Hi All, We had very good outcomes from many of the architecture mailing lists [1]. How about we formally evaluate some of the suggestions we constantly brainstorm in a Airavata Labs repository? We currently use sandbox repository to put code which do not relate directly to the main repo, but they quikcly gets orphaned. Instead, how about we create a persistent labs repo which will be maintained and any research projects can be performed on this experimental setup.
An example would be, we can emulate most of Airavata components and do experiments on what will be an ideal registry implementation. Similar will be to do a academic study of building up on Airavata’s internal implementation for distributed service management (based on zookeeper), or port the current handler based architecture to a Storm or YARN based architecture? Currently we debate on these in mailing list, but it will be good if we can have some emulators to do these experiments and arrive at answers in a more systematic way. Since repository creation is a reversible task, I would like to assume lazy consensus and proceed in requesting a new repo, please voice any objections? Suresh [1] - http://airavata.apache.org/community/mailing-lists.html <http://airavata.apache.org/community/mailing-lists.html>
