mapreduce
	read next range
	split range into list of M items
	for each item in list
		rexec on a randomly chosen node:
			the map function; direct output
			to the partition function which is
			given a list of endpoints
		
	we should get a list of R endpoints
	for each endpoint
		rexec on node
			the reduce worker
			which reads in all data from the endpoint
			sorts the data
			and calls the reduce function with the
			key and list of values.

We'd need a new type for a list of endpoints or endpointsrv.
This is given to the partition function.
/endpointsrv tcp!node0 id

rexec tcp!node1 "{(/fd); /filter $1 "{tq -1rm /n/tdb/tick.bt}} |
	/part {/endpointsrv tcp!node0 id} 

part reads each record hashes the key are writes
the record to the endpoint on endpointsrv identified
by the hash mod R.

/endpointsrv tcp!node0 id |getendpoint "{
	rexec tcp!node1 "{reduce "{tock}}  
}

getendpoint reads the list of endpoints
provided by endpointsrv and call
rexec on each with the reduce function