Search seems like a fairly mundane feature on the internet that no one really cares about. Let’s face the facts, search isn’t that exciting compared to cutting-edge topics like machine learning and DevOps. We suggest you take a moment to rethink search, as we take search for granted these days.
Imagine shopping on Amazon. Just type the title of the book you want in the search bar and hit enter. You’ll see a page of results in just two to three seconds. Amazon sells more than 75 million products, according to one estimate. One search could find what you were looking to in seconds. This is crazy. How can they search such a large product database so quickly? It’s not just Amazon. Google, YouTube, Spotify and any other news site will all provide a user experience that is based on you searching for a string of text and receiving a detailed and accurate list of results.
Imagine if these sites took consistently 10 to 15 seconds to return search results. I’m sure you wouldn’t stay on these sites for long and would move on to a site that was faster. Imagine if you had to go through pages after pages of irrelevant results in order to find what you were looking for. If you can’t find what you are looking for, no store or streaming service will survive. Search is the most important functionality of the internet and a form magic. Instant and relevant results are possible from millions to billions of data rows.
This magic can be done for your site without being as difficult as you might think. AWS offers two services to help you build cost-effective, high-throughput, low latency search solutions. ElasticSearch and CloudSearch. They sound very similar, so they are. There are subtle differences that could affect which service you choose.
AWS Certified Developer-Associate (DVA-C01). Related training from SPOTO
CloudSearch Training
CloudSearch uses Apache Solr to power your search queries. It’s an app that is based on Solr but modified via the AWS console and API to make it easier to manage.
They describe Solr as “highly reliable, scalable, and fault-tolerant. It provides distributed indexing and replication, load-balanced querying and automated failover and recovery. It also offers centralized configuration. Solr powers many of the largest internet sites in the world’s search and navigation features.
However, such power is not easy to achieve. Solr is still required to power your site’s search. This can be done on a dedicated server or in Docker. You can also configure additional nodes if fault tolerance is desired. Plan for the maintenance and upgrades of Solr that will be required.
Most developers don’t have the time or desire to manage such complex applications. They would rather spend their time on something else. CloudSearch is the solution.
CloudSearch is a managed service like EC2 and S3, where AWS runs the application (Solr in this instance). The AWS console and APIs abstract away the application. All aspects of the software’s administration, including upgrades and installation, are done behind the scenes. This allows you to get on with your work faster and with less administrative overhead.
CloudSearch is set up by creating a search domain. This “encapsulates the data you want to search, search instances that process your requests, and a configuration that controls the way your data is indexed or searched.” This basically includes all your data and the options that you have set up around searching for it.
For scalabil
