Currently, ActiveFedora is very tightly coupled to the Solr API which blocks us from using CloudSearch (AWS' Solr like search engine) or ElasticSearch etc. As a result, we need to maintain a SolrCloud quasi-manually in AWS which we do via two separate beanstalk environments (one for zookeeper instances, one for solr instances) - this is a hefty maintenance lift on the infrastructure as solr can be very touchy about instances coming and going, a common occurrence in AWS.
Currently, ActiveFedora is very tightly coupled to the Solr API which blocks us from using CloudSearch (AWS' Solr like search engine) or ElasticSearch etc. As a result, we need to maintain a SolrCloud quasi-manually in AWS which we do via two separate beanstalk environments (one for zookeeper instances, one for solr instances) - this is a hefty maintenance lift on the infrastructure as solr can be very touchy about instances coming and going, a common occurrence in AWS.