EMR vs EC2/Hadoop on AWS

Question

I know that EC2 is more flexible but more work over EMR. However in terms of costs, if using EC2 it probably requires EBS volumes attached to the EC2 instances, whereas AWS just streams in data from S3. So crunching the numbers on the AWS calculator, even though for EMR one must pay for EC2 also, EMR becomes cheaper than EC2 ?? Am i wrong here ? Of course EC2 with EBS is probably faster, but is it worth the cost ?

thanks, Matt

http://stackoverflow.com/a/15178916/610305 – Amar Oct 12 '13 at 20:55 — Amar, Oct 12 '13 at 20:55

score 5 · Answer 1 · edited Oct 22 '13 at 14:11

5

EMR does a lot of things for you that you won't find on standard Hadoop on EC2. Some particularly important ones include

Copying Hadoop logs from your machines to S3. This is very useful for debugging errors after the cluster has been shut down.
Running job flows of multiple MapReduce, Pig, or Hive jobs
Setting sensible configuration defaults based on hardware size you choose
Access to spot instances for cheaper compute
Ability to resize clusters dynamically

You'll also find that the EMR S3 filesystem is faster and more reliable than the standard one packaged with Apache Hadoop. It supports Multipart upload, and streams writes directly to S3 rather than buffering to disk first. For a bit more on this, see Tip #5

Additionally, if you do decide to use EC2 directly, I'd recommend using instance-storage instead of EBS for your nodes. There's really no reason to pay the extra cost of EBS for Hadoop; you'll notice that EMR clusters all run on instance-storage nodes as well.

edited Oct 22 '13 at 14:11

Andre Lombaard

6,985
13
55
96

answered Oct 22 '13 at 13:46

ddaniels888

51
2

thanks - regarding using instance storage: my jobs will be sparsely run, so i want to shut down instances. Streaming the data and results back and forth sounds slow compared to storing everything to EBS? Also, AWS offers either small instance-storage (~8Gig) or super big (48TB), not that great a choice, i have a couple of TB of data .. Another question: do i need to account for Name Node/2ndary NN, Job tracker machines as well or EMR handles this for me ? – matthieu lieber Oct 23 '13 at 17:57
EMR will handle the NN and Job Tracker setup for you. You'll probably find it easier to stream your data in/out from EC2 than try to get it onto the EBS nodes some other way. Also, once data is in Hadoop it is replicated 3X by default, so you should be able to shut nodes down 1 or 2 at a time w/o issue, regardless of the backing storage. – ddaniels888 Nov 11 '13 at 19:10

score 2 · Answer 2 · answered Oct 08 '13 at 12:05

You are correct that EMR uses instance-store backed EC2 instances, rather than EBS. However, there's nothing stopping you from creating an instance-store based instance, packing an AMI and using it for your Hadoop cluster. Using EBS also might not represent a lot of additional costs, depending on your workload and frequency. Also, there's an added cost to the EC2 instance when using it through EMR.

I've been using EMR for two years now and I would highly recommend the service as you don't need to invest time in managing and updating your distribution. If your workload is compatible with EMR (getting data from DynamoDB or S3), I would go for EMR as opposed to EC2/Hadoop.

EMR vs EC2/Hadoop on AWS

2 Answers2

Linked