Hello

I do not recommend viewing NiFi as something to use 'instead of' Storm
or Spark.  NiFi's design and the goals behind it are focused on
managing the flow of information from systems which create data, to
systems that process data, to systems that store data.  Invariably in
the process of connecting the many systems that exist throughout the
enterprise there are mismatches of protocol, format, schema, priority,
interest, authority, etc.. In solving these mismatches NiFi is often
used to do 'data processing'.  Examples of data processing that NiFi
aims to be very compelling for include things like transforming from
one format to another, filtering, sanitization, enrichment,
aggregation, splitting, etc..

In this light there are certainly processing functions done in systems
like Spark or Storm that NiFi would be perfectly fine and perhaps even
better suited for.  But, it is important to keep in mind the focus and
intent of the systems.  In my view, NiFi is oriented around dataflow -
connecting systems within and between datacenters and so on.  Storm
and Spark are focused on processing/analysis of that data.  So these
systems make different design trade-offs and provide different user
experiences centered around their intent.  As a result I think the
more compelling story is about how to leverage the strengths of these
sorts of systems together.

Always tough to answer this question in generic terms.  Happy to talk
through specific use cases.

Thanks
Joe

On Wed, Oct 28, 2015 at 8:02 PM, Adaryl "Bob" Wakefield, MBA
<[email protected]> wrote:
> I've been studying NiFi and there is something I'm not quite understanding.
> Can NiFi be used in place of Storm or Spark Streaming to process streaming
> data?
>
> B.

Reply via email to