The Ukora™ News Search Service is made possible by a variety of interesting, open source technologies. This, in combination with affordable hosting provided by Digital Ocean allows us to provide this service at a relatively low cost. It is our intend to inspire others to test what is possible in this fast evolving landscape.
Ukora WeRock is our proprietary development framework. It allows for re-use of common components and abstract away most of the aspects of web development that make it feel like work, and not creative play.
Our web-crawler is fully independent and built using MongoDB as it's main storage engine. We have a distributed/scalable database architecture containing 500,000,000+ objects which allows us to find news feeds at a fast rate and keep results relevant.
Our crawlers insert searchable records into ElasticSearch, backed by time tested Lucene. This allows for a seamless and scalable full text search implementation
For all things relational, MySql still cannot be beat. The WeRock social CMS relies on MySql for almost all persistent storage.
Memcache and OpCache are leveraged to store objects whose scope crosses multiple requests in memory for lightning fast access.
A trusted web server, when configured correctly apache2 is the right mix of a stable and rich feature set and performance.
The WeRock platform leverages some new PHP features like traits to allow for cleaner design pattern implementations.
We are happily running our crawlers, databases, indexes and web-application with DigitalOcean. Keep up the good work!