Why? Because, the more the data, the longer it takes to index it with sphinx.
We want to index relatively in real-time (or async with background jobs), and continuously.
Do we need to search through backends also?

Requirements

1. searching in more objects: please develop some future goals tmaas-1
2. HA support
3. TLS support
4. Multiple Database support (Oracle, MySQL, Postgresql)
5. Ability to deploy inside and outside of OCP cluster for small and big deployments
6. Real time indexing
7. Fast distributed re-indexing
8. Healthy OpenSource project. Define "healthy" (active contributors, security fixes ...)

Possible replacements

• Postgresql, i.e. using it also as the single supported database – ref: https://www.postgresql.org/docs/10/textsearch.html
• ElasticSearch (industry renowned and very active project)
• SphinxSearch (listing here in case it fulfils all requirements)
• ManticoreSearch (SphinxSearch fork with active development)