Past Meetup

What's New in Solr 5 Security & Solr Custom Collector: The Anti-Score

This Meetup is past

26 people went

Salesforce.com

2001 8th Ave · Seattle, WA

How to find us

Please take the elevators up to the 14th floor and follow the signs to the Meetup space

Location image of event venue

Details

Join us for an evening of networking, food & drinks, and the below presentation from Lucene/Solr committer & PMC member, Anshum Gupta.

What's New in Solr 5 Security, Presented by Anshum Gupta, Lucidworks

Apache Solr has evolved into a highly scalable system, capable of handling a lot of data and high number of queries, but only recently was a mechanism to secure access in Solr provided. Apache Solr 5.2 shipped with pluggable authentication and authorization modules. These modules enable users to write their own plugins for managing security in Solr.

This talk will cover an overview of both the authentication and authorization frameworks, and how they work together within Solr. It will also provide an overview of existing plugins and how to enable them to restrict user access to resources within Solr.

Speaker Bio: Anshum Gupta is a committer and PMC member on the Apache Lucene / Solr project with over 10 years of experience with search and related technologies. He currently works at Lucidworks and spends most of his time working on SolrCloud i.e. the distributed feature set of Apache Solr. Prior to joining Lucidworks, he was a member of the team that developed and launched AWS CloudSearch - The AWS search as a service offering. He was also a key contributor in the search teams at various start ups. Anshum is also the release manager for Solr 5.

Solr Custom Collector: The Anti-Score, Presented by Michael Kosten, Getty Images

Sometimes, you don't want to return just the top scoring documents as your search results. If you have an eCommerce site, you may want to ensure that multiple lines of business are represented. If you incorporate customer interaction in your score, you may want to ensure that newer documents or certain categories are still represented and that your results don't become stale. This requirement could be handled in middleware that post processes the search results, by requesting extra rows and rearranging them or by interleaving multiple queries. A better solution is to implement your own custom collector in Solr, so that search results can be arranged in any order. I'll demonstrate a solution that returns top scoring documents, but grouped within categories. For example, a search for books could interweave the best fiction and non-fiction in a single query result. I'll also demonstrate how to implement a custom priority queue to reduce memory requirements if there are many categories, and how the custom collector can be integrated into Solr without modifying the base distribution.

Speaker Bio: Michael Kosten Principal Engineer at Getty Images attached to Search. He's been working on the Search team at Getty Images for the past 8 years. Before that, he was a software development consultant for 17 years to the insurance and construction industries, both in a solo-practice and as a partner in a consulting firm. Getty Images is a premier site for stock photographs and footage as well for stills and videos from news, sports, fashion and entertainment events.