Hi Danushka, Suresh, Sorry, I missed this email earlier. Both the ideas sound interesting. I would have definitely helped with the Ozzie idea. I can still try to help with the second idea as well, but I feel we may be able to find a better co-mentor for that.
I still didn't get a chance to access the proposal to comment on it, waiting for my approval to join ASF in Melange. thanks, Thilina On Thu, Mar 26, 2015 at 11:50 AM, Danushka Menikkumbura < [email protected]> wrote: > Done. > > Thanks, > Danushka > > On Thu, Mar 26, 2015 at 11:46 AM, Suresh Marru <[email protected]> wrote: > >> Hi Danushka, >> >> Can you make the proposal public and send that url? This is only >> accessible by mentors. >> >> Suresh >> >> On Mar 26, 2015, at 11:35 AM, Danushka Menikkumbura < >> [email protected]> wrote: >> >> Hi, >> >> Please find a rough version of my proposal for topic 2 at [1]. Please >> review and give your feedback. >> >> [1] - >> https://www.google-melange.com/gsoc/proposal/review/student/google/gsoc2015/danushka/5704420943724544 >> >> Thanks, >> Danushka >> >> On Mon, Mar 23, 2015 at 12:29 PM, Suresh Marru <[email protected]> wrote: >> >>> Hi All, >>> >>> Since Danushka is a committer and understand fair amount, please suggest >>> if you have anything to challenge him. >>> >>> Hi Eran, Thilina, & rest of Airavata Community, >>> >>> While we wait for some challenging topics to explore for GSoC, I am >>> debating between two, can you weigh in? >>> >>> * Explore using Airavata to compose Oozie Workflows and catalog the >>> execution states and facilitate further analysis. >>> * Explore using Storm/Spark for Airavata Task Management. >>> >>> Integrating Oozie has a value since Airavata workflow related tools >>> (composer, monitoring, messaging, cataloging results) all are essentially >>> DAG driven. Secondly, a non-science gateway usage will help ensure the >>> architecture is not proliferating science terminology deep into the system. >>> This project might either validate Airavata as a general purpose >>> distributed application management or expose shortcomings in architecture. >>> This is also timely, see few discussions from Shameera for context [1], [2]. >>> >>> Previously in Architecture discussions exploring Storm for Airavata came >>> up [3] [4]. Lahiru has prototyped it [5] and also documented the learnings >>> [6][7]. I wonder if Danushka can build on these and do a academic study of >>> the project (GFac’s distributed task management with external state) and >>> compare and contrast multiple options. >>> >>> I am split between both of them as they both can be valuable additions >>> to Airavata 1.0. So I would suggest we go with a GSoC idea based on who has >>> some time to mentor. I am willing to co-mentor but looking for someone with >>> deep knowledge on these topics to be the primary mentor. >>> >>> Suresh >>> >>> [1] - http://markmail.org/thread/tkpbj3sr4jhg6o6z, >>> [2] - http://markmail.org/thread/vhworkj3736xisn7 >>> [3] - http://airavata.markmail.org/thread/u6my2vxrpgqdviae >>> [4] - http://markmail.org/thread/2ycudn3hvtgz3k3c >>> [5] - https://github.com/apache/airavata/tree/gfac-storm >>> [6] - >>> https://docs.google.com/document/d/1SU1U013Ec_NiW_Y82HRJYVliyoOwNzRH69K4vRli8vU/edit?usp=sharing >>> >>> [7] - >>> https://docs.google.com/document/d/1MigsXsC1VGM-Y7qSblOh9xMO2xdAlYBHUiWppPmK8_A/edit?usp=sharing >>> >>> >>> On Mar 23, 2015, at 11:51 AM, Danushka Menikkumbura < >>> [email protected]> wrote: >>> >>> Hi devs, >>> >>> As I have heard a lot of interesting things are happening (and are about >>> to happen) in Airavata space especially targeting the 1.0 release. Yes its >>> last minute ;-) but if you have got some interesting and challenging topic >>> I am happy to contribute as a GSoC project. But please note that it may go >>> beyond a GSoC project and have some value addition to Airavata in the long >>> run. >>> >>> Thanks, >>> Danushka >>> >>> >>> >> >> > -- https://www.cs.indiana.edu/~tgunarat/ http://www.linkedin.com/in/thilina http://thilina.gunarathne.org
