3 ms·
Hadoop is good at what it's designed for (data processing parallelization built by programmers) and bad at what it's not (ad hoc queries by people who are not p
by spidaman 15y ago
Hadoop is good at what it's designed for (data processing parallelization built by programmers) and bad at what it's not (ad hoc queries by people who are not programmers).
Move along here folks, nothing to see here.
- bunderbunder 15y agoI'd constrain the description of what it's good at even further. The kind of data processing parallelization it's designed to support is map/reduce. That's only one out of a wide array of possible ways to organize a distributed processing job, and not all tasks are well-suited to it.