What is most valuable?
The most important aspect of the product is the ability to work with almost any data source or target. I can readily develop processes that use relational, file, XML, JMS messaging, web and big data sources or targets.
I can control the style of integration through "knowledge modules", and if they don’t do exactly what I want. I can write my own or customize the Oracle supplied ones.
The ability to execute third-party (or in-house developed) Java code by installing JAR files allows a great deal of flexibility; for example, I can add custom processors to do access web APIs that use token based authentication.
Another key feature is that we do not need to pass our data through some form of ETL engine hosted on a server; in some cases for example transforming data within a data warehouse all of the processing is done in single SQL queries thus reducing network traffic.
Finally, the rich SDK supplied with ODI allows developers to create virtually any form of development or deployment automation.
How has it helped my organization?
It allows a single skill set to handle virtually all of the data transport and transformation needs of a company. It moves the ETL processing to where the data resides and saves network traffic and the need for dedicated transformation engines - hardware that needs to be purchased, managed and supported.
What needs improvement?
There needs to better support of external version control software, currently only SVN is present, but I hear the GIT is planned in future patch.
From a DevOps point of view it would be useful to add better separation between ‘code’ and ‘executable’ exports, at the moment a piece of code will contain the embedded executable which bloats any source control objects (this may only be relevant to those the develop their own source and code control processes).
For how long have I used the solution?
I've been using it for the last 10 years.
What was my experience with deployment of the solution?
We've had no issues with deployment.
What do I think about the stability of the solution?
12cR2 is a new version, and there are a few issues with stability, but I expect most to be resolved in the first patch.
What do I think about the scalability of the solution?
There have been no issues with the scalability.
How are customer service and technical support?
I rarely engage with support.
Which solution did I use previously and why did I switch?
As an independent consultant, I work with other ETL products, and switching is requirement of my employers. However, of the products I used recently (Informatica and Talend), I feel ODI gives me the most flexibility.
How was the initial setup?
Set up of ODI Studio and the ODI repository is relatively simple, it is all done through a single JAR file executable. The complexity comes when you need to create ODI agents - there are three flavors of agent and the best choice of agent will depend on your agent management needs and infrastructure. The ODI agent executes ODI code and interacts with the host OS, typically one agent is sufficient, but more may be needed.
What's my experience with pricing, setup cost, and licensing?
The licensing model has changed a few times over the years - read the Oracle price list or speak to sales.
What other advice do I have?
Plan your ODI infrastructure, especially where data is transformed to ensure you get the best balance between license costs and performance. Get your developers trained in best practices so that avoid unnecessary pit falls.