I used Denodo for a few months early this year, from January to March. My main use case for Denodo during that period was visualizing data from Snowflake and other databases across multiple databases and doing various types of analytics on the data inserted in those databases. We used Denodo for getting the data across multiple databases and matching the data in between these databases to have a single entry point for our queries.

For example, we had a bunch of files arriving on Azure Data Lake Storage, and we needed to match when these files arrived to what we needed to do with those files that were stored in a Snowflake instance. So, we used Denodo to cross this data and select queries, and then apply the Azure function that we needed to do to these files that were arriving on our Azure storage.

Denodo offers a great query language, the VQL, which was very easy to understand for someone that only knows SQL. It was easy to work with, and it was easy to integrate various types of databases in general. The usability of the platform is very great. The federated queries were also a feature that I used a lot. The easy usability of Denodo helped my work and made things easier for productivity. We were able to ship code faster than if we needed to use another type of data visualization tool or even an ETL.

When we have multiple databases, the standard or common approach is just to have an ETL that matches the data, and we need to write this whole ETL. With Denodo, we can simplify this whole work to a few queries or setups in the databases and just make a few queries. This was actually easier, and it was faster to deliver a solution for a client instead of building a whole ETL pipeline or simply backfilling one database in another. Denodo simplifies your work and helps you ship code or ship solutions in a faster way. I really like the UI that Denodo has on the web app version. It was very simple and actually very beautiful to see. For a complicated program, it was very easy to understand how it works, even for someone who is just approaching the data engineering career or side of the programming language.

The desktop version of the Denodo app has a difficult UX to understand, touch, and set up. But they have fixed that on the web app version, which is nice. Overall, the platform works great with the current setup that they have. I didn't really get to know the quirks and the bad side of the platform, as I've only used it for three months. In my opinion, it's a great software, and it's not buggy at all. The interface is simple and easy to understand.

I've been working in my current field for eight years right now.

Denodo is stable.

The customer support for Denodo is great. They are very easy to work with, and all my questions and concerns were answered in a very short amount of time.

Positive

I didn't previously use a different solution before Denodo. When I saw Denodo for the first time, I thought it was great, but it was not my choosing to use Denodo; it was the standard platform for the company I was working at.

We use Microsoft Azure as the cloud provider for Denodo's private cloud deployment.

I can't share any metrics with you because I was just a developer. But my general impression is that yes, Denodo saves a lot of employees and development time. Because if you don't need to write a whole ETL to make the data virtualization, you need way fewer people to write a query instead of writing an ETL pipeline.

I didn't really manage the setup cost, licensing, and pricing for Denodo. But in my opinion, it's very expensive. For small companies, it's not a solution that most small companies can afford. But for big companies, the pricing seems fair. It would be nice to have a cheaper version of Denodo for small businesses.

My main advice is to read the documentation for Denodo. It's a very powerful solution, but it has a lot of quirks and a lot of things that can be done and should be done the right way to have better performance when using Denodo. The documentation is very clear and will help you 99% of the times. The review rating is 9.

Denodo is designed for a distributed data fabric. Instead of a centralized data lake, we are in a distributed data mesh. Whenever a team wants to publish some data product, they are publishing data through Denodo, and their data store might be different; it might be in Azure SQL, in an Azure Databricks data table, or in ADLS.

Whenever a user is looking for data, instead of connecting to each individual data source, they can only pull data through Denodo views, which is a virtualization layer. Role-based access control can be implemented. Not everybody can pull data; whoever is looking for data and whatever data they are looking for, that approval we can give, and based on that, they can pull.

For real-time analytics, we have an Azure Kafka endpoint. There are some use cases with the Azure Kafka endpoint; we have transferred data through that, and through Databricks, we have transformed the data and stored it in some persistent layer such as a Databricks data table or any other data store.

One of the best features of Denodo is that it's one unified layer. All programmers or software need to manage different endpoints through one Denodo endpoint. We have custom-developed a tool called Marketplace where we can see whatever data products are available. It is centralized through one portal, a one-stop shop.

Denodo supports SQL base, so if you want to join two tables or two views, you can use SQL, which is an advantage as most developers or business people know SQL. The distributed data mesh functionality allows teams to publish data products through Denodo, regardless of their data store type.

Role-based access control is another valuable feature, allowing selective data access based on approval. The Azure Kafka endpoint enables real-time analytics capabilities.

In terms of improvements for Denodo, regarding performance, in cases where there are multiple virtualizations—such as reading from one Denodo view that is virtualized, and from that view there's also virtualization, and another team is reading from that view—if multiple virtualizations happen with no caching in between, it becomes slow.

This occurs because it is cascading; whenever at the top level someone is reading data, that request is getting cascaded to the nth level, causing issues, especially in cases such as Power BI reports. We need to consider implementing some persistent layer in between.

The scaling process should improve because many things are getting automated. The scale-out part needs to be automated, though I am uncertain whether Denodo has already implemented that feature.

I have been working with Denodo since joining this organization, which is two and a half years. I believe the organization has been using Denodo for five years.

Denodo is scalable and is installed on a cluster, enabling scale-out capabilities. However, it cannot be scaled on the go. For huge data requests, it cannot scale automatically; admin action is required. The scaling process should improve as many things are becoming automated. The scale-out part needs to be automated, though I am uncertain whether Denodo has already implemented that feature.

I would rate Denodo's technical support an eight out of ten because whenever there is planned maintenance or if the Denodo server is down, we receive emails from the support team. If we raise a ticket, it can be resolved or addressed within a reasonable time frame, so support is good.

Positive

I am not certain what was used before Denodo as they had started using it before I joined. At that time, they might have had some legacy systems, and each team had their own ETL pipeline.

We are not using AWS, and a different team is responsible for onboarding or approving new software or licenses. Regarding advanced metadata management features, someone from the Denodo SME or Denodo admin team would be better positioned to comment.

We have used the caching mechanism. For huge data sizes, we are using Snowflake. Denodo has the feature to persist data in any data store. While Denodo doesn't typically persist data as it's a virtualization layer, we can persist data in Snowflake for large datasets.

Denodo can be deployed both on-premises and in the cloud. Initially, they started on-premises with the Denodo server, but have now migrated to the cloud. Denodo is the backbone for the BU I'm working for. If someone wants to share data or bring data, they need to go through Denodo, making it the backbone of this data mesh architecture.

I am not familiar with Denodo's pricing as that is handled by the budget or senior management team. We might be a partner of Denodo, but I'm not certain about this relationship. We are using Databricks Spark, not core Apache Spark.

I would rate the product overall an eight out of ten. I am satisfied, but there is always scope for improvement. If they fix scalability automation and virtualization problems, I would give them a higher rating.

I use Denodo to create a virtualized data environment rather than a physical one. This is helpful when enterprise data is spread across different data sources, such as cloud data warehouses and data lakes, and is used for developing AI and ML strategies. It creates a unified virtualized environment so the algorithms do not need to connect to multiple data sources individually.

Denodo creates a virtualized environment that provides one unified destination for all data sources so that AI and ML algorithms can connect to that virtualized destination without needing to establish connections with multiple data sources.

The most valuable feature in Denodo is data virtualization. This feature focuses on creating a virtualized data environment that allows for simpler connections between data sources and systems. It significantly eases the process of developing AI and ML strategies by centralizing access to various data sources and allowing algorithms to efficiently connect to and utilize data.

Denodo's complexity level is high since it connects multiple sources, sometimes creating glitches in the production environment. The stability of the features can be improved. The system has dependencies on other environments, like JVM, which can affect its performance.

I have been using Denodo for about two to three years.

Denodo shows stability concerns due to its dependency on external environments, such as JVM. If JVM does not function properly, it may cause Denodo to fail to connect to different sources.

I would rate Denodo's scalability as six to seven. Its complexity in configuring and the requirement to install different connectors for different sources affects its scalability.

Denodo's technical support is rated as seven. Although they are proactive in conducting webinars and connecting with customers, they are perceived as less proactive compared to some other vendors.

Neutral

Configuring Denodo involves setting up a Java environment and installing different connectors for various data sources. This can take one or two weeks depending on the number of sources.

Denodo's pricing is not affordable for small companies and is more suitable for medium to large enterprises. On a scale of one to ten, where ten is most expensive, it is rated around three to four in terms of cost.

For new users evaluating Denodo, it is crucial to determine whether you truly need a virtualized environment. If you do not have numerous data sources, Denodo might not be necessary. However, if you do, Denodo makes sense. 

I rate the overall solution as eight out of ten.