The best part about using Spark on HDInsight is the seamless integration within the Azure ecosystem. It allows for effortless scaling of compute power to handle massive datasets. The managed nature of the service means I don't have to worry about the underlying infrastructure overhead, and the Jupyter Notebook integration makes iterative development and data exploration extremely efficient for our engineering team Review collected by and hosted on G2.com.

One downside is the cluster spin-up time, which can feel slow when you need immediate compute for ad-hoc tasks. Additionally, while it is a robust managed service, the pricing can escalate quickly if clusters are not managed or scaled down properly during idle times. It also feels slightly less 'modern' compared to newer alternatives like Azure Databricks, particularly regarding UI and collaborative features Review collected by and hosted on G2.com.

Sparks fast computing capability allows for a more interactive experience. One of the other capabilities of Spark on HDInsight I love involves the Jupiter notebooks which allows for the extensive exploration of data using SQL, Python or Scala. This allows for quick querying and the running of machine learning algorithms.The cons for me using this software are that Spark MLIB covers basic machine learning such as regression or clustering however it doesn't have the capabilities for modeling and training data in natural networks. I also find it hard to run large volumes of data at one time. I wish there were a way to process large amounts of data without having to restart from scratch once it crashes. Review collected by and hosted on G2.com.

The cons for me using this software are that Spark MLIB covers basic machine learning such as regression or clustering however it doesn't have the capabilities for modeling and training data in natural networks. I also find it hard to run large volumes of data at one time. I wish there were a way to process large amounts of data without having to restart from scratch once it crashes. Review collected by and hosted on G2.com.

Apache provides quick overview of data analytics and cluster computing using in memory processing. we use Resource management to create the HDinsight spark cluster. it is good for faster data analytics. Review collected by and hosted on G2.com.

I wish there would be quick solution with one click less then 10 second processing window because all the customers are expecting for the faster result as you know Cloudera is working towards it. Review collected by and hosted on G2.com.

I like the dynamism nature of it by having ability to support multiple languages. Review collected by and hosted on G2.com.

There is no file management system in the system and do needy for integration with other platforms. Review collected by and hosted on G2.com.

What is most helpful about Apache spark for azure hdinsight is the options it provides to make to do easy Review collected by and hosted on G2.com.

8 times I feel like the their process was not user-friendly Review collected by and hosted on G2.com.

I enjoy the environment that the data lives in. Its really clean and intuitive Review collected by and hosted on G2.com.

The way you input the data is easy enough for all users to do and be able to incorporate the data Review collected by and hosted on G2.com.

Its great i would highly recommend for all business needs Review collected by and hosted on G2.com.

There is not anything that i can think of at this moment that i even dislike Review collected by and hosted on G2.com.

Its very easy to implement and integrate the spark jobs on Azure HDInsight cluster and lot of options to load or store your data building spark data pipelines. Review collected by and hosted on G2.com.

The configurations or creating links for various sources to load data. Review collected by and hosted on G2.com.

The most important thing about Apache Spark is that 10X faster than another data analysis platform as Hadoop. Review collected by and hosted on G2.com.

There is nothing in particular, expect how to deal with data object that requires a different approach as you build your data processing workflow. Review collected by and hosted on G2.com.