Comment

Comment by 刘灵燕 on November 22, 2011 at 6:52pm

A outline of transaction

catalogy

 1.      Definition

 2.      Functions

 3.      The ACID properties

   a)      Atomicity

   b)      Consistent

   c)      Isolation

   d)      Durability

 4.      The implements of ACID

   Atomicity

   Consistency

   Isolation

   Durability

5.      Transactions and models

    The flat transaction

    The distributed(or global ) transaction

    Nested transaction

    Multilevel transaction

6.      Transaction processing systems(TPS)

    Transaction processing

    TP monitor

 

 

1.      Definition:

    In computer programming, a transaction usually means a sequence of information exchange and related work (such as database updating) that is treated as a unit for the purposes of satisfying a request and for ensuring database integrity.

2.      Functions:

1:  It can update a database to reflect the occurrence of a real-world event that affacts the states of the enterprise the database is modeling.

2:  it can ensure that one or more real-world evec=nts occur.

3:  it can return information derived from the database

3.      The ACID properties:

a)      Atomicity:

      The system must ensure that the rather the transaction runs to completion or , if it does not completed, it has no effects at all(as it had never been started).

      How a transaction maintain its atomicity?

      When a transaction completes successfully, and the system agrees to preserve its effects, database changes are said to be commit;

      when a transaction does not complete successfully, we say that it has aborted, and the system must ensure the that whatever changes the transaction had made should be rolled back.

b)     Consistent:

      The execution of each transaction in isolation maintains database consistency and moves it to a new state of the enterprise.

      There are two kinds of consistencies:

1: the database must satisfy all integrity constrains.

2:the database must model the state of the real-world enterprise.

c)      Isolation:

      Even though transactions are expected concurrently, the overall effects of the schedule is the same as if the transaction had executed serially in some order.

It’s usually implemented by requiring transactions to obtain locks on the database item they access.

d)     Durability:

      The system must ensure that once the transaction commits, its effects remain in the database even if the computer or the medium on which the database is stored subsequently fails.

Durability can be achieved by storing data redundantly on different backup devices.

It draw forth “transaction schedule”.

 

4.      The implements of ACID:

Atomicity:

       The main operation is commit, abort, rollback

Consistency:

      It is the transaction designer’s job to design consistent transactions.

Isolation:

      This item includes: Schedule ,Serializability and lock.

        a:Schedule :

       A sequences of instructions that specify the chronological order in which instructions of concurrent transactions are executed

       b:Serializability:

      Aschedule is serializable if it is equivalent to a serial schedule.  Different forms of schedule equivalence give rise to the notions of:

      c:Lock:

      Include deadlock, read lock, wait lock, two-phase lock,lock aranularity,

Durability:

       In this part we mainly introduce Recoverability.

5.      Transactions and models:

      To deal with long and complex transactions, many transaction processing systems provide mechanisms for imposing some structure on transactions or breaking up a single task into several related transactions. But how the mechanisms can be implemented?

      The flat transaction:

It involves a database on a single server, and has no internal structure.  

      The distributed(or global ) transaction:

       It can be viewd as a tree. It’s a transaction in legacy system starts at some network site and might access several of the transaction processing systems by initiating the legacy transactions that they export.

     There are two typical model to descript the distributed transaction: hierarchical model and Peer model .

      Nested transaction

      A nested transaction occurs when a new transaction is started by an instruction that is already inside an existing transaction. The new transaction is said to be nested within the existing transaction, hence the term.

     multilevel transaction :

      It’s similar to the distributed and nested transaction, however the motivation for a multilevel transaction is increased performance, it can allow more concurrency in the execution of independent transactions.

6.            Transaction processing systems(TPS):

      Definition : a transaction processing system generally consists of a TP monitor, one or more DBMS(database management systems), and a set of application programs containing transactions.

      There are two important parts of TPS:

      Transaction processing :

      Introduction of the Transaction processing in

1：in a centralized system；such as a single-user system, centralized multiuser system.

2：in a distributed systems; such as sessions and context, queued Transaction processing.

      TP monitor

      It a collection of off-the shelf middleware components that can be used to build a TPS.

      It can provide services such as communication, global atomicity and isolation, load balancing and routing, recoverable queues, security services, threading, nested transaction and so on.