next up previous contents
Next: 4.1 Experimental setup Up: Operation-based Update Propagation in Previous: 3.5 Validation of re-executions

   
4. Evaluation


  
Figure 3: Selected tests and applications
\begin{figure*}
\begin{center}
\begin{scriptsize}
\begin{tabular}{\vert l\vert l...
... marked by bullet
points ($\bullet$ ) in the table.}}\end{center}\end{figure*}

At this time, we do not sufficiently understand client usage patterns to accurately model overall performance improvement. However, we have selected a set of commonly used non-interactive applications that allow us to focus on the following three questions:

1.
Is operation shipping transparent to the applications?
2.
What is the extent of network-traffic reduction that can be achieved by using operation shipping?
3.
What is the extent of elapsed-time reduction that can be achieved by using operation shipping?

We shall answer these questions in the following subsections, after we briefly describe the experimental setup.



 

Copyright 1999 by Y.W. Lee, K.S. Leung, and M. Satyanarayanan