Book Image

QlikView: Advanced Data Visualization

By : Miguel Angel Garcia, Barry Harmsen, Stephen Redmond, Karl Pover
Book Image

QlikView: Advanced Data Visualization

By: Miguel Angel Garcia, Barry Harmsen, Stephen Redmond, Karl Pover

Overview of this book

QlikView is one of the most flexible and powerful business intelligence platforms around, and if you want to transform data into insights, it is one of the best options you have at hand. Use this Learning Path, to explore the many features of QlikView to realize the potential of your data and present it as impactful and engaging visualizations. Each chapter in this Learning Path starts with an understanding of a business requirement and its associated data model and then helps you create insightful analysis and data visualizations around it. You will look at problems that you might encounter while visualizing complex data insights using QlikView, and learn how to troubleshoot these and other not-so-common errors. This Learning Path contains real-world examples from a variety of business domains, such as sales, finance, marketing, and human resources. With all the knowledge that you gain from this Learning Path, you will have all the experience you need to implement your next QlikView project like a pro. This Learning Path includes content from the following Packt products: • QlikView for Developers by Miguel Ángel García, Barry Harmsen • Mastering QlikView by Stephen Redmond • Mastering QlikView Data Visualization by Karl Pover
Table of Contents (25 chapters)
QlikView: Advanced Data Visualization
Contributors
Preface
Index

Chapter 1. Performance Tuning and Scalability

"The way Moore's Law occurs in computing is really unprecedented in other walks of life. If the Boeing 747 obeyed Moore's Law, it would travel a million miles an hour, it would be shrunken down in size, and a trip to New York would cost about five dollars. Those enormous changes just aren't part of our everyday experience."

— Nathan Myhrvold, former Chief Technology Officer at Microsoft, 1995

The way Moore's Law has benefitted QlikView is really unprecedented amongst other BI systems.

QlikView began life in 1993 in Lund, Sweden. Originally titled "QuickView", they had to change things when they couldn't obtain a copyright on that name, and thus "QlikView" was born.

After years of steady growth, something really good happened for QlikView around 2005/2006—the Intel x64 processors became the dominant processors in Windows servers. QlikView had, for a few years, supported the Itanium version of Windows; however, Itanium never became a dominant server processor. Intel and AMD started shipping the x64 processors in 2004 and, by 2006, most servers sold came with an x64 processor—whether the customer wanted 64-bit or not. Because the x64 processors could support either x86 or x64 versions of Windows, the customer didn't even have to know. Even those customers who purchased the x64 version of Windows 2003 didn't really know this because all of their x86 software would run just as well (perhaps with a few tweaks).

But x64 Windows was fantastic for QlikView! Any x86 process is limited to a maximum of 2 GB of physical memory. While 2 GB is quite a lot of memory, it wasn't enough to hold the volume of data that a true enterprise-class BI tool needed to handle. In fact, up until version 9 of QlikView, there was an in-built limitation of about 2 billion rows (actually, 2 to the power of 31) in the number of records that QlikView could load. On x86 processors, QlikView was really confined to the desktop.

x64 was a very different story. Early Intel implementations of x64 could address up to 64 GB of memory. More recent implementations allow up to 256 TB, although Windows Server 2012 can only address 4 TB. Memory is suddenly less of an obstacle to enterprise data volumes.

The other change that happened with processors was the introduction of multi-core architecture. At the time, it was common for a high-end server to come with 2 or 4 processors. Manufacturers came up with a method of putting multiple processors, or cores, on one physical processor. Nowadays, it is not unusual to see a server with 32 cores. High-end servers can have many, many more.

One of QlikView's design features that benefitted from this was that their calculation engine is multithreaded. That means that many of QlikView's calculations will execute across all available processor cores. Unlike many other applications, if you add more cores to your QlikView server, you will, in general, add more performance.

So, when it comes to looking at performance and scalability, very often, the first thing that people look at to improve things is to replace the hardware. This is valid of course! QlikView will almost always work better with newer, faster hardware. But before you go ripping out your racks, you should have a good idea of exactly what is going on with QlikView. Knowledge is power; it will help you tune your implementation to make the best use of the hardware that you already have in place.

The following are the topics we'll be covering in this chapter:

  • Reviewing basic performance tuning techniques

  • Generating test data

  • Understanding how QlikView stores its data

  • Looking at strategies to reduce the data size and to improve performance

  • Using Direct Discovery

  • Testing scalability with JMeter