Smart Computing Article - TrackPoint to transfer rate

Subscribe Today | Contact Us | Register Now

Home | Tech Support | Q&A Board | Article Search | Subscribe & Shop

TrackPoint to transfer rate

Email This

Print This
View My Personal Library

T April 2002 • Vol.6 Issue 4	Add To My Personal Library
TrackPoint to transfer rate

TrackPoint
tractor feed
trade shows
traffic
Traf-O-Data
transaction
transaction processing
transaction server
transcoding
transfer rate

TrackPoint

This IBM keyboard has a TrackPoint device.

TrackPoint

A TrackPoint is a cursor control device. It is a very short joystick that is positioned in the middle of the keyboard. As a result, it takes up virtually no extra space on the keyboard or work area, and it can be used without taking your hands off the keyboard. To move the cursor, you apply pressure to the TrackPoint in the direction you wish to move the cursor. Two buttons are needed on the bottom of the keyboard to give the TrackPoint all the functionality of a mouse.

The TrackPoint was invented by Steve G. Steinberg at IBM and first appeared on IBM ThinkPad notebook computers. The TrackPoint was introduced before touchpads became almost universal on notebook computers, so notebook users were still struggling with external mice as pointing devices. The TrackPoint proved to be so popular with notebook users that IBM tripled its notebook production every month for four months straight after it introduced the TrackPoint. Two other IBM notebooks that were introduced around the same time without a TrackPoint had to be discontinued due to poor sales.

In addition to IBM, the TrackPoint can be found on some Toshiba notebook computers. Although not common, keyboards are available for desktop computers with a TrackPoint built into them.

tractor feed

Tractor feed, also known as pin feed and sprocket feed, is an older method of feeding paper through a printer. A tractor feed printer requires special paper called continuous form paper, which is folded back on itself and, therefore, also called fanfold paper. Tractor feed is found mainly on older dot matrix printers and large, high-speed printers.

Tractor feed printers have a wheel on either side of the paper handling mechanism. Each of these wheels has a number of sprockets or teeth along its outside edge. Continuous form paper has holes on either side that match up to these sprockets. When the printer is printing, the wheels rotate and, as they rotate, the sprockets fit into the holes in the paper and cause the paper to feed through the printer.

In most printers, the wheels are positioned at the end of the print mechanism so they pull the paper through the printer. A few printers have the wheels at the beginning of the print mechanism so they push the paper through the printer.

Some of the paper designed for these printers had perforations along either side so that the small strips with the holes could be torn off after printing. These strips are known as chad. More expensive paper was produced with very small perforations, called microperf, so the paper looked like standard typing paper after the chad was torn off.

When tractor feed printers were commonly used, vendors sold letterhead stationary, custom printed forms, and even envelopes in continuous form. Tractor feed printers can be seen today in many offices where dot matrix printers are still used for their ability to print carbon copy forms.

trade show
Trade shows, such as the 2001 COMDEX Fall tradeshow in Las Vegas, let customers interact with vendors.

trade shows

A trade show is a collection of exhibits by sellers to an industry that is visited by buyers in that industry. Manufacturers and sellers benefit because they can reach a large number of customers at one time in one location. In fact, many manufacturers use their trade show appearance to announce forthcoming products and introduce new products. Buyers benefit because they get to review the products from a large number of venders without traveling to numerous locations and entertaining numerous sales calls by sellers. Trade shows are also well attended by the media who cover that industry. The general media may also cover very large trade shows.

All major industries have trade shows, and the computer industry is no exception. The two largest are COMDEX and PC Expo. In the fall of 1979, a company called The Interface Group started COMDEX, which stands for Computer Dealers Exposition. The first COMDEX had 157 exhibitors and 4,000 attendees. The Fall 2001 COMDEX had more than 2,000 exhibitors and more than 200,000 attendees.

There is both a fall and a spring COMDEX. The Spring COMDEX, started in 1981, is about half the size of the fall COMDEX. The Fall COMDEX is held in Las Vegas. In 2002, the Spring COMDEX will be held in Atlanta.

The other large computer trade show is PC Expo, which was founded in 1983. The first PC Expo had 120 exhibitors and just fewer than 10,000 attendees. The 2001 PC Expo in New York had about 500 exhibitors and almost 50,000 attendees. PC Expo is held in New York City in the summer and in Chicago in the fall.

Unlike other trade shows, in addition to industry buyers, sellers, and media, end users also regularly attend computer trades shows. Computer trade shows are especially well attended by corporate IT (information technology) personnel who are looking to understand emerging technologies and plan for future purchases.
Back to top

traffic

Traffic is the load on a network or other communications system. Traffic is a very general term that can refer to almost any type of information flowing across almost any type of network. The heavier the traffic, the greater the demands that are placed on the network. One of the main tasks of the system administrator is to monitor the level of traffic on the network and take the appropriate action when the traffic becomes too heavy. This is known as traffic engineering.

Traffic on the Internet is growing at over 100% per year. To keep up, ISPs (Internet service providers) are constantly installing new fiber optics and network routers. However, traffic is growing so rapidly that ISPs cannot keep up with the additional traffic. To help, ISPs are turning to traffic management tools such as caches and load balancing.

One way an ISP can speed up Web access for its customers is to temporarily store frequently accessed Web pages locally. Customers can then retrieve the pages from a local site rather than the request having to cross the entire Internet. This temporary, local storage of Web pages is called a cache. The largest caches can store up to 25 million Web pages at once.

Load balancing, widely used at the 1996 Atlanta Olympics to help manage the amount of traffic the Web site received, uses software to split tasks among multiple computers so each one can respond efficiently.

As the Internet becomes more and more congested, tools that monitor and manage traffic will become even more critical.

Back to top

Traf-O-Data

In 1971, a company called Logic Simulation was inventorying the streets in Kent, Wash. They placed traffic counting boxes, which had a hose going from the box across the street, on some streets. Each time a car crossed the hose, the box increased its count. The results were recorded as holes punched into paper tape, one of the main computer storage media of the time.

Bill Gates, a high school student, got a job processing the data. He was to count the holes in the paper tape and produce a report. He also transcribed the results onto punch cards to be transferred into another computer. Rather than doing all the work himself, Gates hired fellow students at Lakeside School.

One of these fellow students was Paul Allen. In 1972, Gates and Allen purchased an Intel 8008 chip for building an automated car-counting machine. They formed a company that Gates would call Traf-O-Data. Its headquarters was Allen's dorm room at Washington State University. Paul Gilbert, a friend of Gates and Allen, designed the hardware while Gates and Allen wrote the software. Because the hardware did not yet exist, they tried to write a program on the Washington State University IBM System 360 to simulate the 8008 chip. That way, they could program on the IBM System 360 and transfer the program to the Traf-O-Data device when it was finished.

Around this time, defense contractor TRW was looking for programmers. Both Gates and Allen were hired. While there, Allen developed the 8008 simulator to run on the TRW PDP-10 computers. This allowed Gates and Allen to begin work on the Traf-O-Data software.

After leaving TRW, Gates was accepted into Harvard University. With everything going on in his life, Traf-O-Data barely limped along. By 1974, Gilbert had finished the hardware, and Gates had mostly finished the software. However, the company had no device to automatically read the paper tapes.

An acquaintance of Gates' father came up with a solution. The device would feel the paper tape with specialized metallic fingers to count the holes. It became known as the squeeze reader. A prototype worked well and a representative of the King County Engineering Department was invited to Gate's parents' home to see a demonstration. The squeeze reader failed completely, and no sale was made.

Traf-O-Data, now a partnership between Gates, Allen, and Gilbert, needed a professional paper tape reader. Gates used $3,400 of his own money to purchase an Enviro-Labs Model GS-311 Paper Tape Reader. Traf-O-Data used the device to process the paper tapes generated by traffic counting boxes and produced some revenue. Eventually, the State of Washington offered free traffic processing services, ending this revenue. In May of 1979, Traf-O-Data sent letters to clients saying they were suspending business. Gates' first business had failed.

Much of the details of Traf-O-Data are lost. This is to be expected since the business was started and run by teen-agers. Gates has claimed that Traf-O-Data had revenues of $20,000 to $30,000 per year over its life. More reliable estimates put the amount closer to $10,000 to $20,000 total.

Some timelines report that Traf-O-Data was renamed Microsoft. This is incorrect. This fallacy is based, no doubt, on the fact that Gates and Allen were involved in starting both Traf-O-Data and Microsoft. Although Traf-O-Data did not become Microsoft, it is no doubt true that the experiences Gates and Allen had with Traf-O-Data helped both of them when they started Microsoft.

Back to top

transaction

In computer programming, a transaction is a sequence of activities that is treated as a single unit for the purpose of both satisfying a user request and for maintaining the integrity of the information in a database.

Transactions have four characteristics:

� Atomicity. A transaction needs to be either completed in its entirety or completely undone.

�Consistency. A transaction needs to preserve all the invariant properties defined on the data. These include all the error checking that is built into the database.

�Isolation. Each transaction needs to be completed independently of any other pending transaction. That is, the effect of running transactions concurrently should be exactly the same as running them one at a time.

�Durability. The results of a completed transaction must always
be persistent.

Taken together, these properties are often referred to as ACID. Their overall effect is to ensure the integrity of the databases as they are acted on by various transactions.

Transactions fall into two broad categories: inquiry and file maintenance. With an inquiry transaction, information is requested from the database but no changes are written back to the database. In other words, the database itself is not modified by an inquiry transaction. An inquiry transaction can be simple, such as a library checking to see if a book is on the shelf, or complex, such as a personnel officer calculating how much of a pension an employee will receive if he works for another 10 years.

With a file maintenance transaction, the contents of one or more database are permanently altered by the transaction. The alteration might be a change in an existing value, the addition of a new record, or the deletion of an existing record. It is not uncommon for a file maintenance transaction to alter the contents of multiple databases and multiple fields within those databases. For example, an online purchase modifies the inventory, accounts receivable, customer, and sales databases.

It is critical that all aspects of the file maintenance transaction take place. That is, systems must be in place to ensure that the information in the databases is permanently altered if, and only if, the file maintenance transaction can be completed in its entirety.

Back to top

transaction processing
The steps taken when a customer places an order online represent a transaction process.

transaction processing

In the broadest sense, transaction processing describes the steps taken by one or more computers in response to an external stimulation, such as a customer placing an order. Transaction processing is made up of data, usually stored in a database, and commands to alter that data. The commands to alter the data are often referred to as transaction code.

Transaction processing is a series of one or more actions that are performed on one or more databases. Because all of the actions must be completed in order for results to be meaningful, a transaction is often referred to as a unit of work. With transaction processing, transactions are completed as soon as they are entered into the computer. Withdraw money from an ATM machine, and your account balance will be instantly reduced by the amount you withdraw. Because results are reported back to the user immediately, transaction processing requires fast and powerful computers.

Not only are powerful computers required, a very high degree of reliability is also required. If the transaction processing system goes down, it is very likely that the business goes down as well. Imagine a bank with ATM machines that cannot verify that you have adequate funds for a withdrawal in your account due to no transaction processing. These ATM machines would also not be able to prevent a customer from making multiple withdrawals at different ATM machines and far exceeding both his daily withdrawal limit and his available balance. Clearly, the bank's business would suffer.

With transaction processing, the state of the database is constantly changing. That is, the data is never static. As a result, it can be difficult to develop a historical perspective unless efforts are made to sample the database periodically and save that information for historical analysis.

Another way to process transactions is to accumulate them and process them in large batches. A bank processing all its deposits at night is an example of this batch processing. Batch processing requires less computational power and will generally not cause a business to go down if the processing goes down.

Transactions may also be completed in hybrid mode, where the processing has characteristics of both transaction and batch processing. For example, in a university online registration system, the student may interact with the system in real-time to select classes. The student's class selections may then be stored in a file, called a shadow file, to later be applied to the class roll database and the financial aid database. In other words, the class rolls and financial aid information may not be updated until a batch process is run that night.

Hybrid processing lets the user get a rapid response while minimizing the processing power required to support the system since many of the changes are not run in real-time. The drawback to this approach is that delaying the updating of some of the databases can result in inappropriate actions. In the university registration system, for example, if the class rolls are only updated once a night, the system may allow enrollment in popular classes to exceed the maximum allowable number of students.

A major problem for transaction processing is making sure that all of the main files are updated correctly. This is often called transaction atomicity. For example, consider a withdrawal at an ATM machine. Some of the steps that are taken include the ATM machine verifying that the customer has enough money in his account, the ATM machine verifying that it has enough cash on hand, the customer's account being reduced by the requested amount, the ATM's cash on hand being reduced by the requested amount, the money being issued to the customer.

If the process fails between steps 4 and 5, the ATM does not issue the cash so the customer does not receive the cash. Both the ATM's cash on hand account and the customer's checking account need to be adjusted to reflect that the transaction failed. This is complex enough with one customer and one ATM. Imagine what happens when you have more than one person withdrawing money from that same account.

Now, imagine a popular Web-based business with a database that has hundreds or even thousands of transactions pending at once. All of these transactions are pending against the same database. In processing all of these transactions, the system must keep all the changes to the database straight so the database can be returned to the appropriate value if any combination of the transaction processes fails.
Back to top

transaction server

The simplest forms of servers are disk and file servers. These servers basically let clients share resources, such as data files and peripherals, across a network. The more advanced forms of servers are application servers, which are set up to run a program (application) for the user that needs to be run remotely, such as e-mail; database servers, which process large amounts of corporate data in a central location; and transaction servers.

With a transaction server, the client invokes a procedure on the transaction server to execute a transaction. That transaction is made up of a series of SQL (Structured Query Language) statements that either all succeed or all fail as a single unit. Essentially, a transaction server is a database server with centralized SQL statements. Running procedures stored on the transaction server is called OLTP (online transaction processing).

OLTP applications tend to be mission critical applications that require a rapid response time with a high degree of control over the integrity and security of the underlying databases. Having the procedures on the transaction server enhances both speed and database integrity and security.

Because the procedures are stored on the transaction server, communication overhead is reduced, and, therefore, speed is enhanced since only simple requests and replies, and not the actual SQL statements, are sent between the server and the client. Having the database and SQL maintained on the server, rather than having multiple SQL statements in multiple database servers, centralizes maintenance and reduces the likelihood of problems.

In the mainframe days, transaction servers were called transaction monitors and were standalone hardware. However, it is becoming rare to see a hardware transaction server sold as a separate product. Today, the transaction server is either part of the server's operating system, or it is part of the middleware, the software that manages the interaction between the client and the server.

Back to top

transcoding
Using a transaction server, a Web designer only has to write one version of a Web page. The transcoding server modifies the Web page to fit the requirements of the device accessing the content.

transcoding

The broadcast industry first used the term transcoding to refer to converting a signal, such as a television program, from one format to another. It was not uncommon to convert a program originally recorded using the American NTSC (National Television System Committee) format to the European PAL (Phase Alternating Line) format. Although the signal needed to be transcoded, the content was not to be altered.

In the computer industry, the term was originally used to describe the operation of converting character data between different character sets. However, recent changes in the way Web pages are viewed has caused the term to take on a greatly expanded meaning.

Back in the old days when Web content was usually displayed in Netscape Navigator, you could write one version of your Web content, and you were done. Once Internet Explorer became popular, you had to worry about the differences between the two in displaying Web content, but those differences were usually minor.

Fast-forward to today. Web content has expanded well beyond a browser running on a PC. Today, you can expect that your Web content may be viewed on WebTV, a PDA (personal digital assistant), or even a cellular phone.

Web content is generally designed to take advantage of the abilities of advanced browsers to display information on large color monitors attached to powerful computers with sound systems. Web content includes not only the text and images you see in the browser itself but also sound, video, Java, JavaScript, XML (Extensible Markup Language), and much more. This complex array of content does not display effectively on the tiny screens available in a PDA or cell phone. Some of these devices have little or no ability to display graphics, some of them use only black and white, and some of them require a different language, such as WML (Wireless Markup Language). Additionally, the users of mobile devices want to access critical data, such as a stock quote, without all the extra details and graphics found on a typical Web page.

Transcoding is the process of adapting content intended for display in a Web browser for display by other devices with limited ability that are operating in a non-PC environment. Transcoding lets Web developers create content once and deliver it anywhere on any type of display device. Because most of the alternative display devices are mobile devices, transcoding a Web site is often referred to as mobilizing it.

The benefits of transcoding are not limited to supporting wireless phones and PDAs or supporting new devices as soon as they emerge. Transcoding can also be used to take large pages and scale them down for users accessing the site with a modem. This can improve the Web experience for all users by letting a site have complex pages for users with fast access and simpler pages for users with slower access. In the future, transcoding may also be used to convert a Web page to speech.

Terms associated with transcoding:

static transcoding: With static content, the designer produces different versions for display on different devices. This can be done either manually or by using software packages to aid in the translation of content from one device to another device. With dynamic content, either different versions of the delivery application are used to provide the information to each device or the application contains programming logic that controls what information is sent and how it is formatted depending on the device requesting the information. Its main advantage with either type of content is that a designer has the final say on the appearance of each page so this approach gives the best looking results. Its main disadvantage, at least for large sites, is the large number of pages that must be converted, tested, debugged, and maintained.

dynamic transcoding: With dynamic transcoding, the content and its format are altered at delivery time based on the presentation requirements of each user and their display device. A transcoding server, such as IBM's WebSphere Transcoding Publisher, can handle dynamic transcoding. Its main advantage is that it is done on the fly without modifying the original source code so updates are easier. Its main drawback is the need to maintain an additional server and the cost of modifying that server as new devices emerge. Hosting transcoding services are available from companies such as Everypath for firms that wish to outsource the transcoding to another company.
Back to top

transfer rate

The transfer rate is the speed at which information or instructions can be moved between locations. Transfer rate can refer to a communication device, memory, or a hard drive.

With a communication device, such as a modem or USB (Universal Serial Bus) port, transfer rate is a measure of the amount of information that can be moved over the device during a period of time. Transfer rate is often measured in megabits (millions of bits) per second (abbreviated Mbps) or megabytes (millions of bytes) per second (abbreviated MBps). Other measures include bits per second (abbreviated bps) and characters per second (abbreviated cps).

For a communications device, transfer rate can be measured in two different ways, either as a raw rate or as an average rate. The average rate includes the gaps between blocks of data as part of the transmission time while raw rate does not. The average rate is closer to what the user will see than is the raw rate.

With memory, transfer rate refers to how quickly the memory can transfer information to the CPU. This is usually measured in megabytes per second; however, newer memory is faster and uses gigabytes per second (GBps) for measuring transfer rates. The current memory standard is SDRAM (synchronous dynamic RAM), which has a peak transfer rate of around 800MBps and an effective transfer rate of around 600MBps. Newer DDR SDRAM (double-data-rate SDRAM) has a peak transfer rate of 2.1GBps and an effective rate of 1.37GBps. The latest RDRAM (Rambus dynamic RAM) has a 1.6GBps transfer rate and an effective rate of 1.36GBps.

Hard drives. With a hard drive, transfer rate refers to how quickly the hard drive can transfer information either internally or to the CPU. It is common to refer to transfer rates for new hard drives as a way of summarizing the speed of different drives. However, there are several different kinds of transfer rates and they do not mean the same thing, so you must be careful:

�Internal transfer rate. This is the speed at which the drive can move data between the drive head and the spinning drive platter. This is the limiting factor on the speed of the drive.

�Burst transfer rate. This is the maximum speed of the interface between the hard drive and the computer. This will be the highest of the transfer rates so it is the one that is usually quoted in hard drive specifications.

�Sustained transfer rate. This is the speed at which the hard drive can supply data to the computer over a long period of time. This is the most meaningful of the transfer rates. It is equal to the internal transfer rate less whatever time the hard drive needs for internal overhead during the transfer.

Hard drive transfer rates are sometimes given as a set of values containing a minimum, maximum, and average sustained rate. The difference comes from the fact that different tracks on the hard drive contain different amounts of data (the outer tracks have more) so the drive head can transfer more data per rotation when reading from an outer track than from an inner track. So, the maximum transfer rate would be for the outer track, the minimum for the inner track, and the average would be for all the tracks.

Back to top

Want more information about a topic you found of interest while reading this article? Type a word or phrase that identifies the topic and click "Search" to find relevant articles from within our editorial database.

Home Copyright & Legal Information Privacy Policy Site Map Contact Us