Hi,
I'm working with a large Asset Manager and they are moving both their API Proxy code (Intelligent layer upon RFA) and their ADS hardware to a virtualized environment. We are wondering, how the persistent connections will behave during a VM Live Migration?
This will be essential in case of a Hypervisor failure during a busy period (EOD, etc.).
Thanks,
David B.
Based on the RFA Java readme file, it supports Virtualization on VMWare ESX 5.X for OMM interfaces with TCP connections on Oracle and RedHat Linux 6 (32 & 64 bit).
What's technology behind the virtualized environment? Is it similar to running a Java application to a virtual machine?
From my opinion, RFA Java underlying layer implements a standard Java socket programming to connect to an infrastructure (e.g. ADS). So, as long as a pure Java application in a virtualized machine can create a socket connection to another machine, RFA Java should be able to do it as well.
By the way, my suggestion is that you can test it with the RFA Java example such as a simple StarterConsumer program to test subscribe data to the server first.
Hi Nipat,
Yes these will be virtualized machines (VMware VMs) and later containers (think Docker/Kubernetes in AWS). The question is not about startup, but rather how does RFA handle transitions (VMotion from one server to another or AWS migration to another DC etc.).
Does that add some color?
-David B.
I'm not familiar with the transitions (VMotion) that you mentioned earlier, so I couldn't confirm it for now.
In order to do some further research, it seems like there are additional things or actions required as a constraint before moving a basic/generic Java application to AWS via Docker or Kubernates, am I correct?
If so, can I have the reference(s), so I can find whether it is also mentioned in any RFA's documents or not?
