Hi,
i want to use Flume to collect logs in my network (~20000 logs/sec).
But now I'm not really sure, what are the differences between Flume and a
MOM like ActiveMQ.
I would say the flume agent on the client side and the flume collector on
the server side build a reliable channel. So i have a reliability from the
moment i give the message to the agent on the client.
And the transfer to a MOM isn't reliable.
Is this the biggest advantage of flume? Do you know any other project like
flume? I have to make a comparison between different transport solutions.
So i could use a simple Socket, JMS, Flume, ...
Regards,
Simon
|