29-12-2009, 06:57 PM
transaction routing in distributed database systems
The transaction routing is a process in which a front- end router selects
an execution node for an incoming transaction.Since response time
estimation and decision making depend on the assumed transaction model
and parameters, it is important to examine the robustness or sensitivity
to the inaccuracy in the assumptions and parameter values.
Though it is robust with respect to change in the number of database calls
per transaction, we find through simulations that the dynamic routing
strategy based strictly on response time is too aggressive in sharing loads
and makes too many non-preferred system routings. Dynamic routing strategy
is relatively sensitive to the distribution of database calls. In this
article, two new refinements are proposed which improve system performance
as well as robustness of routing decisions.
Transaction Routing in Shared Disks Clusters
The cluster technology enables highly available real-time databases which
can be found at the heart of many telecommunication services. By using
internode parallelism, the cluster can reach high performance.
The two major objectives of transaction routing algorithms are:
1) Load balancing: This means to avoid the overloading of individual nodes.
2) Transaction a nity: This means to execute transactions with same or similar
data access patterns on the same node which is called a nity node.
We also have the transaction priority in order to support real time transactions. This involves considering the deadline as a priority and reducing the number of transactions missing their deadlines. The traditional algorithm for cluster allocation is called DACA(dynamic a nity cluster allocation.)
The transaction routing is a process in which a front- end router selects
an execution node for an incoming transaction.Since response time
estimation and decision making depend on the assumed transaction model
and parameters, it is important to examine the robustness or sensitivity
to the inaccuracy in the assumptions and parameter values.
Though it is robust with respect to change in the number of database calls
per transaction, we find through simulations that the dynamic routing
strategy based strictly on response time is too aggressive in sharing loads
and makes too many non-preferred system routings. Dynamic routing strategy
is relatively sensitive to the distribution of database calls. In this
article, two new refinements are proposed which improve system performance
as well as robustness of routing decisions.
Transaction Routing in Shared Disks Clusters
The cluster technology enables highly available real-time databases which
can be found at the heart of many telecommunication services. By using
internode parallelism, the cluster can reach high performance.
The two major objectives of transaction routing algorithms are:
1) Load balancing: This means to avoid the overloading of individual nodes.
2) Transaction a nity: This means to execute transactions with same or similar
data access patterns on the same node which is called a nity node.
We also have the transaction priority in order to support real time transactions. This involves considering the deadline as a priority and reducing the number of transactions missing their deadlines. The traditional algorithm for cluster allocation is called DACA(dynamic a nity cluster allocation.)