Enough with worthless 2014 predictions. We can't control what the rest of the world will do — but we are the captains of our own fate. In case you hadn't heard, 2014 is the year we stop talking and start doing in earnest. IMHO, here's the stuff you should be rolling out.
1. Document databases
IT develops lots of systems that basically consist of writing data structures to structured storage, while requiring high concurrency. Document databases have been around for decades, and IT knew one as Lotus Domino's Notes Storage Facility, but the NoSQL and big data revolution has sparked new interest. MongoDB andCouchbase are the leaders.
2. Key-value stores
Sometimes you have what amounts to a really giant table that could fit in memory, if only you could get enough memory. Well, if you have a grid, you can put the table in memory distributed across multiple boxes to make writes faster. If it's a smaller table that's read-mostly, you can replicate it across all nodes so that reads are a memory thing. Either way, key-value stores deserve a loop on your toolbelt. Nearly all key-value stores allow you to create custom cache loader or cache stores to read/write to an RDBMS or another source. Many allow you to "write behind" or queue writes to the database. The leaders in the space are Couchbase, Memcached, Infinispan, and GemFire.
3. Graph databases
From recommendation engines to social networks to geographic analysis to even bioanalytics, graph databases give you more bang for the buck. Friend-of-a-friend queries are not too efficient in the traditional RDBMS, even with all of the latest features, because the structure is still wrong. While graph databases have been around for decades, the recent data explosion and interest in personalization has made them more popular. Check out Neo4j and Apache Giraph.
4. Google Drive/Apps
5. On-premises search
I still run into people who actually compose massive and/or/likequeries in SQL. Get with it — not only does this result in horrible performance, but also terribly unclear code and usually a user-hostile interface. Elasticsearch has been catching fire, and I'm seeing it everywhere. Think of it as Google for your own data, whether it resides in the database, documents, or various file systems.Apache Solr is worth looking into, as well.
Sign up for Computerworld eNewsletters.