Build your own search solution around your specific needs using one of the great open source alternatives solutions out there (Apache Solr/Elasticsearch) and stop worrying about what your search vendor will offer (or not) in the future.
If you’re somehow related to the Search Enterprise world you must have noticed by now that Google is discontinuing the Google Search Appliance (GSA) product, and that it will not be available for sale from 2017. Miles Kehoe and Martin White have written a very good analysis and some follow up steps as what to do next. I think that this proves once again the traction that Open Source Search alternatives are gaining and not as Laurent Fanichet of Sinequa says that the ends of GSA “seals the end of the era of commoditized search”, in my opinion this commoditization of search products is a consequence of the rapid growth of open source search options. It’s true that neither Apache Solr or Elasticsearch are drop-in replacements for GSA, but both of this solutions provide a solid foundation for any enterprise or home cooked solution.
In my opinion this situation, right now, is a very good reminder on what’s wrong with closed vendor/solutions like GSA/Endeca and FAST ESP (for instance), you develop your business around this key technologies (yes, search is a critical feature) and you wake up one day to find out that is no longer available or supported or actively developed. Sure, building search products is not easy but rather soon than later if will payoff, first you’ll be constructing a solution around a technology that is not very likely to disappear, Apache/Solr has been around for ages, and is being developed under the Apache Software Foundation, which means that a lot of great people is putting great ideas and code behind it; and the community around it will ensure that it will never go away (I’m not trying to be an absolutist here, but … you get the point). Sure both Apache Solr and Elasticsearch have companies supporting the development of the products (Lucidworks and Elastic) respectively, but the licensing options allow you to put your head in the pillow without any worries about whats happening tomorrow, perhaps the company will go away (not likely) but the products as they’re today will remain available and you could develop your own “fork” if you need to, or hire developers to work on your particular needs.
Migrating to something new is always a scary process, but there are a lot of great consultancy firms out there willing to help you: Open Source Connections, Flax, just to say a few, even the very same company that is behind the Solr development Lucidworks or Elastic which supports Elasticsearch development.
Also, know that you’re not alone, other companies have done the switch already and with happy results, let’s take for instance the case of Career Builder and other companies that once they made the change are developing stronger products making search a more essential part of their business. If you want to see who is using Solr and for what, check out the videos of the great Lucene Solr Revolution conference (in any of the years), this conference is a gold mine of Solr use cases.
One particular product developed by Lucidworks is Fusion, which is something like “Solr on steroids”. I talked previously about Fusion and Solr, and the short version is that if you don’t want to build a full search team / search product to support your business, then you can use something like Fusion or DataSax, that provides enterprise features ready to use. The key difference is that this products are built around Open Source technology and you won’t be tied up to a particular vendor. A lot of options exist, I really advice you to use something that is baked up by Open Source and if you could build your own solution around this technologies, even better.
As for the future, I think we’ll see a couple of companies writing “connectors” or interfaces to ease the transition out of GSA into their own particular plattforms.
So wrapping things up, I’m not saying anything new, but people get out of one closed vendor to the next one and only remember that bad things happen in times like this, when a product is discontinued. Moving to Apache Solr/Elasticsearch in 2016 is no that hard, consultancy options exists, the community is out there willing to help for free, and there are plenty of success stories to present to your stakeholders to convince them that the migration is worth the effort.
And finally, let’s add the two cents on free marketing :)
I’m currently working as a search engineer/consultant so if you think that I can help you, please get in touch.