Feeds

Aggregates: the not-so-forgotten DBA issue

Reducing aggregates can relieve the headache

Securing Web Applications Made Simple and Scalable

Aggregates are probably the second biggest headache for database administrators in data warehouses after indexes and the tuning thereof.

However, while there has been lots of discussion about indexes there has been very little about aggregates. For example, the data warehouse appliance vendors remove indexes more or less completely, while the traditional suppliers have incrementally added more features to help the DBA with indexes and tuning, so that advisors will now suggest the building of new indexes when it is appropriate and then build them for you automatically.

Of course, the removal of indexes makes the database storage requirements significantly smaller, which is a major additional benefit but, nevertheless, apart from materialised views (which are the best part of a decade old) not much has been done to help DBAs with aggregates.

Perhaps I had better explain what aggregates are. Put simply, they are what it says on the tin. For example, the aggregate of all sales in a particular store over a particular period. In other words, exactly what you might store in an OLAP cube. However, for a variety of reasons (which we need not go into now), it is often advantageous to store these aggregates directly within the data warehouse and without using OLAP technology such as Analysis Services or Essbase.

What this means for the DBA is that he or she has to define and maintain all the hierarchies and dimensions along which aggregates have to be calculated. This is not only complex to establish in the first place, it is also a major ongoing headache. For example, every time a new product is launched or a new store is opened, or there is a company re-organisation, all the relevant aggregates have to be re-defined and amended.

What got me thinking about this was Netezza's recent user conference. Now, I was already aware that a number of companies have implemented a data warehouse appliance purely for the purposes of calculating aggregates as a front-end to a Teradata warehouse, but it came as more of a surprise to hear that a number of companies, of which Catalina Marketing and Carphone Warehouse are examples, have stopped using aggregates altogether as a result of implementing a Netezza solution.

Why? Simply, because Netezza performed so well for the relevant queries that the companies no long felt it necessary to pre-calculate them.

It turned out that there was another advantage as well. More than one company at the conference reported that they had a longer than expected testing cycle when they first implemented Netezza. Why? Because the results they were getting were different from those that they had previously got. I spoke to one of these companies about the reasons. After investigation, it turned out that the aggregates they had previously used were sometimes incorrect.

And if you think about this, it is hardly surprising: with a highly complex set of dimensions and hierarchies, not to mention all that maintenance, it is not unlikely that error will creep in at some point. Interestingly, the company also remarked that although the database it was previously using was supposed to be aggregate aware, in practice it found that the optimiser did not always use the aggregates that were in place.

To conclude: data warehouse appliances in general, and Netezza in particular, allow you to reduce or eliminate the use of aggregates if you want to. I would not go so far as to recommend their elimination (it will depend on circumstances), but the ability to minimise them should be a boon for DBAs.

Copyright © 2006, IT-Analysis.com

Bridging the IT gap between rising business demands and ageing tools

More from The Register

next story
NO MORE ALL CAPS and other pleasures of Visual Studio 14
Unpicking a packed preview that breaks down ASP.NET
Cheer up, Nokia fans. It can start making mobes again in 18 months
The real winner of the Nokia sale is *drumroll* ... Nokia
Mozilla fixes CRITICAL security holes in Firefox, urges v31 upgrade
Misc memory hazards 'could be exploited' - and guess what, one's a Javascript vuln
Put down that Oracle database patch: It could cost $23,000 per CPU
On-by-default INMEMORY tech a boon for developers ... as long as they can afford it
Google shows off new Chrome OS look
Athena springs full-grown from Chromium project's head
Apple: We'll unleash OS X Yosemite beta on the MASSES on 24 July
Starting today, regular fanbois will be guinea pigs, it tells Reg
HIDDEN packet sniffer spy tech in MILLIONS of iPhones, iPads – expert
Don't panic though – Apple's backdoor is not wide open to all, guru tells us
prev story

Whitepapers

Designing a Defense for Mobile Applications
Learn about the various considerations for defending mobile applications - from the application architecture itself to the myriad testing technologies.
Implementing global e-invoicing with guaranteed legal certainty
Explaining the role local tax compliance plays in successful supply chain management and e-business and how leading global brands are addressing this.
Top 8 considerations to enable and simplify mobility
In this whitepaper learn how to successfully add mobile capabilities simply and cost effectively.
Seven Steps to Software Security
Seven practical steps you can begin to take today to secure your applications and prevent the damages a successful cyber-attack can cause.
Boost IT visibility and business value
How building a great service catalog relieves pressure points and demonstrates the value of IT service management.