I apologize for the miss-send.
Hi all,
Here is my Mid-term update for the GSoC project "Map-Reduce support
for Apache Tuscany." I will briefly explain the progress I have made
so far, as well as the next steps I will take during the second phase
of the program.
Progress thus far:
I have spent time getting to know the Tuscany code base. I have gone
through the POJO sample implementation and have a good understanding
of what is needed in order to complete a component implementation.
I have spent time getting to know the Hadoop Code base. The current
Map-Reduce API supported by Hadoop uses a JAR file to pass the
necessary classes from node to node. This requires all the functions
to be put in a JAR file before the job is submitted. I have been
looking through the code base to see if there is a way to execute a
Map-Reduce job without a JAR file. I have not found a way so far. I
have also created two test functions that interact with Hadoop and
submit Map-Reduce jobs without the use of a shell script.
This first phase was marked by a large amount of reading documentation
as well as code.
Plans for the second phase:
Setting aside Hadoop for the moment, the first step will be to
represent the Map-Reduce programing model as a composite of java
components using a simple example. This will be a good step in
establishing how the model will be represented in Tuscany. I will then
be able to receive feedback on wether the design is simple enough to
make Map-Reduce programing easier, as well as flexible enough to
handle important use cases. Once an acceptable design is implemented,
I will work to create component types for each java component in the
composite. Once the basic types are implemented, I will revisit Hadoop
and start the integration with HDFS.
Finally, I would like to thank my mentor Jean-Sebastien, Luciano, and
Raymond, as well as the rest of the Tuscany community, for their
tremendous support during the first phase of the program.
Any questions or comments are always welcome.
Chris Trezzo
On Jul 12, 2008, at 12:30 AM, Chris Trezzo wrote:
Hi all,
Here is my Mid-term update for the GSoC project "Map-Reduce support
for Apache Tuscany." I will briefly explain the progress I have made
so far, as well as the next steps I will take during the second
portion of the program.
Progress thus far:
On Jul 9, 2008, at 9:25 AM, Luciano Resende wrote:
To help Mentors with mid-term evaluation, I'd like to suggest
students
to send a mid-term update describing high level
progress/acomplishments towards the original project proposal.
Mentors : Any thoughts on this ? other suggestions ?
--
Luciano Resende
Apache Tuscany Committer
http://people.apache.org/~lresende
http://lresende.blogspot.com/
