Network news in linux

Computer Store News The origins of online news go back to 197. At that time, two college students, Tom Tr s c o t t and Jim Ellis, thought of using U U C P to connect computers for each user to exchange information. As a result, they facilitated the establishment of a small network of three computers at the University of North Carolina. Initially, the transmission of information was handled with a large number of shell scripts (later rewritten in C), but these scripts were never publicly released. Because they were quickly replaced by “A” News, the first publicly released news software.
When designing the news software, there was no plan to have it deal with a large number of articles in each newsgroup every day. But as the amount of information continues to grow, it is already overwhelmed, so Mark Horton and Matt Glickman rewrote it, and they call it "B& quo" (also known as B n e w s). The first version of B n e w s was 2.1 released in 1928. Later, people continued to expand it and added many new features. Its current version is Bnews 2.11. As its last official maintenance personnel turned to I N N, it was gradually abandoned. Another rewrite of the “A” news software was completed and published in 1987, which was rewritten by G e ff Collyer and Henry Spencer, which is called “C””. Later, many C-N e w s patches appeared, the most prominent being C-N e sPerformance Release. For sites that run many newsgroups, the cost of frequently calling transit news is very large because it is responsible for distributing all incoming newsgroup articles. Performance Release adds an option to transit news that allows administrators to run it in daemon mode, where the program runs itself in the background. Performance Release is the latest C-N e w s version. All news in the C version is primarily used in the U U C P network, but it can also be used in other environments. A new solution is needed to effectively transmit news on T C P /I P, D E C N e t or related networks. So N N T P (Network News Protocol) came into being in 1986. The protocol is based on a network connection and specifies a number of commands for interactive transmission and access to news. There are many N N T P based applications online. One of them is the n n t p d package, which was written by Brian Barber and Phil Lapsley to provide news reading services for a large number of hosts on the LAN. The purpose of designing n n t p d is to perfect news packets such as B n e w s and C - N e w s, giving them N N T P characteristics. Another different N N T P package is I N N, also known as Internet News. It is not just a front end but an independent news system. It contains a sophisticated news relay daemon that effectively maintains several concurrent N N T P links, so it is also the preferred news server for many Internet sites. 15.1 What is Usenet About U s e n e t, the most unexpected is that it does not belong to any organization, and there is no core network administrator. In fact, apart from the technical description, it is only part of the knowledge of U s e n e t, so you can't define it exactly as ……, it can only be said that it is not ……. If you have Brendan Kehoe's "The Zen and the Art in the Internet" (located in ww w. cs. indiana. edu /docproject /zen /zen - 1. 0 toc /html); or see his homepage (ww w Zen. o rg /~ bren -dan), you can fully understand the characteristics of the US senet. Nonetheless, U s e n e t is still defined as a group of independent sites that exchange U s e n e t news with each other. To become a U s e n e t site, you only need to find another U s e n e t site, agree with its owner and maintenance staff, and allow them to exchange U s e n e t news with you. To provide news for another site, it is called “send, distribute or exchange news for it”, and use another familiar U s e n e t quote: “provide news, you can go to U s e n e t”. The basic unit of U s e n e t news is the article. That is, the user writes or “delivers” to the online message. In order to enable the news system to process them, some administrative information is added in front of these articles, which is called “article head”. It is very similar to the header specified in the Internet mail standard R F C - 8 2 2 and consists of several lines of text. The format of each line is this: field name: field value. Note that the Usenet news message format is specified in RFC-1036 & nbsp;USENET Message Exchange Standard". Newsgroup articles are submitted to one or more newsgroups. We can think of a newsgroup as a place to gather articles related to a topic. All newsgroups are organized in a hierarchical structure, and the group names of each group represent its position in this structure. In this way, you can see at a glance what the theme of a newsgroup is. For example, anyone can know from the newsgroup name c o m p . o s . l i n u x . a n n o u n c e, the newsgroup is used to publish a message related to the computer operating system named L i n u x . After submission, these newsgroup articles were exchanged between all of the sites that were willing to disseminate news for the group. After the two sites agree to exchange news, they can exchange newsgroups freely according to their own preferences, and even join their own local news structure. For example, g r o u c h o . e d u has a point to b a r n y a r d . e d u (which is a major news feed site) and several links to secondary news feed sites. Now, B a r n y a r d college may have received all the newsgroups, and G M U only wants to transmit a few major news structures such as (s c i, c o m p, r e c, etc.). Some downstream sites, such as a U U C P site called b r e w h q, intend to transmit fewer newsgroups than the former because they lack network and hardware resources. On the other hand, b r e w h q may want to receive newsgroups in the f j structure, but G M U does not transmit these newsgroups. Therefore, it maintains another link to g a rg l e b l a s t e r. c o m, which will transmit all f j newsgroup structures and send them to b r e w h q. 15.2 How Usenet Controls News Today, the growth rate of U s e n e t is increasing. For example, in the case of sites that transmit entire online news, the amount of data they transmit per day is generally between 3 and 5 G B. This of course requires more control over the news. Therefore, it is necessary to talk about how most systems handle U s e n e t news. Note If you want to learn about the latest U s e n i x theme (some very interesting topics are involved), be sure to check it out here: www.infosys.tuwien.ac.at/staff/pooh/papers/NewsCacheHP/. The distribution of news between networks is done by different transmission points. U U C P is commonly used in the past, and now it is mainly transmitted by Internet sites. The route used is called "diffusion routing" (f l o o d i n g). Each site maintains many links to other sites (newsletters), and any article sent or received by the local news system will be forwarded to the linked site, but if they are already on the linked site, they will be throw away. By looking at the path: the article header field, the site can find out which sites the article has gone through. The article header contains a list of all the systems, and the b a n g path expression shows which systems the article has experienced. In order to distinguish between the article and identify whether the article is duplicated, the U senet article must transmit a message ID (specified in the Message -ID: header field), which is a combination of the sending site name and a serial number, ie < serial @ site> . For each article processed, the news system records this I D in a history file. To check all new articles, check out this history file. The communication flow between any two sites is subject to two standards: First, an article corresponds to an allocation (defined in the D istribution : header field) to limit the transmission of the article to a specific site group. Second, the newsgroups involved in the exchange are subject to restrictions by both the sending and receiving systems. The set of newsgroups and assignments allowed to be sent to a site are usually kept in the sys file. The absolute numbering of the article usually requires the aforementioned scheme. Improvement. The UUCP network itself is used to collect articles on a regular basis and to group the collected articles into a single file. The file is then compressed and sent to the remote site. This is called “batch processing” (batching). Another approach is to use the ihave /sendme protocol, which prevents repetitive articles from being transmitted locally to remote sites, saving network bandwidth. It does not take all the articles into the batch file and sends them together. Remote site, but only combine the message ID of the article into a large "ld";i Hav e” message, then send it to the remote site. Then, read this message, compare it with its own history file, and list the list of articles you need in the returned “sendm e” message Finally, only those articles in the message are sent to the remote site. Of course, ihave /sendme only applies to this situation: the two large sites that exchange news each receive several articles sent by the sending site, which are frequently exchanged. Articles. Sites on the Internet generally rely on TCP/IP-based software for transmission. It uses the “News Transfer Protocol” (NNTP, see RFC - 9 7 7 for details). It is news on both sides of the news. Transmit and provide access to individual users on remote hosts. NNTP recognizes three news transmission methods. One is the real-time version of ihave /sendme, also known as "push" (pushing) news. The second is &ldquo [pulling] news, in this way, the client requests a list of specified news Or a list of articles in the structure (these articles arrive at the server site after the specified date) and select those articles that are not in their own history files. The third way is for interactive reading, allowing you or your news to read. The machine gets the article of the specified newsgroup and the article with incomplete post information. On each site, the news is stored in a directory structure under /var/spool/news, each article is in a separate file. Each newsgroup is in a separate directory. The directory name consists of the newsgroup name and path part. Therefore, c o m p . o s . l i n u x . m i s c is stored in /v a r /s p o o l /n e w s /c o m p /o s /l i n u x /m i s c. Articles in the newsgroup are numbered according to their order of arrival. This number is used as the file name. Currently, the number range of online articles is stored in a file called a c t i v e, which also acts as a list of newsgroups known to your site. Since disk space is limited, some articles should be deleted at intervals. This is called "expiration". Often, articles from specific newsgroups and structures expire on the day they arrive. This period can be rewritten by the delivery person by specifying the expiration date in the E x p i r e s : f i e l d field at the beginning of the article.