NoSQL Job Trends – August 2011
Continuing the August job trends month we have another installment. Today, we are looking at NoSQL job trends. In this edition, I have not changed the list of NoSQL offerings, so the same 9 tools are included. So, the list includes Cassandra, Redis, Voldemort, SimpleDB, CouchDB, MongoDB, HBase, Hypertable and Riak. As was stated previously, Hadoop continues to be the clear leader in demand and tends to flattens the trends of other solutions. Hadoop is still the clear leader in demand, but other tools are closing that gap. So, Hadoop could return to the analysis in the next update.
First, we look at the trends from Indeed:
MongoDB has overtaken Cassandra in terms of demand, and has a significant lead already. MongoDB has grown rapidly in the past 18 months and has received a significant amount of publicity. Cassandra actually looks to be leveling off, with a slight decline in the past few months. HBase is still growing, with a better upward trend than it had during the previous 12 months. Redis is showing nice growth in the past 6 months, with a significantly better trend than previous months. CouchDB is still growing, but has slowed a bit lately. SimpleDB and Voldemort both have shown basically no growth in the past year, possibly being overshadowed by the other tools. It may be difficult to see in the graph, but Riak has a very solid growth trend and passing SimpleDB and Voldemort in the process. Hypertable demand is completely flat and barely registers which means it may not be in the next update. One minor note for these trends, Cassandra has some noise in the job postings and I am looking for a way to filter properly.
Now, let’s look at the short term trends of SimpleHired:
SimplyHired mostly agrees with Indeed, showing Cassandra and MongoDB leading and HBase growing at a much slower pace. Redis shows similar growth to the Indeed trend. SimplyHired does show a flat trend the past few months for CouchDB, which does not agree with Indeed. The trend for SimpleDB is also different, showing nice growth during the Spring. Voldemort is showing a flat trend or a slight gain, and Riak shows some solid growth just like the Indeed trends. SimplyHired agrees with Indeed on the Hypertable demand, or lack of it. Overall, the graph shows solid growth for most of the tools.
Now, let’s look at the relative growth from Indeed:
I love the relative growth chart, but sometimes it can be a little misleading if one of the data points is exploding. Here we can see that demand for MongoDB is going through the roof. About 55,000% growth is crazy, and just makes the growth of other tools seem minor. However, Redis is also showing a great growth trend just above 30,000%. HBase continues strong growth as well, with Cassandra a little behind. Interestingly, Cassandra growth is strong but it seems to be leveling off a bit. The growth for the rest of the tools is difficult to see given the scale of the graph. CouchDB continues with solid upward trend and Riak is showing similar growth. SimpleDB and Voldemort both seem to have hit a plateau, with SimpleDB near 2500%. Lastly, Hypertable barely registers any growth in this view.
MongoDB is clearly gaining a lot of popularity, even overshadowing Cassandra. It will be interesting to see the longer term trend for Cassandra to determine whether its growth was only due to Facebook or more widespread adoption. Redis and Riak are definitely tools to watch as their growth is really starting to take hold. Given its Hadoop roots, HBase cannot be ignored and continues to grow on its own. CouchDB is another tool to watch, and it has received a solid amount of press coverage, but obviously that does not translate directly to job growth. It could just be a matter of time before CouchDB shows similar growth to the other tools. For the other tools in the list, the job trends look like a recommendation to wait for more adoption before spending a lot of time learning about them.
As always, if there are other NoSQL tools that should be included, please let me know in the comments.
(Note: Opinions expressed in this article and its replies are the opinions of their respective authors and not those of DZone, Inc.)