Flexibility in terms of options of software patches. Review collected by and hosted on G2.com.

Overhead delay in starting clusters which can cause problems. Review collected by and hosted on G2.com.

The ease of use and creating new clusters in seconds Review collected by and hosted on G2.com.

The user interface could use a facelift to clean clutter Review collected by and hosted on G2.com.

EMR used to develop big data solutions. I developed many big data solution using AWS EMR Review collected by and hosted on G2.com.

Need more documentation on the Amazon site to know about EMR Review collected by and hosted on G2.com.

the configuration and node managements are flexible and the performance is good

it takes care of all installation and setup of

hive,

spark,

all other big data tools

the resizing feature is good, we can add or remove core and task nodes on a button click

EMR internally uses EC2 , which allows us to use other features that EC2 supports such as node vice monitoring , clowning etc

the monitoring dashboards has some use-full KPI's Review collected by and hosted on G2.com.

the app logs are difficult to view, could implement centralised logging mechanism instead of YARN

we may need a way to terminate unused clusters

could introduce terminate by duration feature, so that the EMR gets terminated by set duration (should be flexible enough so that we should be able to cancel or edit it at any point if time) Review collected by and hosted on G2.com.

The distributed computation of the calculations. Review collected by and hosted on G2.com.

The need to scale up the memory and choosing the EC2 instances Review collected by and hosted on G2.com.

On demand transient clusters for big data processing. I like its availability in different cost tires makes it extremely elastic for different scale users. Can be pre-installed with any Big data tools like Hive, Spark, Pig etc. Detailed cluster monitoring helps to track several metrics, in-turn helps to reduce cost. Review collected by and hosted on G2.com.

Charges of ERM are like EC2 charges and then data processing cost added to that. Together cost becomes bit on higher side.

For persistent EMR clusters, some time master nodes goes without notice, making entire cluster unusable. Review collected by and hosted on G2.com.

.It allows to work on large amount of data an on different framework. EMR offers the expandable low-configuration service as an easier alternative to running in-house cluster computing Review collected by and hosted on G2.com.

Its complex to use in initially and i think there should be proper Userinterface for real time. Review collected by and hosted on G2.com.

EMR is simple to use, configure, and maintain. The number of components available in EMR is very good and serves almost all the usecases and rarely you will have to install a software on your own. The ability to mix instance lifecycles (spot, on-demand, reserved) is very helpful to optimize the costs Review collected by and hosted on G2.com.

There isn't an easy way for setting up authorization and authentication in EMR. Cloudera does a better job in maintaining that. Review collected by and hosted on G2.com.

We were able to easily crunch over 400million events every few hours. We can use HIVE which makes the learning curve much easier. Review collected by and hosted on G2.com.

IF you forget to turn off your instance.. you can expect a nice heavy bill. The sizing of the cluster is not easy to use, we tend to just guess. Review collected by and hosted on G2.com.

the ability to spin up cluster size based upon the needs of each job Review collected by and hosted on G2.com.

error logging is sometimes hard to track down root cause Review collected by and hosted on G2.com.