基于 Apache Solr 的 Spring Boot 应用搜索功能

Question

We are in the planning phase for building a search feature with autocomplete for our Spring boot application which will be very similar to search feature on Pinterest. We are considering, developing a search engine based on Apache Solr. That said, our team(2 people) has no prior experience with search engines and would be implementing the solution as they learn.

Although, I have not provided our detailed requirement for the search engine. I would like to get input by other members before we start investing more resources on this. Kindly give your input on the following points

1. What is the learning curve for Apache Solr for a 7-8 years experienced Java developer?
2. What should be the effort estimation(in hours) for building such a search engine. And delivery estimation given two people are working on it?
3. What additional infrastructure cost to consider for Apache Solr?

Thanks for your inputs.

Update: Somebody has voted for closing the question because answer to the question is opinion based. Well, this question does require other members to share their opinion so we can make a better decision about the specific scenario described in the question. That should not be a problem.

Answer 1

I tried to provide the answer to your queries. One cannot give you the exact answer until all the questions are resolved. But if you have answers to question I have raised, it would help you to do the analysis in detailed way.

1. What is the learning curve for Apache Solr for a 7-8 years experienced Java developer?

   One does not need any java background as such for the leanrning solr. The solr is developed on lucene. The lucene is a search library developed in java. Solr is configuration based. But still it all depends on how you want to use it. There are many option by which you can index the data by posting the data in the form of xml,json etc. You can indexed the structured data like rdbms by using solr DIH feature. You can also index the data using solrJ api which are for Java developer.

2. What should be the effort estimation(in hours) for building such a search engine. And delivery estimation given two people are working on it?

   It all depends on the scope of the search engine. How you want to index the data? How you want to query the data? What all features you want to implement for example autosuggest, faceting, highlighting, filtering etc. Is there any requirement of re-indexing the data periodically.

3. What additional infrastructure cost to consider for Apache Solr? Solr come with solr cloud feature. Where solr can be used in clustered or distributed mode. Here sharding and replication features can be used. If you want to use solr cloud then installing zookeeper and solr nodes need to be considered. If you want to run solr in standalone mode then not much infra is needed. but again it all depends on the data size which is been indexed. How frequently it is updated? How many users are querying it etc.