04-24-2020 02:52 AM
Hi - if I may, two questions regarding APOLLO
1 - I have seen that for Windows Server 2012, an m4.large instance was recommended (https://hexgeo.co/2S1LYj6) but what about for Windows Server 2016 or Windows Server 2019?? This is for one APOLLO installation utilising dropbox, acting mainly as a catalog for upto 100 users accessing data and as a back end catalog to a LuciadRIA web portal with 1000's of users, doing preset searches. There may be some geoporcessing usage, from the 100 users.
2 - can the APOLLO geoprocessing server be seperated from the APOLLO server? So similar to the M.App Enterprise architecture.
04-24-2020 06:11 AM
Concerning question 2.:
You can't dissociate the APOLLO server from its geoprocessing capabilities.
What you can do is configuring two APOLLO instances in a cluster (2 lic needed). The 1st node of the cluster will deal with all the requests except the Geoprocessing and 2nd node will only process the geoprocessing requests. You just need to configure the load balancer properly to send only the WPS requests on the 2nd node.
04-24-2020 09:34 AM
Based on Apollo 2020 system spec, m4.large fit system specification.
Checked question 1 with Apollo development team, and they would suggest EC2 t2.xlarge as a minimum... EC2 t2,2xlarge might handle the geoprocessing workload better. Data storage requirements should also be considered, like whether all data will be local to the server. For external storage, I would recommend AWS FSx over S3 for APOLLO and I would configure FSx throughput to be >= 256 MBps.
hope it helps.