Denodo allows us to create and combine new views to create a virtual repository and APIs without a single line of code. It is excellent because it can present connectors with a view format for downstream consumers by flattening a JSON file. Reading or connecting to various sources and displaying a tabular view is an excellent feature. The product's technical data catalog is well-organized.
Well suited for my big data related project or a static data set analysis especially for uploading huge dataset to the cluster.
But had some issues with connecting IoT real-time data and feeding to Power BI. It might be my understanding please take it as a mere comment rather than a suggestion.
Caching - but I am sure it will be improved by now. There were times when we expected the cache to be refreshed but it was stale.
Schema generation of endpoints from API response was sometimes incomplete as not all API calls returned all the fields. Will be good to have an ability to load the schema itself (XSD/JSON/Soap XML etc).
Denodo exposed web services were in preliminary stage when we used; I'm sure it will be improved by now.
Export/Import deployment, while it was helpful, there were unexpected issues without any errors during deployment. Issues were only identified during testing. Some views were not created properly and did not work. If it was working in the environment from where it was exported from, it should work in the environment where it is imported.
Easier pricing and plug-and-play like you see with AWS and Azure, it would be nice from a budgeting and billing standpoint, as well as better support for the administration.
Bundling of the Cloud Object Storage should be included with the Analytics Engine.
The inability to add your own Hadoop stack components has made some transfers a little more complex.
Denodo is a tool to rapidly mash data sources together and create meaningful datasets. It does have its downfalls though. When you create larger, more complex datasets, you will most likely need to cache your datasets, regardless of how proper your joins are set up. Since DV takes data from multiple environments, you are taxing the corporate network, so you need to be conscious of how much data you are sending through the network and truly understand how and when to join datasets due to this.
We initially wanted to go with Google BigQuery, mainly for the name recognition. However, the pricing and support structure led us to seek alternatives, which pointed us to IBM. Apache Spark was also in the running, but here IBM's domination in the industry made the choice a no-brainer. As previously stated, the support received was not quite what we expected, but was adequate.
This product has allowed us to gather analytics data across multiple platforms so we can view and analyze the data from different workflows, all in one place.
IBM Analytics has allowed us to scale on demand which allows us to capture more and more data, thus increasing our ROI.
The convenience of the ability to access and administer the product via multiple interfaces has allowed our administrators to ensure that the application is making a positive ROI for our business users and partners.
