Meguro: A Map/Reduce Framework
frameworks
server
tracemonkey
Do you love server-side JavaScript? Do you need map/reduce? Do you use Tokyo Cabinet? If you answered “Yes” to any of these questions then Meguro might be exactly what you’re looking for!
The JavaScript part of Meguro is inspired by CouchDB, which is a very easy API to work with — two functions, zero hassle. Tokyo Cabinet is an interesting choice for storage because it suits certain classes of problems very well. I’ve used Tokyo Cabinet for a project that accesses millions of small records (like logs), and I was very pleased with its efficiency in terms of storage space and speed.
The author, Curtis Spencer, notes that Meguro is good for projects where you need performance but perhaps not the empire-building scalability of Hadoop.
Both the map and reduce functions take a key and value, and API methods are called to emit or save values:
function map(key, value) {
if (a condition is met) {
Meguro.emit('some important value', '1');
}
}
function reduce(key, values) {
Meguro.save(key, values.length);
}
If you’re new to map/reduce but think it might be useful for a project, Meguro might be a good start because it’s easy to set up, and code could be carried over to other platforms like CouchDB with only minor changes.
Meguro is also on GitHub: jubos/meguro.
