- One reason for separating S4 app and adapter is so that the adapter
only focuses on converting external streams into streams of S4 events,
while the other S4 app(s) focus on business logic
- Another reason is that the adapter benefits from properties of S4
apps, in particular scalability, being an S4 application itself.
- Last, using separate clusters for separate apps allows independent
deployment and administration.
However, you may as well implement some business logic in the adapter
app itself. It really depends on the complexity of the app. Implications
on static load balancing are really dependent upon the stream and the
application.
Hope this helps,
Matthieu
On 10/10/12 8:26 AM, Frank Zheng wrote:
Hi All,
Why do we need to put adapter and PE in separated clusters now?
Is it possible to put adapter and PE nodes in one cluster?
And how to do load balancing when adapter and PE nodes are in the same
cluster?
Thanks
Yu
