Apache Hadoop vs. Apache Spark vs. IBM Cloudant

Altogether, I want to say that Apache Hadoop is well-suited to a larger and unstructured data flow like an aggregation of web traffic or even advertising. I think Apache Hadoop is great when you literally have petabytes of data that need to be stored and processed on an ongoing basis. Also, I would recommend that the software should be supplemented with a faster and interactive database for a better querying service. Lastly, it's very cost-effective so it is good to give it a shot before coming to any conclusion.

Incentivized

Well suited: To most of the local run of datasets and non-prod systems - scalability is not a problem at all. Including data from multiple types of data sources is an added advantage. MLlib is a decently nice built-in library that can be used for most of the ML tasks. Less appropriate: We had to work on a RecSys where the music dataset that we used was around 300+Gb in size. We faced memory-based issues. Few times we also got memory errors. Also the MLlib library does not have support for advanced analytics and deep-learning frameworks support. Understanding the internals of the working of Apache Spark for beginners is highly not possible.

Incentivized

Our organization found Cloudant most suitable if One, a fixed pricing structure would make the most sense, for example in a situation where the project Cloudant is being used in makes its revenue in procurement or fixed retainer — thus the predictability of costs is paramount; Two, where you need to frequently edit the data and/or share access to the query engine to non-engineers — this is where the GUI shines.

Incentivized

• Handles large amounts of unstructured data well, for business level purposes
• Is a good catchall because of this design, i.e. what does not fit into our vertical tables fits here.
• Decent for large ETL pipelines and logging free-for-alls because of this, also.

Incentivized

• Rich APIs for data transformation making for very each to transform and prepare data in a distributed environment without worrying about memory issues
• Faster in execution times compare to Hadoop and PIG Latin
• Easy SQL interface to the same data set for people who are comfortable to explore data in a declarative manner
• Interoperability between SQL and Scala / Python style of munging data

Incentivized

• For us, performance and scalability is the key, and Cloudant DB backed by CouchDB is scalable and performant.
• IBM Cloudant dB is very easy to provision for sandbox, development, QA as well as production.
• Support for Java for CouchDB app server analytics enables a greater control for over developers.
• Schema free oriented very easy to program and build applications on it.
• We love it!!

Incentivized

• Less organizational support system. Bugs need to be fixed and outside help take a long time to push updates
• Not for small data sets
• Data security needs to be ramped up
• Failure in NameNode has no replication which takes a lot of time to recover

Incentivized

• Memory management. Very weak on that.
• PySpark not as robust as scala with spark.
• spark master HA is needed. Not as HA as it should be.
• Locality should not be a necessity, but does help improvement. But would prefer no locality

Incentivized

• It was only after we went with the cloud-based solution that IBM rolled out an on-premise version.
• We found that a 3rd-party ODBC driver was required for a few applications that needed to pull data out of Cloudant.
• The sales process was difficult because the salesperson we used was not as versed on Cloudant as I had hoped.

Incentivized

Hadoop is organization-independent and can be used for various purposes ranging from archiving to reporting and can make use of economic, commodity hardware. There is also a lot of saving in terms of licensing costs - since most of the Hadoop ecosystem is available as open-source and is free

Capacity of computing data in cluster and fast speed.

the flexibility of NoSQL allow us to modify and upgrade our apps very fast and in a convenient way. Having the solution hosted by IBM is also giving us the chance to focus on features and the improvement of our apps. It's one thing less to be worried about

Incentivized

As Hadoop enterprise licensed version is quite fine tuned and easy to use makes it good choice for Hadoop administrators. It’s scalability and integration with Kerberos is good option for authentication and authorisation. installation can be improved. logging can be improved so that it become easier for debugging purposes. parallel processing of data is achieved easily.

Incentivized

If the team looking to use Apache Spark is not used to debug and tweak settings for jobs to ensure maximum optimizations, it can be frustrating. However, the documentation and the support of the community on the internet can help resolve most issues. Moreover, it is highly configurable and it integrates with different tools (eg: it can be used by dbt core), which increase the scenarios where it can be used

Incentivized

It's mostly just a straight forward API to a data store. I knock one off for the full text search thing, but I don't need it much anyways. Also, the dashboard UI they give is pretty nice to use. It provides syntax-highlighting for writing views and queries are easy to test. I wish other DBs had a UI like this.

Incentivized

it is a highly available solution in the IBM cloud portfolio and hence we have never had any issues with the data base being available - we also do continuous replication to be on the safer side just in case some thing goes awry. We also perform twice a year disaster recovery tests.

Incentivized

very easy to get started and is very developer friendly given that it uses couchDB analytics. It is a cloud based solution and hence there is no hardware investment in a server and staging the server to get started and the associated delays/bureaucracy involved to get started. Good documentation is also available.

Incentivized

It's a great value for what you pay, and most Data Base Administrators (DBAs) can walk in and use it without substantial training. I tend to dabble on the analyst side, so querying the data I need feels like it can take forever, especially on higher traffic days like Monday.

Incentivized

1. It integrates very well with scala or python. 2. It's very easy to understand SQL interoperability. 3. Apache is way faster than the other competitive technologies. 4. The support from the Apache community is very huge for Spark. 5. Execution times are faster as compared to others. 6. There are a large number of forums available for Apache Spark. 7. The code availability for Apache Spark is simpler and easy to gain access to. 8. Many organizations use Apache Spark, so many solutions are available for existing applications.

Very happy by the commitment given by the team which has been really good over the last 7 years of usage.

Incentivized

Hadoop is a complex topic and best suited for classrom training. Online training are a waste of time and money.

online resources are good enough to understand but there is nothing like testing. In our case, we discovered some not documented behavior that we take in count now. Also, the experience in NodeJs is critical. Also, take in count that most of the "good practices" with cloudant are not in online courses but in blogs and pages from independent developers

Incentivized

• Test the architecture on CouchDB helped us to address initial design flaws.
• The migration to Cloudant as such was very painless.
• We have migrate our replication system to Cloudant Android Sync for mobile devices.
• We have regular informal contact with the Cloudant leadership to discuss our use cases and implementation strategies.

Incentivized

Not used any other product than Hadoop and I don't think our company will switch to any other product, as Hadoop is providing excellent results. Our company is growing rapidly, Hadoop helps to keep up our performance and meet customer expectations. We also use HDFS which provides very high bandwidth to support MapReduce workloads.

Incentivized

Spark in comparison to similar technologies ends up being a one stop shop. You can achieve so much with this one framework instead of having to stitch and weave multiple technologies from the Hadoop stack, all while getting incredibility performance, minimal boilerplate, and getting the ability to write your application in the language of your choosing.

Incentivized

The feature-set, including security, is very comparable. Overall, IBM's services added to the product are mature and stable, although product support and engineers could be a little better. Global availability is improving, and Disaster Recover Capabilities are great. Overall, it's very comparable to MongoDB as a DBaaS offer, available globally and with great documentation.

Incentivized

The service scales incredibly well. As you would expect from CloudDB and IBM combination. The only reason I wouldn't score it a 10 is the fact that document trees can get nested and nested very quickly if you are attempting to do very complex datasets. Which makes your code that much more complex to deal. Its very possible we could find a solution to this problem with better database planning to begin with, but one of the reasons we chose a service over a self-hosted solution was so we could set it up quick and forget about it. So we weren't going to dedicate a team to architecture optimization.

Incentivized

• There are many advantages of Hadoop as first it has made the management and processing of extremely colossal data very easy and has simplified the lives of so many people including me.
• Hadoop is quite interesting due to its new and improved features plus innovative functions.

Incentivized

• Business leaders are able to take data driven decisions
• Business users are able access to data in near real time now . Before using spark, they had to wait for at least 24 hours for data to be available
• Business is able come up with new product ideas

Incentivized

• IBM Cloudant is very secure and we never have to worry about losing data/unauthorized access
• It is one of the best data backup system and works well
• Global availability means it is easy to connect to the nearest data center and this reduces load time which is great.