I have created a mini-cloud of Hadoop in my lab on few machines, when I executed a mapreduce job on the mini-cloud and on single VM(independent, not in cloud) , i found that the time taken on single VM is much more than the mapreduce job executed on mini-cloud.
If anyone can suggest me the reasons behind it. Thanks in advance.
On the basis of whatever is mentioned in question, It is obvious that execution time on a single machine will be more than that on more than one machine i.e. few machines.
Dear,
Map functions, when you distribute them on multi VM, you run them on parallel.
and when you run them on one VM you lose the benefits of parallel.
Regards!
yes , I thought the same!
But was searching for few papers based on this concept so that can use in the literature survey. So, if anyone has them please post here.
Thanks in advance.
https://www.researchgate.net/publication/303250528_Observations_on_Factors_Affecting_Performance_of_MapReduce_based_Apriori_on_Hadoop_Cluster
This paper might match your searching. Use case is Apriori algorithm but some cases can be applicable to others as well.
Conference Paper Observations on Factors Affecting Performance of MapReduce b...
Thanks Sudhakar and Ahmed Hassan Abase for sharing the docs.
They would be really helpful.
- Badges
- Science method