James Nichola Gray

Research and Development

Gray was 1944, a Ph.D. in the University of California University Berkeley Branch. Its doctoral thesis is the theory of analytical analysis of priority. After learning, he worked in Bell Lab, IBM, TANDEM, DEC and other companies, and research directions to the database field.

During IBM, he participated in and hosted the development of IMS, SYSTEM R, SQL / DS, DB2 and other projects, which except that SYSTEM R used as a research prototype, did not become a product, and several other IBM has an influential product on the database market.

During TANDEM, Gray has improved and expanded the company's main database product EnCompass, and participated in the development of system dictionary, parallel sort, distributed SQL, NonStop SQL and other projects.

in DEC, he is still responsible for the technology of database products. When Gear enters the database, the basic theory of relational database has matured, but the major companies have encountered a series of technical issues in realization and product development of the relational database management system (RDBMS), mainly in the scale of the database. Big, the structure of the database is getting more and more complicated, and how to protect the integrity, security, parallelism (ConcURRENCY), and failure in the case of integrity, parallelism The database is how to recover from the fault (Recovery). These issues can not be accepted by users if they cannot be successfully resolved, no matter which company's database products cannot enter practical use. It is in the process of solving these major technical issues, making DBMS matures and smoothly enters the market, Gray played a very critical role in his intelligence.

Each DBMS solves the main technical means and methods of the above problems as follows:

1. The operation of the database is divided into a "transaction" (or "transaction element", transaction) one atom. The transaction is the basic execution unit of Transaction Process, that is, the operation in a transaction is either executed, or all do not execute, that is, the principle of so-called All or None. A transaction typically begins with a "start" statement, first removes some data from the database, and then performs the required processing, and finally ends with the "Submit" statement (CommIT). If an exception occurs in the transaction, use the "Avantly Termination" statement (ABORT) or "Rollback" statement (ROLLBACK) to revoke all updates (ie, so-called undo), restore the database to the transaction The correct state is to protect the integrity, consistency of data.

2. When the user issues an operation request for the database, the system is "lock" (Locking) of Gran Guity (Locking), and the locking data will be temporarily disabled from accessing other users. Here is only a simplified interpretation. In fact, according to the different nature of the user's data request, the lock data is to treat another user's request, which is complex, for example, if the locking data will be modified, then It is absolutely disabled from other users to access; and if the locking data is only used for read, other users' read requests will also be allowed. This is managed by the so-called "lock compatibility matrix" --Lock Compatibility Matrix and control). Unlocking after the operation is complete. This mechanism is used to maintain "concurrency" between the transactions and guarantee the "integrity" of the data.

3. Establish System Run Logs (LOG), record the initiation points, end points, and reform status of the page updated in the transaction, so that the database is broken when the system is broken. , Can restore the database to the correct state before the database is added to the database periodically or not periodically for the database, and the last backup can be reserved for the revision of the database.

4. Any update of the database is submitted two-phase (Two-Phase Commital). This is especially important in distributed systems based on a transaction, may also involve two different database systems.

The above and other methods can be generally referred to as "Transaction Processing Technique). Gray has created thinking and pioneering work in transactional technology, making him a recognized authority in the field of technology. His research results are reflected in a series of papers and research reports, the last crystallization is a thick monograph Transaction Processing: Concepts and Techniquhers, 1993, and another author is A. University of Stuttgart, Germany. Professor Reuter). Although the transaction technology is born in database research, data management and communication in the Client / Server structure, which is equally important for fault-tolerant and high reliability systems.


In order to fully realize the ideal of the above three scientific giants, Gray calls on the US government to pay attention to support for long-term research on IT technology, and believe that its significance is no less than 200 years ago. Jefferson (Thomas Jefferson, 1743-1826, "Independent Declaration" drafters, US third president, 1801-1809 in place) Decided to buy Louis Anna Terminal from the French government in his hands in his hand (Louisiana Territory, this is Located between the Mississippi River and the Rocky Mountains, the large-scale land of the Northern Canada, the South Dames, the area of ​​2,070,000km2) This is a famous historical event called Louisiana Purchase, and then dispatched to Captain, Captain Meriwether Lewis. "Corps for Discovery", "Corps for Discovery", to the Western Adventure until the Pacific Coast, laid the foundation for the final formation of the United States today. Gray believes that a good IT long-term goal should have the following five key Sex:

1. Apprecient objectives should be simple to express and have been understood.

2. How to achieve the goal is not very obvious.

3. The use is widely used for computer scientists, but is useful to most people.

4. Testability to check the progress of the project and know if the goal is reached.

5 There are several milestones in progressive in progress to check the progress of the project and encourage researchers to dry.


Under the above argument, the long-term study of several IT technology proposed by Gray The goal is as follows:

1. Scale scalability.

2. Through Tulex test.

3. Voice to text Conversion (Speech TO text).

4. Text to speech.

5. Machine vision, can identify objects and exercise like people.

< P> 6. Personal "Mei Michax", can record everything you can see and hear, quickly retrieve it quickly.

7. The "Mei Michex" of the world, ie Establish text, music, image, art, film "Corpus", answering any questions, as fast as human experts, doing abstracts.

8. Virtual reality (Gray used the word Telepresenee, see the introduction of the 1969 Tu Leon winner Minsky).

9. Trouble-free systems.

10. Safety System (SEC URE systems.

11. High availability system (ALWAYSUP).

12. Automatic Programming.

Another copy of Jim Gray Gray is the Benchmark Handbook: for Database and Transaction Processing Systems, 1st edition published in 1993, It is also published by Morgan Kaufmann Press. Gray is also the editor of the "Data Management System Series" of the Press.

Gray is very active in the database academia. Some important database academic conferences held by the international regular or irregular period, such as VLDB, Sigmod, can see his figure and hear his voice. In addition to the company's research and development, he also partred in the University of Burkeley, Stanford University, and Budapest's abroad teaching and lectures. In 1992, the VLDB Journal was published, he was edited.

Gret is one of the 6 winners of the Software System Reated Software System R in 1988, the other 5 people are Donald Chamberlin, Raymond Lorie, Gianfranco Putzolu, Patricia Selinger and Irving Traiger. It is because of Gray in the reputation of database technology, "Giant Mac" Microsoft, in the software industry, decided to enter the large-scale relational database market in 1993, and did not cherish Gray from DEC. Because Gear doesn't like the rainy Seattle of Microsoft Headquarters, I will stay in the sunny San Francisco, Microsoft specially opened up the second Microsoft Research Institute Bay Area Research Center Barc (Bay Area Research Center), and arranged the Director of the Institute of Grass. . Gray did not expect, leading a development team to develop MS SQL Server 7.0, becoming a milestone version in Microsoft, and also became the leader in today's relational database market.

Gret is accepted on the ACM National Conference held at Atlanta on May 4, 1999. Gray issued the speech of "Information Technology's Goal" (WHAT? A dozen Remaining It Problems, in the case of several directional issues related to information technology development. Later, after the cultural passbook, on Sigmod's will be published again with what next? A dozen it research Goals. Gray Spelling In the development of computing technology, the Barles Babage, 1791-1871, in the 19th century, is basically realized today; American mathematician Bush (Vannevar) Bush, 1890-1974, scientific adviser of Roosevelt President, the "Memory" MEMEX "Memory Extender" envisioned in the 1940s is currently close to realization; the intelligence proposed by Tuling There is also a distance from the machine to the implementation, and the computer is difficult to pass the "Tulex Test".


Jim Gray But this life is committed to losing contact with everyone after driving a ship in the sea. After several days of search, the US Coast Guard is still nothing.

On January 28, 2007, Gray drove a 40-foot yacht, and arrived at 25 miles from the San Francisco Golden Gate Bridge. He told his family that he had to sprinkle ash for a 97-year-old mother who passed away and he intended to return it to the night.

10 o'clock in the morning, his wife and daughter received his phone, and everything was as usual, he gave admired that the weather was very good, there were dolphins to swim around the boat, and there were several He will not receive the phone signal.

then, Gray disappeared. By the 31st, the coastal guards rushed along the Pacific Ocean, searched from Monteree Bay to Oregon, the sea area, neither Gray, and did not find his boat. What did the rescue personnel think about it? What did Gear encountered?

It is said that the weather is very good on the 28th, the sea is small, and the visibility is excellent. Gray's flight is only 60 miles. He started from a commercial canal. After passing through the Golden Gate Bridge, it was a wide sea surface, and it took only one or two to arrive at the destination.

If he suddenly has a health problem on the road, such as heart disease, his ship should also be discovered.

If the yacht has encountered underwater reef or with the whale collision, Gray should have enough time to send a rescue signal, as long as there is a ship nearly 22 to 23 miles, It can be accepted. However, Gray disappeared, did not leave any clues.

Related Articles