HPE Ezmeral Data Fabric Reviews

It has several valuable features, among them being the file system, High-Availability services, and NFS.

This product enabled us opening up endless possibilities in data analytics, IOE/IOT, and predictive analysis.

It'd like to see file system auditing, data encryption, and certification of other vendors' tools.

I've used it for three years.

We've had no problems with deployment.

It hasn't be unstable for us.

It's scaled for our needs.

I rate customer service a 7 out of 10.

Technical support is a 7 out of 10.

We were using another solution and switched because of MapR's highly available services, and the robust file system.

The initial setup was straightforward.

We implemented it in-house and it was pretty easy.

I can't tell what the ROI is at this moment.

Make sure the third-party applications you are trying to implement with it are certified with this product. If not, please start discussing with the MapR team that they can start a partnership with the third-party application vendor.

The fact that the heavy computation is required on Big Data can be distributed across many nodes in a cluster, makes this solution a winner. Of course the same concept is available across any solution based on the Hadoop architecture, but MapR provides bundled services and UI driven configurations which would otherwise take a long time to understand and implement.

We implemented this for our client where data from sensors was to be analyzed and sense to be made of this data. Sensor data is huge, and earlier there was no meaningful early warning raised against the discrepancies observed on Sensor’s Data. With the help of our Big Data solution, we have been consistently raising alerts as required.

• Installations and setups are still a bit cryptic and can be improved.
• Skilled resource base in Big Data Tools is generally low and hence project costing is that much higher.

Please make sure following answers are clearly known:

• Define the Objectives, and what you are not currently able to achieve without Big Data Tools.
• Why is the data big - because of velocity of data pouring in. th number of years of data, or a sudden increase in business operations
• Define a sample use case / example of the expectations from Big Data analytics
• Carry out a POC & review the objectives.

The most valuable features for us are--

• MapR-DB
• Multi-tenancy

It's allowed us to use Twitter as a data source to monitor the effectiveness of our marketing campaigns.

The UI for administration still has a lot of manual work to set up the cluster and get it running.

We've used it for six months.

No issues encountered with deployment.

No issues encountered with stability.

No issues encountered with scalability.

Customer service is good.

Technical support is good.

We switched because we were able to get better value for the money for support.

Setup wasn't complex.

We did it in-house. You should start small, and iterate often.

We don't have a quantifiable ROI yet. We still need to complete our proof of concept.

I highly recommend MapR.

MapR is one of three leading Hadoop distributions. The key differentiators I found the most valuable are:

• Lack of namenode (no SPOF)
• Performance – the base components are implemented in C
• NFS functionality – copying to MapR FS is as easy as copying files to a directory

I haven't yet used it in production environments, but I've used it extensively in two proofs-of-concept.

It would be nice to have new developments in the Apache space (Spark, Storm, etc.).

I've had no issues with deployment so far.

It's been stable so far.

There have been no scalability issues so far.

I haven't yet had to use customer service.

I haven't yet had to use technical service.

I tested both Cloudera and Hortonworks, and I rate it similarly to Hortonworks and slightly higher than Cloudera.

Initial setup is rather straightforward thanks to detailed documentation covering all the bases. More advanced installations required manual deployment to an extent, but this too is well-documented.

I implemented it myself.

I haven't put it into production yet.

I don't have any advice because the installation documentation is superb.