This is the second of a six-part review of the Pentaho BI suite. In each part of the review, we will take a look at the components that make up the BI suite, according to how they would be used in the real world.
In this second part, we'll be discussing the Pentaho BI Server from which all of the reports, dashboards, and analytic tools are served to the users. A BI suite usually has a central place where users log in using their assigned credentials. In this case, the server is a standalone web server (an Apache Tomcat instance) that is augmented by various tools that provides the functionalities – most of these tools are written by Webdetails (webdetails.pt). We'll visit these tools in subsequent review parts, for now, let's focus on the server itself.
In the case of Pentaho BI Server, it has two components:
The Pentaho User Console (a.k.a PUC) – this is what we usually associate with the main BI Server in the Pentaho world; where users would spend the majority of their time generating reports (both real-time or scheduled), using the analytic tools, build and publish dashboards, etc. This is also where administrator users can manage who can access which reports either by User or by Role – obviously, Role-based ACL is cleaner and easier to maintain.
The Administration Console (a.k.a PAC) – this is where admin users go to create new Users, Roles, and schedule jobs. It is another standalone web server that can be started and stopped when needed, it is totally independent of the main PUC server.
Is it Corporate-Ready?
BI servers are considered ready for corporate “demands” based on the number of users they can support, and the facilities to manage them. The Pentaho BI Suite Enterprise Edition is without a doubt ready for corporate use because it comes with the support that will make sure that is the case.
The Community Edition is more interesting, it is definitely corporate ready, but the personnels who set it up needs to be intimately familiar with the ins and outs of the server itself. Having installed three of these, I am confident that the BI Server, due to its built in ACL management is ready for prime time in the corporate world.
Although the Pentaho BI server includes a scheduler, another “corporate” feature, I find myself using cron (or Windows Task Scheduler) for the most part. The built-in scheduler is based on the Quartz library for Java. It is a good facility with decent UI to schedule reports or ETL from within the PUC.
Is it Easy to Use?
The PAC is very easy to use. The UI interface is simple enough due to the minimum numbers of menus and options. In a sense, it's a simple facility to manage user/role and scheduling – not ACL, just users and roles.
The PUC is more involved, but adopting the familiar file folder look and feel on the left panel, it is quite easy to get into and start using. Administrators would love the way they can set who can Execute, Edit, Schedule each reports, saved analytic views, and dashboards – by the way, Pentaho calls these: Solutions.
Setting up the BI server is better left to the consultants who are used to doing it. Or if there are in-house personnels who would be doing this, it is worth the time to participate in the training webinars that Pentaho held periodically. The steps to setup a BI server far from being simple, but that is the case for all BI servers, regardless the brand.
The collapsible left panel serves as the directory of the solutions, with the top part shows the folders, and the bottom part shows the individual solution. The bigger panel on the right is where you actually see the content of the solutions. And in some cases, that's where you'd create a Dashboard using the CDE tool (we'll revisit this in later review part).
Is it Easy to Create Solutions?
Remember that the concept “solution” here refer to the different types of reports, dashboards, analytic views. Pentaho BI server employs a “glue” scripting facility called the xactions. These are XML documents that contain some sequence of actions that can do various things like:
Asking users for input parameters
Issuing a SQL query based on user input
Trigger an ETL that produce reports
Once you are familiar with this facility, it is not that hard to start producing solutions, but it pays to install the included examples and study them to find out how to do certain things with xaction and/or to copy snippets into your own scripts.
On the PUC, we can build these solutions:
Dashboards using CDE
Ad-hoc reports and data model using the built in Model generator (very handy for accessing those BI tables that are populated by ETL runs)
Analytic Views using tools like Saiku or its equivalent for the Professional and Enterprise edition. NOTE: This requires a pre-published schema which is built using another tool called the schema-workbench (we will see this in the latter parts of this review series)
Is it Customizable?
Being the user-facing tool, one of the requirement would be the ability to customize the appearance via themes, at the very least, a BI server need to allow companies to change the logo into their own.
The good news is, you can do all that with Pentaho BI Server. If you opt for the Professional and Enterprise editions, you can rely on the support that you already paid for. For those using the Community Edition, customizing the appearance requires knowledge on how a typical Java Web Server is structured. Again, any good BI consultant should be able to tackle this without too much difficulties.
Here is an example of a customized PUC login page:
In case you are wondering, yes, you can customize the PUC interface also, and it even comes with a theme structure in which you can assign your graphic artists to redefine the CSS elements.
The Pentaho BI server, is the central place where users are going to interact with Pentaho BI Suite. It brings together solutions (what Pentaho call contents) produced by the other tools in the suite, and expose it to the user while being protected by a robust ACL.
On the balance between ease-of-use and the ability to customize, the Pentaho BI Server scores well provided that the personnel in charge is familiar with the Java Enterprise environment. To illustrate this, in one project, I managed to tweak the security framework to make the PUC part of a single-sign-on Liferay portal, along with other applications such as Opentaps and Alfresco.
Next in part-three, we will discuss the wide array of Pentaho Reporting tools.