Embedded business intelligence is a new requirement for custom and commercial applications. Developers have different reasons for choosing to integrate commercial or open source BI. This article on searchDataManagement.com
examines what those reasons are.
Developers that want to embed business intelligence (BI) and reporting features into applications must find a balance among requirements, cost and development time.
Embedded reports are a standard requirement of most applications, according to the developers interviewed for this article. But users are increasingly demanding more sophisticated reporting from applications -- seeking such features as custom report design, ad hoc report creation, analytics, and even performance management. Independent software vendors (ISVs) and application developers have choices ranging from open source tools to high-end commercial BI platforms.
When it comes to BI, are you using commercial or open source? How has your experience with open source BI been?
At the mortgage company I currently work for the reporting is king. We ended up going with a comcercial solution (Oracle Reports and Discoverer) for two main reasons first. For the first two years of getting the company going the report requests averaged about 5 a week (The application now consists of over 200+ reports). This doesnt seem like a lot but that leads me to reason two. At least half of the reports were based on complex criteria and had very specific layout requirements.
We wanted to be able to have the business anaylists be able to write the reports as they were the ones that best understood the data and user requirements. From our testing we found that Oracle BI tools were comprehensive enough to allow the business anaylists to do most of the work.
Im not going to defend the fact that the Oracle Reports Designer tool can be unstable at times and there have been a few times that reports needed to be blown away and recreated from scratch. This being said the amount of money/time we saved from freeing up developer resources was worth it.
Besides the mortgage company I have found Jasper to be a great asset and have used it on in amost every case besides the one listed above.
Why we went opensource
My company (loyaltymatrix) started out using a closed source solution on top of MS olap. We ended up being the QA team for that closed source product - and we paid them! At first they sent us custom patches to address issues. Eventually we got ignored. Very frustrated we sought out opensource replacements. Jpivot, mondrian, and jasper were all key components. Nothing brought them together at the time, so we had to do that work, then opened it up as openi.
Depends on how "embedded" you want you BI application. You can get some mileage out of embedding closed source BI suites as long as that embedding/integration is already available. (the article mentions onyx CRM + cognos). If you are embedding a licensed product inside of another licensed product, you are incurring significant license fees before you even implement a single report.
But, if you have your own custom application, opensource provides most of the components you need: mondrian, jpivot, jasperreports, jfreecharts. Or use a project like openi which tries to bring these together projects together bundled with additional security/portal features.
If support is a key decision point, this is still available with all of your opensource BI players - jasper (jaspersoft), openi (loyaltymatrix), pentaho all have support contract options available. Use opensource, buy some support, save some money. Use that savings towards DBA, business analyst resources: just because you have business intelligence software, does not mean you have intelligence. There is no software available today that will give you intelligence without a smart team creating reports, and getting meaning out of those reports.
so we have reached the commoditization point?
Pay yourself or pay someone else.
At "Yellowfin Reporting", Our experience with embedded BI probably was typical for most developers – we needed a relatively sophisticated reporting front end for both ad hoc and pre-canned reports for our commercial HR products – but commercial products at the time were either too costly or really did not provide the seemless integration we were looking for. So naively we we embarked to build a reporting platform as we wanted it to be.
Why did we do this?
1. Open Source Options do not address a typical business users needs –for embedded BI the number one priority has to be usability – can the end user author a report, especially an ad hoc one?
2. Commercial products could be integrated but not embedded – and cost a small fortune.
3. As a hosted solution we had significant concerns about security and needed to control access to data.
4. Flexibility – we knew that what ever we delivered our customers were going to want to change the report content or layout.
What we did:
We designed our ideal reporting application, used some open source components such as JFreeChart, iText, and did integrate "Jasper" as one of the ‘advanced’ authoring options. Developing the interface is where we spent a huge amount of time and effort – far more than the integration of components. What we ended up with was an entirely new product – a Reporting application that not only could be embedded into our app but into any other one as well.
Our vision is that Yellowfin provides a viable commercial alternative – cost effective, highly embeddable BI solution.
Our next release will also integrate BIRT to further extend the products capabilities.