Companies have to be creative when it comes to the Hadoop talent they need for Big Data projects. With so few people with those skills, the ones out there can pretty much name their own salaries.
One potential solution calls for using the staff you have already. (What a cool idea!) The Hadoop support structure Hive, meanwhile, is similar to standard SQL. As Stefan Groschupf, CEO of Datameer, put it at ZDNet:
A lot more people know SQL than can write Hadoop’s native MapReduce code, which makes use of Hive an attractive/cheaper alternative to hiring new talent, or making developers learn Java and MapReduce programming patterns.
Hive has its limitations, though, he goes on to say, including that it allows you to query only structured data and that its response time is pretty slow.
You can use Hive within applications written in C++, Java, PHP, Python or Ruby, much like you can use these languages with embedded SQL, IBM says in a post that points out basic differences. It, too, notes that the latency makes it inappropriate for applications that need a rapid response. And since it’s read-only, it won’t work for transaction processing that requires a lot of write responses.
Still, with its limitations in mind, Hive might be a useful and affordable Big Data talent solution. I found this demo for more information with IBM Big Data development lead Anshul Dawra.