Tape Hits a Terabyte

One might think magnetic tape is fast becoming a technological fossil. But researchers at IBM Corp. disagree. They’ve come up with a way to store a terabyte (1000 GB) of uncompressed data on a standard tape, an accomplishment that represents an order of magnitude leap in storage capacity over the last generation.

The new technology uses more tracks and packs more data per millimeter of tape. IBM’s tape techies believe that it is possible to increase tape storage by still an order of magnitude more. But that will require overcoming some significant engineering hurdles. In the meantime, when it’s available commercially, the new tape technology will enable the medium to continue to beat the cost of storing large volumes of data on hard disks.

Six Orders of Magnitude in Linear Tape Technology: The One-Terabyte Project , by E. R. Childers, et al., IBM Journal of Research and Development , July 2003, pp. 471–82.

It’s been assumed that telecommuting, besides being attractive to employees, greatly reduces energy use and greenhouse gas emissions. Well, it’s not really that simple, say a pair of environmental engineers from the University of California, Berkeley. They found that telecommuting isn’t always as big a plus as one would expect. For instance, telecommuting in California reduces carbon dioxide emissions more in the summer than in the winter. The reason for the differences is that telecommuters often use energy they would not have used if they were at the office.

Workers, for example, may use more air conditioning or heating while working at home than if they had spent the day in their cubicles. There are also several appliances, such as stereos and TVs, which contribute to the extra energy consumed by teleworkers. While the data used for their report was based on emissions in the United States, their findings should hold for other countries as well.

Energy-Related Emissions from Telework , by E. Kitou and A. Horvath, Environmental Science and Technology , August 2003, pp. 3467–75.

Transistors are practically the definition of solid-state devices, but scientists at Bell Labs and the University of Illinois at Urbana Champaign recently fabricated a liquid-state version. The device consists of an organic semiconductor and mercury-filled microfluidic channels that act as electrode contacts to the transistor’s source and drain. Changing the amount of mercury in the channels dynamically tunes the transistor’s electrical characteristics by changing its effective gate length and width.

The researchers envision that these devices could serve as sensitive couplings between electrical and microfluidic systems, such as on-chip chemistry laboratories, because hydraulic events in the chip could easily control the mercury flow. The fluid transistor’s semiconductor, a film of the organic chemical pentacene, is mechanically flexible as well, opening up even more possibilities for unusual bendable microfluidic sensors.

Tunable Organic Transistors that Use Microfluidic Source and Drain Electrodes , by G. Maltezos, et al., Applied Physics Letters , 8 September 2003, pp. 2067–69.

Imagine that your bank account was accidentally reduced to zero, when a bug in electronic bill paying software wrote the amount you owe the power company in the spot that should contain your checking account balance. That’s just the kind of so-called wild write database error that a group of researchers from the United States and India is trying to prevent. Databases, financial and otherwise, are integrated more tightly than ever with third-party software applications that have direct access to the data. In effect, data corruption can now come from both database management software and outside software.

For this reason, the researchers have developed a suite of mechanisms to detect incorrect writes to a database and even to allow recovery from some errors. The detection uses a scheme for writing data that forms a code word, a value computed from a set of data that indicates unintended changes. From their research, the group believes that the error-detection and correction schemes they provide are fast enough to work in large systems.

Detection and Recovery Techniques for Database Corruption , by P. Bohannon, et al., IEEE Transactions on Knowledge and Data Engineering , September 2003, pp. 1120–36.