"Y" Networking Definitions & Concepts...

Yahoo .. to .. Ymodem

Search for Information Technology Items

Y et A nother H ierarchically O diferous O racle, an extendible collection of subjects in a very large world wide arena. This is WWW search site that lets you search by subject, via its subject libraries that are linked across the World.

Like many other aspects of the computer age, Yahoo! began as an idea, grew into a hobby and lately has turned into a full-time passion. The two developers of Yahoo!, David Filo and Jerry Yang, Ph.D. candidates in Electrical Engineering at Stanford University, started their guide in April 1994 as a way to keep track of their personal interests on the Internet. Before long they found that their home-brewed lists were becoming too long and unwieldy. Gradually they began to spend more and more time on Yahoo!.

During 1994 they converted Yahoo! into a customized database designed to serve the needs of the thousands of users that began to use the service through the closely bound Internet community. They developed customized software to help them efficiently locate, identify and edit material stored on the Internet. The name Yahoo! is supposed to stand for "Yet Another Hierarchical Officious Oracle" but Filo and Yang insist they selected the name because they considered themselves yahoos. Yahoo! itself first resided on Yang's student workstation, "akebono" while the search engine was lodged on Filo's computer, "konishiki." (These machines were named after legendary Hawaiian sumo wrestlers.)

In early 1995 Marc Andreessen, co-founder of Netscape Communications in Mountain View, Ca. (and the developer of the two most popular Web-browsers) invited Filo and Yang to move their files over to larger computers housed at Netscape. As a result Stanford's computer network returned to normal, and both parties benefited. Today, Yahoo! contains organized information on tens of thousands of computers linked to the Web. The San Jose Mercury news recently noted that "Yahoo is closest in spirit to the work of Linnaeus, the 18th century botanist whoseclassification system organized the natural world."

Yahoo Features:

• What's Cool: On this page, they present entries they think are cool. In particular, they have selected specific categories within Yahoo! in which they find amusing, extraordinary, or especially useful content.
• What's New: You can see what they've added in the last week, along with the part of the hierarchy in which they reside.
• Random Link: a script provided by Ed Kubatis which picks a random link from the Yahoo! database - give it a click and see where it takes you!
• Headlines: Yahoo! provides the top headline news hourly, plus news in specific categories such as Business, Entertainment, Sports, International, and U.S. Politics.
• Picks of the Week: Updated on a weekly basis, their Picks section contains sites that they think are good for various reasons - timely, informative, wacky, you name it. Included here is their weekly World Tour, highlighting some of the great stuff found in their regional categories.
• Yahoo! for the Day: Every day, they present a selection of sites from their new additions that stand out as noteworthy.

The Yellow Pages (YP) protocol, like the NFS (Network File System), MOUNT or port mapper protocols, is an RPC (Remote Procedure Call) service of the application layer. As the name Yellow Pages indicates, YP provides a directory service. YP has recently been renamed Network Information Service (NIS), since the name Yellow Pages is a registered trademark. It is a registered trademark in the United Kingdom of British Telecommunications plc for their (paper) commercial telephone directory. Sun changed the name of their system to NIS, though all the commands and functions still start with "yp", e.g. ypcat, ypmatch, ypwhich.

The YP protocol and the implementation of the YP service over server and client offers the possibility of managing the mappings in a central location in the network (the master server), thereby attempting to rule out overlaps of the user IDs. However, as before, programs running on other computers must have access over YP to the information held centrally. This is catered for by client functions in the form of YP library routines and some YP utilities.

It is possible to use a service such as YP for information other than the distribution of user and group IDs. Fortunately, this service is flexible and the network administrator may also use it for any other tables.

YP Mode of operation

YP is a distributed system: it operates with several servers on different intercommunicating computers in the network. One of the servers is designated as the master server by the network administrator; the master server possesses the original version of a table. In order to be fail-safe, and possibly to spread the load, other slave servers are started up on other computers. If the contents of the tables are altered, the slave servers are brought up to date with the latest version by the master server.

The information in the servers is organized in maps, which are each valid only within an administrative area (or domain). Thus, if is possible to distinguish between several domains in the same network, for example, different departments within a company. Since there is no feedback between the DNS (Domain Name Server) and the YP systems, the YP domains are not necessarily identical with the DNS domains and may actually be organized in a different way.

Every YP map contains data records consisting of an index key and the corresponding information record (usually an ASCII character string). To make this clearer, we shall consider an example form the /etc/passwd file where the index is cleverly constructed from the user name. A query with the key santi would produce the data record:

santi:xjieQuhd:123:Michael Santifaller:/user/santi:

As this example also shows, data records in the passwd map and in most other YP maps are in the same format as the corresponding line of the original file. This means that programs which previously accessed the /etc files directly remain virtually unchanged. In addition the /etc/passwd and /etc/group files needed to assign access rights in NFS operation, for technical administrative reasons, the following files are available on most NFS systems as YP maps:

• /etc/passwd
• /etc/group
• /etc/hosts
• /etc/ethers
• /etc/networks
• /etc/rpc
• /etc/services
• /etc/protocols
• /etc/netgroup
• /etc/lib/aliases

In the case of some of the files, a central management structure seems to be possible. The information in a file such as /etc/hosts is equally valid for all computers in the network and YP guarantees that the latest version is available throughout the network.

YP Procedures

The RPC identification for YP is as shown in the table above. Naturally, there is a reason for the very surprising fact that an official service such as YP has two differnt and equally valid(!) version numbers. Since the first version of the protocol had serious design faults which led to problems on reconfiguration after a server crash, a second version had to be created very rapidly. For upward compatibility, both versions are supported by the server (which is possible using the version number in the RPC protocol).

The YP protocol incorporates procedures to search for YP server in the network, to interrogate data records using a key, to transfer complete maps and to interrogate management information.

The YPBIND protocol, which is not discussed in detail here, is associated with the YP service. It helps the YP client to find a YP server in the network.

Use of YP

Correctly used, there is no doubt that YP brings with it a simplification of NFS management. The installation and maintenance of a network-wide YP system is a task which always requires a certain amount of discipline. In particular, linking into the NFS operation on existing machines is comparatively demanding. Thus, many networks (particularly small ones) do not use YP at all. Since there is a tendency towards even larger networks, it is to be expected that management costs without the use of YP will one day no loger be tolerable. At the least, the problem of overlapping user IDs must be resolved in some way (possibly manually), since otherwise the security of NFS cannot be guaranteed.

Although the YP system has proved itself overy many years, it is not uncontroversial, and rightly so. Its suitablity as a general naming and authentication service (in some ways similar to DNS) is not only limited by ist narrow reliance on UNIX specifics. Another factor is that the manner in which access rights are assigned in UNIX (and in the version of NFS described here) is insufficient for network operation. In addition, the implementation also includes a serious security loophole.

Better procedures, such as the Kerberos system (Miller et al., 1987) or secure NFS, are already in use in some places. The OSF (Open Software Foundation) in its Distrbuted Computing Environment (DCE) offering has included Kerberos and a service called PasswdEtc, which provides programs with distributed access to information normally kept in /etc/passwd. As previously mentioned, secure NFS is a component of the System V Release 4.

Ymodem is a popular file transfer protocol available in many off-the-shelf and shareware communications packages, as well as on many bulletin board systems (BBSs). Ymodem is a variation of the Xmodem protocal.

This protocol divides the data to be transmitted into blocks. Each block consists of the start-of-header character, a block number, 1 kilobyte of data, and a checksum. Ymodem also incorporates the capabilities to send multiple files in the same session and to abort file transfer during the transmission.

Ymodems's larger data block results in less overhead for error control than required by Xmodem; however, if the block must be retransmitted because the protocol detects an error, there is more data to resend. Also see Kemit; Xmodem; Zmodem.

Search for Information Technology Items

Return back to Network & Concepts Index

Networking "Y" Definition and Concepts

This site is brought to you by
Bob Betterton; 2001 - 2011.

This page was last updated on 09/18/2005
Copyright, RDB Prime Engineering