Written by GoodData Author |
Editor's Note: This article was originally featured as a guest post on the BI Software Insight blog on 5/27/14.
With today’s increasing variety of data, business intelligence solutions are challenged in ways not previously imagined as they strive to provide value to organizations. Many BI solutions will suffer when challenged with integration and openness requirements that allow data, application and workflow sharing that was simply not defined nor understood when the product was first written.
With so many integration alternatives, each offering specific functionality like data access, visualization extension, or functional automation, what are the key capabilities to look for when a vendor says, “Open and Flexible.”
Of course, programmers recognize these terms as meaning, “You can program against our APIs.” But what about limitations or shortcomings? Beyond re-skinning and white labeling, what else should we use to evaluate the value of vendor APIs?
Every vendor says, Any Data… But is that Any Data in or All Data Everywhere?
Most vendors rely heavily on standard data access protocols like JDBC, certainly dependable for most on-premise data. Additionally, they will offer web-services, or REST-based APIs to capture data flows from cloud or other API-accessed sources.
Interestingly, most cloud-based sources are accessible only via API, which creates possible problems in increasing fragility and management complexity as each cloud system is added.
For example, who is responsible when a third party API changes? Here, cloud vendors have an advantage in that they are likely managing connections to cloud sources like Salesforce, Yammer or Twitter on behalf of more than one customer in one place, monitored around the clock.
The most interesting data-related API function, however, is a BI environment’s ability to act as a data source itself for some other use. Most vendors are reluctant to provide access to data in their systems, as this reduces their control over the customer’s implementation. So, look deeper into vendors’ data sharing features, and control all your data.
2) Find APIs that foster a developer ecosystem
When a vendor says, “You can do that with our platform…,” what they really mean is, “we have not implemented that yet, but you can program it yourself against our APIs.”
Most platforms offer programmatic capabilities, either by including a programming tool within their environment or integrating with one. In both of these cases the program extensions that are written are usually executed externally, perhaps in an external JVM.
The atypical case is when a vendor is able to actually execute programs from directly within their platform. The benefit here is that each program inherits the configuration, tenancy, security and access rights directly from the platform, which helps reduce the complexity of the program, and increases it’s portability.
Ideally, your BI platform should have the ability to create its own developer ecosystem from a collection of third party works against the platform’s automation APIs.
Data visualization techniques are changing constantly.
Thanks to pioneers like USA Today and Steven Few, communicating information via visualizations is mainstream. The explosion in popularity of HTML5 has only made integrating new visualizers more popular. There are many, but our favorites are d3.js, Highcharts and bonsaiJS.
Platforms that offer the ability to extend their visualization capabilities easily will allow your organization to be as expressive as you can imagine.
Open platforms are nothing new, but in today’s cloud-centric world, they are certainly evolving to accommodate ideas from everywhere.
Written by GoodData Author |