Group items tagged

Schema on write This is the traditional approach for Business Intelligence. A model, often dimensional, is built as part of the design process. This model is an abstraction of the complexity of the underlying systems, put in business terms. The purpose of the model is to allow the business users to interrogate the data in a way they understand.

The model is instantiated through physical database tables and the date is loaded through an ETL (extract, transform and load) process that takes data from one or more source systems and transforms it to fit the model, then loads it into the model.

The key thing is that the model is determined before the data is finally written and the users are very much guided or driven by the model in how they query the data and what results they can get from the system. The designer must anticipate the queries and requests in advance of the user asking the questions.

OBIEE 11g skills, both in terms of new functionality (mapping, analyses, KPIs and Scorecards etc) and new infrastructure (WebLogic, EM, OPSS etc) A smattering of Essbase skills, focused mainly on the integration with OBIEE and Essbase (and the many workarounds and gotchas) Good ODI skills, both in terms of the basics, but also being able to write knowledge modules, integrate with OBIEE, deployment and migration Solid database skills – OBIEE gave the illusion through aggregates etc that database tuning was redundant, but time has shown it’s by far the biggest success factor in a project – get the database design and optimisation wrong, and your project is toast. You need to know partitioning, materialized views, index types, and increasingly, you need to get yourself on an Exadata project as customers are buying the technology but you can’t teach it to yourself at home BI Apps skills, but watch out for everything changing when BI Apps 11g comes out, and be prepared to learn the Fusion Apps and JDeveloper if you want to stay in the game Looking to the future, keep an eye on technologies such as in-memory (TimesTen), mid-tier caching (Coherence), plus technologies such as Business Activity Monitoring (BAM), “big data” (Hadoop, large data sets, NoSQL), complex event processing and maybe products such as Qlikview, just in case Oracle buys them, or at least to know what the competition are up to, or more importantly pitching to your boss

The other thing to bear in mind of course, if you’re an Oracle BI developer, is that you need to have great business, communication and data modeling skills.

The best practice in business intelligence delivery is always to build a data warehouse.

Pure transactional reporting is problematic. There are, of course, the usual performance issues. Equally troublesome is the difficulty in distilling a physical model down to a format that is easy for business users to understand. Dimensional models are typically the way business users envision their business: simple, inclusive structures for each entity. The standard OLTP data model that takes two of the four walls in the conference room to display will never make sense to your average business user.

f you are running a large Oracle data warehouse you should be using parallel

Like all tools you have to use parallel correctly; no more would we think of using a wrench to hammer a nail then should you think parallel is the answer to all performance problems. Sometimes parallel will make things worse, sometimes parallel will make performance less predictable.

Parallel introduces additional work to a query, simplistically we need to: split the query into multiple parallel processes, execute them, wait for the processes to complete and finally coordinate the results. This all takes time to do. Our time saving comes from being able to process multiple smaller chunks of data simultaneously. If the time to execute the step in parallel is not significantly faster than doing it without parallel then the additional overhead may make parallel processing a slower option; this is typically the case with small tables where a full tablescan or an indexed access is fast. Use too few parallel processes and we will not gain much in performance, too many and we risk starving the database of resource for other work or even slow our own process as it waits for resource. If you have implemented some form of CPU resource management on your system you may find that you experience delays as your parallel slaves ‘wait their turn’

What this means in practice is that when the BI Server component starts up, it creates and reserves a number of threads in advance, determined by a number of parameters including SERVER_THREAD_RANGE.

You can see these threads running and ready to perform tasks for the BI Server component by using a tool such as Process Explorer for Windows

Thinking it through a bit, any given single query is, to a certain extent, only really going to use a small part of the total amount of CPUs available on a server, because it’s not the BI Server that runs queries in parallel, it’s the underlying database. For example, a single analysis against a single Oracle Database datasource would only really need a single BI Server thread to handle the query request, but when the underlying database receives the query, it might use a large number of its CPUs to process the query, returning results back to the BI Server to then pass back to the Presentation Server for display to the user.

OBIEE, at the BI Server level. lets you define query limits that either warn or stop users from exceeding certain elapsed query times or number of rows returned. Assuming you define a “standard” group for most OBIEE users, you might want to stop them from displaying reports (requests) that return more than 50,000 rows, whilst you might want to warn them if their query takes over five minutes to run.