11:01 a.m.
Hi Ashish.
Sorry for the delay in my response. I'll do my best to answer your
questions.
1. We want to have single API Manager for our consumers in
"on-premise"
and "cloud" (AWS/Azure..etc). So, can we deploy couple of instance of "
Apiman Gateway" in on-premise and couple of on cloud ? If yes, what
would be possible deployment architecture ? (best place to put RDBMS
on-prem/cloud, best place to put keyclock server/ ES server..etc ?)
There are a very large number of choices here! It depends a lot on
exactly what you're trying to achieve. Two different configurations
jump to mind.
Config 1 - A single "logical" API Gateway
In this configuration, every API Gateway node (regardless of whether it
is on-prem or in the cloud) shares a common data store. This allows
every node to service the *same* set of APIs. In this case the single
gateway storage (database or elasticsearch) will need to be accessible
by all gateway nodes.
In the UI, there is only *one* gateway, so there are no extra steps that
API developers need take when publishing APIs. However, every Gateway
node must be able to access every back-end endpoint at the same URL.
Config 2 - Two "logical" API Gateways
In this config, the on-prem gateway node(s) and the in-cloud gateway
node(s) each have their own gateway storage (database and/or
elasticsearch). When set up this way, the API Manager must be
configured with *both* gateways, and each gateway should have a sensible
name, such as "On Premise Gateway" and "Cloud Gateway".
In the UI, the API Developers will need to choose which of the two API
Gateways their API should be published to (they must choose *at least
one* gateway). This allows different gateways to service different
APIs, offering more flexibility.
2. production guide having wildfly version, however we would prefer
tomcat / vert.x, any specific guide changes for tomcat / vert.x ?
We only have documentation for Wildfly at the moment, but if you wish to
use Tomcat it should be relatively simple to extrapolate the
documentation to that platform. The vert.x implementation is
experimental at this point, and should not be used in production.
3. Can we have companies using apiman on production ? any
benchmarking
stats and case study material ?
I'm sorry to say that we don't currently have reference users or
production benchmarks yet. These materials will surely be coming in the
future as the project matures.
Thanks!
-Eric