The Fiorano Peer-to-Peer distributed model provides unmatched flexibility in deciding installation topology. Users can decide on an installation topology based on various parameters:
- Availability of hardware
- Hardware configuration – RAM (Random Access Memory) and, CPU processing
- Expected system performance
- Number of licenses available
The exact topology architecture for each solution varies and is determined based on specific customer requirements.
Typical Development Topology
Fiorano Platform (Enterprise Edition) can be installed on a single machine. The default Peer Server profile is configured to work with the Enterprise Server available on the local machine. As such, no further configuration (other than JVM settings, like setting the heap size) is needed to launch the Servers. The Fiorano eStudio also comes pre-configured with the local machine enterprise server connectivity information.
Single machine installation is the simplest and fastest way to get acquainted with the software. If the single machine has more than the recommended RAM and processing capabilities, then the servers can be deployed on the same machine. However, if a large number of processes are required to run, whereas the available hardware configuration does not support the memory requirements of the business process, then users have to consider distributed deployment of servers.
Typical Production/Testing Topology
Typical production or a testing environment topology of Fiorano Servers has all the servers configured in High Availability mode on different machines. There can be multiple peers in HA mode, which are distributed across other machines. The machines can be in the same network or different networks or even in cloud provided the latencies are manageable. With this approach, it is possible to distribute the load across multiple mid-range machines as compared to using a single high-end machine.
In the case of DMZ deployments, it is typically required to host the components such as Stubs and JMS Replier components in a separate Peer Server hosted in the DMZ. All other secure systems are communicating with microservices running on peer server present behind the firewall. Exceptions have to be defined in the firewall to allow communication with Enterprise server and other peer servers.
For disaster recovery, a separate set of peers can be deployed, and if needed, scripts can be deployed to move the processing from DR machines to actual peers and back. The diagram below shows a topology where disaster recovery servers are launched as standalone servers.