days, I will continue to pay attention to the current protocols that may cause or need a major modification.
Agency. Since then, the focus will be concentrated on modified words, specific implementation, protocol expansion, and utilization. I am in UCLA, I have to be with me.
Contact, you can give me a secretary Benita Kristel, Ms. (213) 825-2368. Moreover, we welcome each
The man submitted to the NWG / RFC series, Benita will set a unique number.
"Stop bit on the link device" is a method of changing RFNM's by receiving host so as to compress data traffic. A generation
The way is the host to control the command with the host.
[This RFC WAS PUT INTO MACHINE READABLE FORM for Entry]
[INTO THE ONLINE RFC Archives By Ron fitzherbert 1/97]
RFC37 ---- NetWork Meeting Epilogue, ETC. After the online meeting,
4
RFC Document Chinese Translation Program
