Hortonworks Data Platform Reviews

We use this solution to look at and manage big data. It's mostly historical data that we offload from our data warehouse, as well as from other databases in other platforms.

We have two different installations. The first one is based on IBM POWER CPUs, and the other one is based on Intel CPUs. Our data center is on-premise. There is some thought on moving to a private could, or a private IBM cloud, but we have not proceeded with that as of yet.

This solution is a cheaper way for us to offload the otherwise expensive data. We can move data from outdated database versions, such as Oracle 10. It is now out of support, but still hosts some of our historical data. This solution has helped us move our data to the current version.

Previously, we had our data on more expensive platforms. Now, using this solution, it is much cheaper to have all of the data available for searching, not in real-time, but whenever there is a pending request.

We have had problems with the backup and with services that require a disaster site. We are still struggling with some of these issues.

We are having trouble with Active Directory and Hive integration.

I would like to see more support for containers such as Docker and OpenShift.

We have had some issues with the code, but it's mostly from the developers. From our side, we don't see any issues with stability, although it may be that we have a lot of unused CPU capacity.

We have not acquired any additional hardware since our initial purchase. However, we expect more use cases to be added, at which point we may have performance or scalability problems.

The initial setup is not very difficult. The configuration is not easy, but somebody with some experience is able to set it up. We had users for which we had to set up quotas and queues. For us, the basic installation was completed within a matter of a week.

We had IBM set up both of our installations. 

This is a good product, but we still have some issues with backup, and the performance monitors that we install on every system. There may be solutions, but we're struggling to integrate them.

This is a product that I recommend. It's a solution that comes at a lower price, and it works well if you don't have expectations that it will behave like a much more expensive system.

I would rate this solution an eight out of ten.

Hortonworks actually provides a complete solution and just one user interface that can manage all the packages. It can monitor all the requirements, all the versions and additionally all the quays and all the hardware-dependent services. What I want is a useful user interface which is the reason why I currently prefer to use Hortonworks.

One of the most valuable features is that you can configure your data nodes in the big data. Whatever you want. Normally, for example, if you are testing websites and Hash clouds and other sites, in most of them you must manage more than three or four requirements. For example, you must install each feature, you must compile some additional things, and also you must manage more than three configuration files to enable all the nodes to work together.

In the Hortonworks solution, you just need the service, and you just want to install it once to get started on projects easily. You can just click run and it's already installed and you can create and communicate between your services.

I work a lot with banking, IT and communications customers. Hortonworks must improve or must upgrade their services for these sectors.

Each customer has different requirements. From the IT side, someone who has some experience of the cluster, computer clustering, computer networking, different fire defense, for example, it is so important that they have some additional graphics, some additional service reports added to the Hortonworks current user interface, which could provide easy images. Especially if they are using it without any experience first.

The Hortonworks solution is very stable. It works as a production system, without any error, without any downtime. If I have downtime, it is mostly caused by the hardware of the computers.

The solution is very scalable. In our procedure, there are two we are using in production without any downtime. The most important thing is the hardware from the computer cluster. I can work on more than 1,000 servers at the same time. On the communication solution, currently, 50 people use it at the same time.

The initial setup is so easy. You can just watch a video. It's handy. If you have some knowledge of computer networking and computer clusters, it is so easy. Deployment time depends on the project and the project size. Sometimes it takes more than three hours to complete.

The solution is comprehensible but it also depends on the customers and the customer's stability requirements. I know that Hortonworks is stable, but sometimes when you are talking with the customers, they wonder if Hortonworks is free, how can it be enterprise. But I explain that Hortonworks is open-source.

The solution is an open-source project. If you don't want to use their professional support services, you don't pay anything for Hortonworks and its solutions. When you want to call there or use its user interface, it's paid. This is why prefer Hortonworks solutions in my projects.

I would rate this solution 10 out of 10.

We use it for data science activities.

Data is now available.

I have no preferences towards any feature.

• Security
• Performance
• Workload management

A few of them, namely: Hive/Tez, HBase, Ranger, Yarn and Ambari. Ambari helps managing the platform, Hive is very easy to use. Ranger for security; with Ranger we can manager user’s permissions/access controls very easily.

We have successfully ported a Microsoft SSIS product application into Hadoop, that saved millions of dollars for the company and, at the same time, they are getting better performance. Also, we implemented fraud detection, as quickly as possible, for the online orders. (Fraudulent orders became a big headache for our company. The early detection of fraud is saving the company a lot of money).

Hive performance. If Hive performance increased, Hadoop would replace (not everywhere) traditional databases (Oracle/Teradata, etc.), which would save a lot of money for the company.

I have been working on this HDP platform since Jan 2015.

No, our company is a satisfied customer.

No, not at all.

Product is good. Reason I gave a rating of eight is that their community is very large and relatively very quick in bug fixes.

• Ambari Web UI: user-friendly
• Views for Hive, Tez, Pig
• Spark and Ranger

It has helped our organisation cater to clients who are using Big Data for data storage and analysis combined with our security product.

Deleting any service requires a lot of clean up, unlike Cloudera.

Five years.

Not until now.

No.

Very supportive, prompt responses.

We didn't use a previous solution.

The Ambari upgrade is not very user-friendly.

Not applicable.

Cloudera and MapR.

It's a great company with a great product employing dedicated people.

We evaluated Cloudera and Hortonworks. Based on our evaluation and actual experience in production of 60 nodes and development of 12 nodes, the most valuable features of Hortonworks are:

• 100% open
• No lock-in like Cloudera
• Fast and accurate support instantly
• Largest number of committers to Hadoop by any means
• Hive is better in performance and ease of use compared to Impala

It helps a lot in data in motion (ingestion and manage in real time). We are able to do 3rd-party data monetization of our data within a t+20 minute time frame to our end customers.

• Cost
• Reliability
• Speed
• Ease of use

I have used it for three years.

I initially encountered deployment issues, but they were very good in resolving them.

I have not encountered stability issues.

I have not encountered any scalability issues at all. That's the key reason we picked HDP over Cloudera, as Cloudera have issues & don't support compression of Hive in ORC format. They push only their products (not good).

Customer service has been excellent from the day one until now... and our Admin is comfortable with the SLA and turnaround time.

Technical support is very good and proactive with SmartSense.

We previously used a different solution. We switched from Cloudera. Initially, we went with Cloudera due to it being a popular choice in the market, etc, then realized it was bad choice. Before we scaled from 6 nodes to 12 nodes and before we went livein production, we scrapped it due to Impala's performance and lock-in.

Using Ambari, it was easy to set up and we even tried the AWS for a test cluster.

An in-house team implemented it: two admins, seven developers, one data scientist, one PM and 22 business users at the customer (end-user side).

ROI is 300%.

Hortonworks is the best, comparing all three flavors. If all is well, we might use open source alone in the next three years; others you can't due to lock-in...

Before choosing this product, we also evaluate Cloudera.

It is the best in terms of product vision and actual delivery.

• It's the one and only complete open source big data platform
• Ambari-managed admin configuration for HDFS, YARN, Hive, HBase, etc.
• Customized dashboards
• Web-based HDFS browser
• SQL editor for Hive
• Apache Phoenix - OLTP and operational analytics on Hadoop
• Apache Zeppelin - A web-based notebook that enables interactive data analytics

• Maintenance of our own data lake in the enterprise-level
• Storage and analysis of server logs
• Applying Operational Intelligence in the enterprise-level based on the analysis of various department units data
• Semantic analysis based on the feeds from social networking data, clickstream data, etc.

• Rolling upgrade
• Disaster recovery features such as mirroring should be supported

We've used it for one year.

No issues encountered.

No issues encountered.

No issues encountered.

3/10

3/10

No previous solution was in place.

It's easy to setup.

We did it in-house.

Completely use the community edition along with other features that can be implemented on top.

No other solutions were looked at.

Study, analyze, and compare with other big data platforms features according to your requirements before choosing the appropriate one.

It has a powerful, user-friendly interface called Ambari which allowed us to administrate our cluster easily.

It allows us to performa data lake implementation to handle/treat huge amounts of data, or what we call the "terrible bytes”.

Integrate a complete hive web client (Ambari views), like Hue Today, in the next release.

I've used it for three years.

The HDP v2.3 is stable release. But we have encountered some issues linked to hbase (fixed by: hbase server and region server should not be installed on the same node).

It has good documentation, but it's not fully complete for complex security needs (knox/ranger with cluster).

We have installed our first cluster using native Apache repositories.

The setup of Hadoop was easy thanks to Ambari, but installing the security components was complex.

Our cluster has been implemented in-house. We have automated the entire installation of Hadoop, set-up and configuration included.

Hadoop/Cloudera is still much cheaper than Oracle's RDBM system, if you want to handle a huge amount of data and make complex analytics. 

It's 40,000€ for 10 Hadoop nodes vs 1.7 million Euros  for an Oracle server with 40 cores.

In short, I recommend this product simply because Hortonworks is the only distribution that runs on Linux and Windows Servers.

Hortonworks is 100% Open Source. Hortonworks does a great job in managing all different components of Hadoop.

We've done multiple implementations of it. It enables customers to perform sentimental analysis from social media data to engineering analytics.

Security- Although they support Knox and Ranger and Kerberos, they are still missing attribute-level encryption features.

Name Node High Availability is still not stable (memory issues).

• Ambari
• Hive
• Sqoop
• Flume
• Spark

The Hadoop value proposition is in expanded functionality, linear scalability, and reduced software and infrastructure costs. Hadoop offers several generic frameworks for batch, real-time, and iterative processing, such as map-reduce, spark, and spark streaming. Additionally, these frameworks provide libraries for predictive analytics and machine learning. This type of expanded functionality is not easily achieved on any other single platform.

File system to provide indexed access to individual records with in-place update/delete. Also, Security integration through a common interface for authentication, authorization, disk encryption, network encryption, data access layer, data masking, etc.

Hadoop, does not provide improved performance, compared to traditional RDBMS, unless processing batches in the TB-PB range, or if the Hadoop platform has significantly more resources available.

I have implemented various flavors of Hadoop over the past five years, including platform configuration and application development.

Deployment is improving rapidly, but like other flavors of Hadoop there are always issues.

Stability is improving rapidly, but like other flavors of Hadoop there are always issues.

5/10 - Responsive, but like all flavors of Hadoop, there are too many tickets to be reasonably triaged and supported.

I have implemented various flavors of Hadoop such as Hortonworks and Cloudera over the past five years, including platform configuration and application development.

Straightforward once you know what you’re doing.

I work for a vendor team.

ROI is one of the main reasons organization pursue Hadoop. Cost per TB is a compelling factor.

Hadoop is complex. It takes a dedicated approach from individuals with a broad range of technology skills and commitment to overcome challenges that do not normally present themselves in well-established technologies.

Hadoop is complex. It takes a dedicated approach from individuals with a broad range of technology skills and commitment to overcome challenges that do not normally present themselves in well-established technologies.

The features I've found most valuable are--

• Ambari UI
• Hive
• Pig
• Hive
• Also integrated Tableau with this distribution

It's easy to deploy and we've used this distribution for some of our recommendation and trend analysis use cases.

I've used it for almost one year.

No issues encountered.

No issues encountered.

We faced some issues while upgrading to newer versions with current distributions, but with their support we solved it.

Customer service is great.

Technical support is great.

No, we did not use a previous solution.

Initial setup was straightforward.

We implemented it with our in-house team.

From a product standpoint, their Ambari UI is incredibly valuable for cluster monitoring. It simplifies the deployment and maintenance of hosts, and we can provision, configure and test Hadoop services.

From an overall perspective, Hortonworks support is crucial to our operations.

As this is open source, there are certain features that need tuning, such as the Hue UI. More stability on this would be helpful.

I've used it for one year.

As this is all new technology, we face issues at every level. However, hardware support and documentation have been instrumental in helping us resolve the majority of those issues.

We've had some issues with stability, but hardware support and documentation have helped us resolve most of those.

No issues with scalability.

They have outstanding customer support. Their responses are prompt, and they resolve issues quickly.

I have not used a solution of this nature before.

The set up is straightforward enough, but at every level there are many parameters to be tuned. Ensuring all these parameters are set is the complex part, as poorly set parameters can cause unwanted issues.

We have an in-house team to do implementations. I would advise that all implementations get seen through all the way to having users smoke test applications to ensure correct functionality.

I would suggest that if you are implementing this at an enterprise level, the support is compulsory. Additionally having a high degree of patience is key, as this is open source and road bumps can be frequent when moving at a fast pace.

There are several features that are most valuable for us--

• Hue
• Hive
• Spark
• S3

With it, we have faster processing times for our apps.

It needs to be quicker and to have the ability to automate deployment on multiple nodes.

I've used it for two years.

Sometimes there were issues.

Sometimes there were issues.

Sometimes there were issues.

I've not had to use it.

No solution had been used previously, but we are using it alongside AWS EMR.

It was complex to configure.

It was done in-house.

We provide services for product implementation, so people looking for such products can contact me.

Ease of deployment and management of the Hadoop cluster are features we've found most valuable.

It allows our organization to collect data from databases that are different, and where the data is similar, it allows for a detailed analysis from a single data store.

The ability to update data is an area where the product could improve.

I've used it for one year.

We had an issue during deployment. You have to be sure that your base image is perfect and that your infrastructure is properly configured or issues will occur.

We don't have their paid support, but I have had discussions with their engineers and they have been extremely helpful. So based on that, I would give them 8/10.

The initial setup is complex. It mainly stems from small issues that typically pop up and also a lack of experience in deploying the product. I highly suggest taking the Hortonworks Training prior to deploying.

We used an in-house team. Take your time and utilize the free resources provided by Hortonworks.

At this point I don't believe I could provide a ROI as we aren't fully utilizing the product.

If possible, I would suggest paying for the professional services which would give you on-site engineers to help deploy the cluster.

We did look at Cloudera, but due to having literally no money to spend for the project, we chose Hortonworks due to its being completely free and open source.

Take your time and script as much as you can so that all base images are the same.

• Open-source
• Big community

It is a different paradigm than standard relational databases. We can also process different tasks then just those related to the standard database world. That said, we are capable of processing various data science tasks, e.g. natural language processing or log processing.

• Stability
• It needs to be more mature
• Security
• User friendliness

I've used it for three years alongside MapR and Cloudera.

Almost every part of the Hadoop ecosystem has its problems and bugs.

Almost every part of the Hadoop ecosystem has its problems and bugs.

Almost every part of the Hadoop ecosystem has its problems and bugs.

The paid service is pretty good, but if you don't pay, there is documentation available in the community which is pretty good.

I slightly experienced Cloudera which is very similar to Hortonworks, but there are parts which are not open source. I'm working more with Hortnoworks because all its parts are open source. and my company has a longer partnership with Hortnoworks

It is easy if you have good administrators. It is also easy if you want to just play with it on your laptop. For real work and stability, I definitely recommend some paid support.

I was involved in multiple projects. Usually, it was done in-house with paid support.

Every project is different. Since Hadoop is an infrastructure for a long period there is no simple ROI. Also, each customer has different expectations.

Its flexibility is the most valuable feature because you can leverage any Hadoop component and take full advantage of its open source capabilities.

We're able to perform sentiment analysis on Twitter data.

It needs a better UI.

I used it for five months, one year ago.

It requires too much coding work; we're not good Java and Python developers.

No issues.

No issues.

They are detailed and informational.

We got stuck with Java job development and were able to get assistance from tech support.

No previous solution was used.

It was straightforward since we have Linux resources.

We did it in-house.

• Datameer
• Cloudera

You need to be tech savvy.

Its ability to scale out seamlessly with little to no effort is very valuable to us. All the tools in the stack are built from the ground up to support massive amounts of data.

It allows us to provide our customers with data insights that they previously were unable to obtain.

There have been some governance initiatives, but they are far from production ready. I would like to see a big improvement in that space, as governance is critical in many regulated industries.

I've been using it for one year.

Stability is good if configured properly, but for some tools such as for instance HBase, configuration is extremely hard to get right.

Scalability is superb.

I never interacted with customer support.

Cloudera and vanilla Big Data tech. We continue to use them alongside HortonWorks, depending on our clients preferences and needs.

Cloudera and vanilla Big Data tech. We continue to use them alongside Hortonworks, depending on our clients preferences and needs.

With Ambari, it is pretty straightforward, but I have no idea why they prefer FQDN over IP.

My colleagues and I are the implementation team. The general advice is to start out with a small enough scope. Try to get an MVP up and running before bringing out the big guns./

Licensing is on a per node basis and it encourages people to scale vertically rather than horizontally yet the whole purpose of the tools they sell is to scale horizontally. I do like that everything is also available freely for those that do not require support.

Make sure you understand what happens under the hood. Out-of-the-box tools are sub-par. Customisation is the way to go for now.

The HDFS (Java-based file system) and Hive Utilities are proving to be most useful.

Hortonworks has allowed my organization to increase the amount of data that we regularly store from sensor data and weblogs, which in turn gives us a greater scope of data to analyze.

I would like to see an increase in usability for the Apache Storm engine within the data platform.

I have been using it for less than a year.

When initializing our cluster, we did not allocate enough space to our VAR partition and that ended up causing some issues with the networking to our onsite Tomcat server.

It's fairly low customer service.

It's fairly low technical support.

We started off with this product.

Both straightforward and complex, everything was easy to set up, but a lot of the behind the scenes configuration changes for customization could be rather time consuming.

We used an in-house team. My advice is to study hard and read all the documentation thoroughly before starting any implementation. It is paramount that one understands the system before implementing it.

Current ROI is none as we are still in the POC phase with most of our products.

Be sure that the product is necessary for the situation.

There’s not only one, the all-stack of Hadoop is valuable, the distributed file system HDFS, Spark, Kafka, HBase, etc. Hortonworks has certainly got the most up-to-date version of each component of Hadoop.

Compared to the other Hadoop distributions, the Ambari server provides the user an easy way to manage, to administrate and to configure their cluster. Ambari also provides a single view that gives you the possibility to use different Hadoop components from the same web interface.

This product gives the possibility to the organization to easily and quickly install and configure a Hadoop cluster. With this cluster, the organization will be able to store and process their data and bring out some specificity on it. For example, unknown common points between their clients or key elements that will increase or decrease the churn of the client.

It would be interesting to have an easy way to implement multi-tenant for HDFS with federation. At the moment, you have to do it manually in command line.

Also, it needs to support having more than two HDFS namenodes. HDFS supports more than 2 namenodes, but Hortonworks doesn't.

I work with it in different projects and POCs for two years now.

The only issue that I had was when I tried to reinstall the software on every node. You have to manually clean up everything, as Hortonworks doesn’t provide the ability to perform a clean uninstall (software, library, log, configuration files, etc). In some case, it can generate some problems if the uninstall has not done correctly.

I never had to create a case at the support, so I don’t know. I always find the answers to my questions on the web (forum or blog). There’s a big community that can support you.

I also used Cloudera, MapR, and Microsoft HD Insight.

The first time, I didn’t know anything about Big Data and Hadoop, so yes it was difficult because I did not clearly understand what I was doing.

The implementation was at the clients datacenter. My advice is to perform a POC on premise or via a virtual machine to learn how to use it and how to tune the configuration of each Hadoop component.

When implementing it in production, firstly you need to have a clear view of the requirements you need to perform the install. For example, if you are using a local repository to install the software, it has to be updated with Hortonworks sources, especially if there are security rules (firewall access, root access limitation, etc.).

My last piece of advice is that if you have a heavy load, it is really important to implement the solution on premise, not in a virtualized environment. If you do both, you will see the difference in performance.

The use of Hortonworks is free there’s no license but if you want there’s a support. It’s up to you to see if you need it (certainly) and to maybe negotiate it.

I did not really made the choice, as the client made it dependent on their experience, functionality of each distribution, privacy of the data and the licensing/support price.

Firstly perform a POC to learn and to get an idea of the load of your future applications. Then, you should be able to correctly design the need infrastructure.