A bank in Boston has decided to update its image by changing its logo. Diane, the graphic design artist, has designed the new logo and it now prints out on all her correspondence. Jeff, who uses the same server and same application to conduct his correspondence, still has the old logo on his work. They are in a client/server environment and share the same application. Diane changed the logo on her machine and on the server. Why do they have different logos?
The logo is stored locally because is usually changes very infrequently. Although Diane changed her default logo and the one on the server, Jeffs did not change because he was accessing the logo locally on his own PC.
In terms of a hybrid architecture, what does local autonomy mean? What are its pros and cons?
Local autonomy means that the local sites are responsible for the local data. It allows the local sites to control the security, integrity, and storage of local data. Absolute autonomy can pose a problem for a hybrid database because the various sites must cooperate in order for transparency to be feasible.
A trading company has offices in Caracas, Venezuela and Paris, France. The company wishes to use one global database and would like to continue to have relatively low communications costs. The database has many records that are accessed in only one site, but it has some records that are processed in both sites. What type of database architecture would you suggest and why?
A distributed database should be developed where the database is fragmented and those records pertaining to Venezuela should be stored in Venezuela and those that pertain to France should be stored in France. Since both offices occasionally update the records associated with the other office, the records that reside at the remote database used in the update may be replicated in both sites.
4.
Data can be stored and processed locally and remotely. Create a 2x2 table using these dimensions and place the four basic architectures into the table. Briefly describe each basic architecture.
Remote Job entry is when data are stored locally and sent over a communications link to a remote computer for processing and output is sent back. A personal database is when the user stores and processes data locally. Host/Terminal is when data are stored and processed remotely, typically on a mainframe. Client/Server usually processes data locally but stores it remotely.
5.
What is a hybrid architecture? Describe two fundamental principles that a hybrid architecture should satisfy.
It is a mix of data processing architectures that have evolved over time. There are 12 fundamental principles:
Transparency - the user sees the database system as a seamless system that readily provides needed information. The location of the data, its storage format, and access method should be invisible to the user.
No reliance on a central site - reliance on a central site can create bottlenecks and if that site fails the entire system fails. Those requests should not be routed through a central site.
Local autonomy - local sites are responsible for the local data. It is allows the local sites to control the security, integrity, and storage of local data. Absolute autonomy can pose a problem for a hybrid database because the various sites must cooperate in order for transparency to be feasible.
Continuous operation - the system must be accessible when required.
Distributed query processing - the time taken to execute a query should be independent of the location from which it was submitted. The system, not the user, should decide the most efficient way to process the query.
Distributed transaction processing - the system should ensure that a transaction is successfully executed for all sites. Partial updating of files can cause inconsistencies.
Fragmentation independence - any table can be broken into fragments and then stored in separate locations. The table should be able to be broken down by field or by record.
Replication independence - the replication of fragmented data that has been copied and physically stored at multiple sites happens behind the scenes. The user should not need to be aware of this replication. Replication can be done either synchronously or asynchronously.
Hardware independence - a hybrid architecture should support hardware from multiple vendors without affecting the user.
Operating system independence - a hybrid architecture should support a range of operating systems on a variety of hardware.
Network independence - a hybrid architecture should support a range of network configurations
DBMS independence - a hybrid architecture should support an organizational standard that allows for a variety of DBMSs to work together.
6.
What type of architecture is most used for large scale transaction processing
Answer: c. Host/terminal
7.
When using a hybrid architecture what does it mean for the data to be transparent?
Transparency in this context means that the location of the data, its storage format, and access method are invisible to the user. That is, the systems use should be independent of the location of the data.
8.
Why is it so important for hybrid architectures to be in continuous operation?
People who are geographically distributed and may reside in various time zones often use hybrid architectures. Therefore, business hours for one user may be in the middle of the night for other users.
9.
What are a few key parameters to consider when comparing host/terminal with client/server architectures?
10.
What are the two main reasons a hybrid architecture should not rely on a central site?
A central site has two major problems. First, it can create a bottleneck for the system because all requests are routed through it. Second, it is too vulnerable to disasters (natural and man-made).
11.
A database is slowing down, and you decide you can help with its access time by vertically and horizontally partitioning the customer table. You have offices in Chicago, Calgary, Paris and Madrid. Although you can have customers from all over the world most of them are from one of these four locations. Each customer has contact information, date of birth, preferences (five different fields entered in by customers when they initially register with the site) credit information, a rating assigned to them by the credit industry, a credit limit, and a credit card number). When a customer orders something the firm needs to access the credit and contact information. Once a month the firm runs a birthday agent that sends an e-mail to all customers giving them a special offer that is determined by their preferences. How would you partition these tables?
Answer: You could vertically partition into two tables one that has the customers contact information and credit information,and the other having customers, name, email address, date of birth, and preferences. You could horizontally partition t by location, putting all customers from one country (note that this is the criteria to use not city) into one of the four sites (dependent on proximity).
This page is part of the promotional and support
material for Data Management (open edition) by Richard T. Watson For questions and comments please contact the author |