Share your experience using Oracle Product Hub

My main use case for Cloudera Data Platform is for data analytics and AI workload.

We have different data sources where the data is coming in tabular format or CSV, semi-structured or structured, unstructured, and some sort of Kafka streaming messages. We use to store it and then we process and transform, apply the business logic, and then make the data ready for the consumer to consume.

Cloudera Data Platform offers excellent architectures in terms of decoupling the storage layer from the compute. It is flexible in terms of scaling to your storage account or compute. Additionally, we have different streaming services as part of the ecosystem, and they have added Ranger for security controls, which is a valuable feature.

Decoupling storage from compute has helped my team significantly. Before using Cloudera Data Platform, we were using Cloudera Distribution for Hadoop (CDH), where we had to have on-premises virtual machines or Linux boxes to add to the cluster, which required lots of effort. We had defined authorized maximum storage per system; for example, one computer can have a maximum of 8 TB, and scaling up to add more compute to the cluster was very challenging. In the current Cloudera Data Platform, the backend storage is a data lake that auto-scales, so we don't have to add more storage. In terms of security, we used to use Sentry in traditional CDH, but in Cloudera Data Platform, Ranger provides more granular level of security, allowing us to manage who can access data at different levels, maybe at a tabular level or column level.

Streaming services are provided by NiFi, which is one of the best ecosystems for streaming and ETL support.

Cloudera Data Platform has positively impacted our organization by reducing overall manual intervention, requiring fewer efforts and resources to build a big data cluster compared to traditional methods. It is also cost-effective and more stable than the traditional ways of handling big data workload.

In terms of resources, we have reduced from ten resources to four or five resources, making it an effective reduction in manual effort. Regarding cost saving, since we are in the cloud, we are saving significant money compared to maintaining infrastructure on-premises.

Cloudera Data Platform could improve by innovating more in terms of full-fledged support for AI workloads, enriching machine learning or LLM, as there haven't been updates in that aspect over the last one and a half years.

I have been using Cloudera Data Platform for almost four years.

Cloudera Data Platform's scalability is very good.

Customer support is good. However, having a common chat channel between firms and service providers would make communication faster and more efficient.

Neutral

My advice to others looking into using Cloudera Data Platform is that if they are looking for big data workloads on the cloud where they can do analysis and achieve cost savings and resource reductions, it is definitely a good use case. It can vary based on business needs, but it is a good option for big data workloads.

I rated Cloudera Data Platform a six out of ten because I wish that it would keep up with market trends and release AI technology and AI-enabled workloads. Sometimes we struggle to get support, and having a common chat channel between firms and service providers would make communication and support more effective, especially in production.

My main use case for Cloudera Data Platform is data analytics and AI.

For data analytics and AI in my day-to-day work, we have a multi-source system where the data keeps coming from different source systems, from RDBMS, in tabular format, or semi-structured, or streaming data from Kafka. We process and store data in the backend ADLS, then apply business rule logic to create a golden table which is published for business or end users who consume the data for analytics. Some AI engineers develop or run that code, Python code, or LLM against those data to gain insights.

The most unique feature I love about Cloudera Data Platform is its integration with Ranger services. Ranger is more flexible compared to Cloudera's previous data distribution component, Sentry, making it more reliable and allowing for access control at a more granular level.

The Ranger integration makes it more flexible and reliable for me by allowing control over data access, specifying who can access at what level, such as table level, masking, or data layer level. This is crucial for managing all data inside the farm.

In terms of integration, it is very easy with Cloudera Data Platform. We just hook it up since it comes with the package when we install the CDP runtime, allowing us to select the ecosystem we want in our farm depending on our use cases. It is not a standalone installation requirement; it is an easy job. Scalability and flexibility are very good.

From a holistic view in the market, I have not seen enough innovation in Cloudera Data Platform, particularly in support for machine learning. It supports it, but not to a robust extent compared to other tech providers, such as Databricks, which are more flexible and in tune with current trends in AI and machine learning. I wish Cloudera would innovate and keep pace with market demands.

Regarding the user interface of Cloudera Data Platform, I have not faced any challenges, though we definitely look forward to innovation to support varied data models and scalability.

I have been using Cloudera Data Platform for almost four years.

Cloudera Data Platform is generally stable; however, we occasionally face minor network connectivity issues as confirmed by the vendor. Sometimes a node goes down, but it automatically returns to a healthy state.

Cloudera Data Platform has positively impacted my organization by eliminating challenges we faced with CDH, which had not been supported for a cloud journey. When adding scalability, such as horizontal scalability to our existing cluster, the process was time-consuming and required upfront costs for procuring servers. In contrast, CDP allows for easy, mostly automated scalability where I can schedule job workflows, fine-tune system resource metrics, and add nodes with just a click.

Customer support depends on the case severity, but from my experience, it is great. Cloudera support is timely and responsive, adhering to the SLAs they provide.

Neutral

Previously, we used Cloudera Data Distribution, known as CDH, which was on-premises and required more manual efforts among multiple teams, taking almost a month to set up a cluster. We switched primarily for cost-effectiveness, flexibility, and the reduced time required for setup.

Cloudera Data Platform has positively impacted my organization by eliminating challenges we faced with CDH, which had not been supported for a cloud journey. When adding scalability, such as horizontal scalability to our existing cluster, the process was time-consuming and required upfront costs for procuring servers. In contrast, CDP allows for easy, mostly automated scalability where I can schedule job workflows, fine-tune system resource metrics, and add nodes with just a click.

A solution architect from the vendor helps us resolve any ongoing issues such as bugs or vulnerabilities, and we appreciate the flexibility of the cloud journey.

In terms of return on investment, I see great changes in operational effectiveness measured by RTO when comparing on-premises solutions with cloud solutions. The difference is notable.

I have not been involved overall in cost negotiation, but we find Cloudera Data Platform to be cost-effective. We work with the Cloudera vendor to secure one or two-year licenses upfront for discounts.

We evaluated Databricks three years ago, but it was not up to market standards in feature support at that time, particularly lacking an account console, which was introduced afterward. We have seen clients migrating from Cloudera to Databricks since the rollout of that console.

My advice for those considering Cloudera Data Platform is to evaluate their business use case and budget, as these two factors are crucial. If the organization does not need advanced features such as LLM or machine learning, Cloudera Data Platform may be suitable. However, based on the current market, if rating between Databricks and Cloudera, I would give Databricks a one and Cloudera a two.

There are lots of challenges I face while using Cloudera Data Platform. Sometimes, vulnerabilities depend on which version of CDP runtime I am using, so we work with the Cloudera vendor side to remediate any vulnerabilities based on that version. Along with that, we use it for data audit purposes, gathering all inflow data such as how data is being used, who has access, and how many times.

In terms of cost savings with Cloudera Data Platform, moving from on-premises to cloud is very cost-effective. We can use bare metal servers or on-spot servers, which makes it economical. In performance, it is superior to previous versions since multiple Spark versions are added to the CDP runtime, improving data distribution, handling, and fault tolerance, requiring no code fine-tuning.

I rate Cloudera Data Platform six out of ten.