Thanks Sterling, I can whip up a talk about the general premise of what
I implemented and what I believe the merits of the approach are.
I'm planning on expanding on the idea to support bi-directional data.
Will probably look into making a little app in node with Express as well
to support retrieval and display. Perhaps I'll also make an index over
time in addition to the data the app currently creates.
Depending on your timeline and my productivity at home, I can give a
talk about the latter in addition to the former.
- Josh
On 05/03/2012 08:42 PM, Sterling Foster wrote:
Josh,
Very nice. Would you be willing to give a talk at the next cloud meeting?
Sterling
Sent from my iPad
On May 2, 2012, at 9:46 PM, Josh Elser<[email protected]> wrote:
A while back, I wrote some code as a proof of concept to use node.js (an
event-driven, server-side Javascript environment) as a way to ingest data into
Accumulo. I think it's a cool example of alternative ways to get data into
Accumulo, so I cleaned up what I wrote and posted it to Github.
https://github.com/joshelser/node-accumulo
Using RabbitMQ (open source message broker which implements the AMQP standard)
and a node.js plugin for AMQP, I was able to send data from a node.js process,
through RabbitMQ, into a background Java process which inserts the data into
Accumulo.
The example is a contrived website hit-tracking application (think Google
Analytics). When a user visits a website, an HTTP request -- with the client
IP, server IP, and current timestamp in the query string -- is made to an HTTP
server running inside of the node.js process. This information then makes its
way into Accumulo as a key-value pair.
I hope to expand on this some more, but I wanted to share the existing concept
with everyone. Instructions and additional information can be found on the
project page. Feedback is welcome!
- Josh
